Implementation Plan

Last updated: 2026-06-06

This file now tracks only active repo-level work. Detailed closure narratives live in the dedicated subsystem audits under docs/reverse-engineering/. Historical task headings that existing parity gates still check are preserved in the appendix as compact archival anchors instead of repeated full narratives.

Strategic goal

The long-term parity target remains the same: the reconstructed engine should, in theory, be able to replace the retail Quake Live engine, load retail cgamex86.dll, qagamex86.dll, and uix86.dll, present the retail main menu, and interoperate with the retail Windows host/runtime contracts. Quake Live-only online services remain behind QL_BUILD_ONLINE_SERVICES, default disabled, until a documented open replacement path exists.

Current planning baseline

• Treat AUDIT.md and docs/reverse-engineering/repo-wide-parity-audit-2026-04-21.md as the current cross-subsystem truth.
• Treat the 2026-04-10 engine-wide 100% report as the strict-retail Windows closure milestone, not as a whole-repo all-green claim.
• The strict-retail Windows replacement target remains 100% on the current worktree, and the current repo-wide parity estimate is 99% after the non-Windows portability lanes were closed as explicit compatibility-only containment. Remaining repo-wide uncertainty is validation freshness rather than active non-Windows source parity debt.
• The new file-by-file audit campaign now lives in docs/reverse-engineering/source-file-parity-ledger-2026-04-22.md and docs/reverse-engineering/source-file-parity-audit-plan-2026-04-22.md. It tracks 567 source entries, seeds concrete per-file notes for the documented RW-G01 divergence owners plus the active RW-G02 gap owners, and keeps the older subsystem ledgers as inherited baselines until each file is walked again.
• Top-level planning is reopened in this file because the 2026-04-21 repo-wide audit identified active gaps outside the strict-retail score.
• The older broad planning notes under docs/parity-plan.md, docs/ui_deltas.md, and docs/ui_followup_issues.md are historical snapshots, not current gap ledgers.
• The ownerdraw index checklists live in docs/reverse-engineering/ui-ownerdrawtype-parity-index.md and docs/reverse-engineering/cg-ownerdrawtype-parity-index.md; use them to pick and close remaining src/ui/menudef.h ownerdraw IDs.

Active work

Task A334: Pin qagame ai_team modal order and leader dispatcher wiring [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_qagame_ai_team_modal_order_parity.py, docs/reverse-engineering/botlib-qagame-ai-team-modal-order-mapping-2026-06-06.md Parity estimate: before 0% -> after 100% for focused qagame ai_team.c Obelisk/Harvester/leader/BotTeamAI alias coverage, before 88% -> after 98% for focused source-owned modal team-order reconstruction confidence, before 94% -> after 98% for focused leader/chat/EA/import-wiring confidence, and overall botlib plus qagame AI wiring confidence 98.3% -> 98.6%.

Completed work:

1. Promoted both FUN_... and sub_... aliases for BotObeliskOrders, BotHarvesterOrders, FindHumanTeamLeader, and BotTeamAI at 0x10029df0..0x1002aef0.
2. Cross-checked the checked-in source against HLIL evidence for defend/attack/harvest order splits, passive/aggressive caps, human leader discovery, whoisteamleader / iamteamleader chat, startleader voice emission, and the gametype-specific order dispatcher cadence.
3. Added a parity regression covering aliases, Ghidra function inventory, symbol-map signatures/comments, source anchors, Binary Ninja HLIL entry/flow anchors, Ghidra top-function anchors, and the configstring/AAS/chat/EA import bridge used by this band.
4. Documented that the source reconstruction already matches this retail slice, so no C source edits were required in this round.

Task A333: Pin ZMQ idZMQ host lifecycle and source wiring [COMPLETED]

Priority: High Primary areas: docs/reverse-engineering/quakelive_steam_mapping_round_420.md, tests/test_platform_services.py, src/code/server/sv_zmq.c Parity estimate: before 76% -> after 95% for focused idZMQ host lifecycle/publication/RCON wiring confidence, before 94.1% -> after 94.2% for ZMQ-related source reconstruction confidence, and overall Quake Live source parity 55.83% -> 55.84%.

Completed work:

1. Rechecked the older retail idZMQ host aliases from Rounds 94 through 97 against current HLIL, function inventory, and the source reconstruction.
2. Re-pinned the server-facing match-report/player-event wrappers, retained object methods, thin global wrappers, RCON broadcast/pump helpers, shutdown split, and peer-table ownership helpers.
3. Tied the retail wrapper and string evidence back to the current sv_zmq.c, sv_init.c, sv_main.c, sv_game.c, and common.c call sites without changing C source behavior.
4. Documented the CZMQ/libzmq inference boundary: retail uses embedded helper bands, while this repository keeps Quake Live's server integration source and dynamically resolves public zmq_* symbols for opted-in live transport.

Task A332: Pin qagame ai_team one-flag order and team dispatcher wiring [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_qagame_ai_team_oneflag_order_parity.py, docs/reverse-engineering/botlib-qagame-ai-team-oneflag-order-mapping-2026-06-06.md Parity estimate: before 0% -> after 100% for focused qagame ai_team.c dispatcher/group/1FCTF alias coverage, before 87% -> after 97% for focused source-owned one-flag and team-order reconstruction confidence, before 94% -> after 98% for focused Cvar/configstring/chat/EA import-wiring confidence, and overall botlib plus qagame AI wiring confidence 98.0% -> 98.3%.

Completed work:

1. Promoted both FUN_... and sub_... aliases for BotCTFOrders, BotCreateGroup, BotTeamOrders, the four Bot1FCTFOrders_* helpers, and Bot1FCTFOrders at 0x10027750..0x10029d90.
2. Cross-checked the checked-in source against HLIL evidence for the CTF flag-status dispatcher, group leader follow commands, cached sv_maxclients team enumeration, 1FCTF percentage splits, return/getflag voice order quirks, and neutralflagstatus dispatch.
3. Added a parity regression covering aliases, Ghidra function inventory, symbol-map signatures/comments, source anchors, Binary Ninja HLIL entry/flow anchors, Ghidra top-function anchors, BotTeamAI gametype dispatch, and the Cvar/configstring/chat/EA import bridge used by the band.
4. Documented that the source reconstruction already matches this retail slice, so no C source edits were required in this round.

Task A331: Pin ZMQ option switch source-constant wiring [COMPLETED]

Priority: High Primary areas: docs/reverse-engineering/quakelive_steam_mapping_round_419.md, tests/test_platform_services.py, src/code/server/sv_zmq.c Parity estimate: before 68% -> after 93% for focused ZMQ option-switch/source-constant linkage confidence, before 94.0% -> after 94.1% for ZMQ-related source reconstruction confidence, and overall Quake Live source parity 55.82% -> 55.83%.

Completed work:

1. Rechecked the retail options_t::setsockopt / options_t::getsockopt switch tables at 0x40E0E0 and 0x40E690 against the committed HLIL.
2. Pinned the option range, jump/lookup-table shapes, EINVAL rejection path, and the entries for ROUTER mandatory delivery (0x21), PLAIN server mode (0x2c), and ZAP domain strings (0x37).
3. Tied those retail option ids to the reconstructed QL_ZMQ_* constants used by the server-owned RCON and stats socket setup in src/code/server/sv_zmq.c.
4. Added a focused regression so future ZMQ edits cannot drift the source constants away from the retail option switch evidence.

Task A330: Classify ZMQ dependency ownership boundary [COMPLETED]

Priority: High Primary areas: docs/reverse-engineering/zmq-dependency-ownership-2026-06-06.md, tests/test_platform_services.py, README.md Parity estimate: before 61% -> after 92% for focused ZMQ dependency ownership confidence, before 94.0% -> after 94.0% for ZMQ-related source reconstruction confidence, and overall Quake Live source parity 55.82% -> 55.82%.

Completed work:

1. Investigated whether writable source contains actual libzmq implementation code or only the Quake Live server-owned idZMQ integration.
2. Confirmed the only ZMQ-named source file under src/code/ is src/code/server/sv_zmq.c, and that it dynamically resolves external libzmq symbols for opted-in online-service builds.
3. Documented the conclusion that no dependency should be installed under src/libs/ in this pass and no sv_zmq.c code should be cut from src/code/.
4. Added a regression guard against accidentally treating src/code as a vendored libzmq implementation space.

Task A329: Pin qagame ai_team CTF order and botlib wiring [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_qagame_ai_team_ctf_order_parity.py, docs/reverse-engineering/botlib-qagame-ai-team-ctf-order-mapping-2026-06-06.md Parity estimate: before 0% -> after 100% for focused qagame ai_team.c CTF order-band alias coverage, before 88% -> after 97% for focused source-owned team-order reconstruction confidence, before 94% -> after 98% for focused AAS/chat/EA import-wiring confidence, and overall botlib plus qagame AI wiring confidence 97.6% -> 98.0%.

Completed work:

1. Promoted both FUN_... and sub_... aliases for the retail teammate counting, base travel-time sorting, CTF task-preference, team-order chat, voice-chat, and first four CTF order helpers at 0x10025390..0x10026f20.
2. Cross-checked the source reconstruction against HLIL evidence for sv_maxclients caching, player configstring walks, AAS travel-time calls, task-preference name validation, defender/roamer/attacker partitioning, vsay/vtell and vosay/votell EA formatting, and the BotCTFOrders dispatcher edges.
3. Added a parity regression covering aliases, Ghidra function sizes, symbol-map signatures/comments, source anchors, Binary Ninja HLIL entry/flow anchors, Ghidra decompiler anchors, and the game/server import bridge for the AAS, chat, and EA calls used by the slice.
4. Documented FUN_100269d0 as a medium-high confidence mapping because HLIL reports unresolved stack usage, while the surrounding dispatcher, command-string, sort/preference, and symbol-map evidence still strongly support BotCTFOrders_EnemyFlagNotAtBase.

Task A328: Pin ZMQ clock and thread runtime wiring [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_418.md Parity estimate: before 48% -> after 93% for focused ZMQ clock/thread runtime wiring confidence, before 93.8% -> after 94.0% for ZMQ-related source reconstruction confidence, and overall Quake Live source parity 55.81% -> 55.82%.

Completed work:

1. Rechecked the cached millisecond clock helper, Windows tick-count fallback, thread trampoline, start wrapper, stop wrapper, and fallback lock atexit cleanup against HLIL around 0x40B950..0x40BB30 and 0x52A9E0..0x52B070.
2. Added aliases for the address-backed clock/thread helpers that connect socket_base_t timeout loops, select_t, io_thread_t, and reaper_t to the retail runtime scaffolding.
3. Documented the GetTickCount64 resolver/fallback path while leaving sub_40B950 as supporting evidence because it is not exposed as a normal Ghidra inventory row.
4. Added a parity regression covering alias rows, clock cache anchors, thread assertion anchors, select/io-thread/reaper callsites, and static initializer evidence for the fallback tick-count lock.

Task A327: Pin qagame ai_main lifecycle and training helper wiring [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_qagame_ai_main_lifecycle_training_parity.py, docs/reverse-engineering/botlib-qagame-ai-main-lifecycle-training-mapping-2026-06-06.md Parity estimate: before 8% -> after 100% for focused qagame ai_main.c lifecycle and training-helper alias coverage, before 92% -> after 98% for focused source-owned lifecycle reconstruction confidence, before 58% -> after 87% for focused retail-only training/dynamic-skill helper mapping confidence, and overall botlib plus qagame AI execution wiring confidence 97.1% -> 97.6%.

Completed work:

1. Promoted both FUN_... and sub_... aliases for the source-owned BotTestAAS, interbreed, team-leader, view-angle, input translation, per-bot think, scheduling, session, setup, shutdown, reset, map-load, botlib setup, and botlib shutdown rows.
2. Promoted mapped training and tutorial helper aliases for BotEntityBoundsGap, BotSetIdealViewAnglesToPoint, BotSetTrainingBotState, BotAppendDynamicSkillSample, BotUpdateDynamicSkill, BotApplyBeyondRealityTravelFlags, local/bot client selectors, training cvar helpers, and the training-state tail.
3. Added a parity regression covering aliases, Ghidra row sizes, symbol-map signatures/comments, reconstructed source anchors, Binary Ninja HLIL entry/flow anchors, Ghidra decompiler anchors for the larger retail-only helpers, and the botlib lifecycle, EA, chat, move, goal, weapon, entity, and user-command import wiring used by the slice.
4. Documented the source-reconstruction boundary for the retail-only dynamic-skill island, where names and call sites are now pinned but exact C reconstruction remains open.

Task A326: Pin ZMQ own_t owned pointer-set lifecycle wiring [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_417.md Parity estimate: before 42% -> after 94% for focused ZMQ own_t owned pointer-set lifecycle confidence, before 93.6% -> after 93.8% for ZMQ-related source reconstruction confidence, and overall Quake Live source parity 55.80% -> 55.81%.

Completed work:

1. Rechecked own_t construction, destruction, process_own, process_term_req, and process_term against HLIL around 0x40F050..0x40FAE0.
2. Added aliases for the pointer-node create/find/predecessor/rightmost, subtree free, right-rotation, range-erase, and full-tree clear helpers that back the owned child set.
3. Preserved the Round 148 shared pointer-set insert/erase/rebalance aliases while tying the remaining helper surface to the observed own.cpp lifecycle flow.
4. Documented the reconstructed container-level source shape and the inference boundary for the inline command opcode 0xB termination send sequence.

Task A325: Pin ZMQ object default command and poll-event assertion wiring [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_416.md Parity estimate: before 35% -> after 96% for focused ZMQ object-command default callback surface confidence, before 93.4% -> after 93.6% for ZMQ-related source reconstruction confidence, and overall Quake Live source parity 55.79% -> 55.80%.

Completed work:

1. Rechecked object_t, io_thread_t, and reaper_t default callback slots against HLIL around 0x40CF60..0x40DF00.
2. Added aliases for io-thread/reaper out_event and timer_event assertion callbacks with their source-file-specific assertion anchors.
3. Added aliases for all 16 base object_t::process_* fail-fast command handlers, matching the zmq_command_t_process opcode dispatch table and existing command send helpers.
4. Documented the reconstructed command vtable surface and the inference boundary for the process_attach slot.

Task A324: Pin qagame DMQ3 deathmatch setup and ai_main wrapper wiring [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, references/symbol-maps/qagame.json, tests/test_botlib_qagame_ai_dmq3_deathmatch_setup_parity.py, docs/reverse-engineering/botlib-qagame-ai-dmq3-deathmatch-setup-mapping-2026-06-06.md Parity estimate: before 0% -> after 100% for focused qagame deathmatch/setup and adjacent ai_main alias coverage, before 91% -> after 98% for focused source-owned deathmatch/setup/wrapper parity confidence, before 45% -> after 82% for focused retail-only tutorial helper mapping confidence, and overall botlib plus qagame AI execution wiring confidence 96.7% -> 97.1%.

Completed work:

1. Promoted both FUN_... and sub_... aliases for the source-owned BotDeathmatchAI, BotSetEntityNumForGoal, BotSetupDeathmatchAI, BotAI_Print, BotAI_Trace, BotAI_GetClientState, BotClientHasNoTargetFlag, BotAI_GetEntityState, and BotAI_BotInitialChat rows.
2. Promoted mapped-only retail helper aliases for BotSelectTormentTarget, BotResolveTourPoint, BotAcceptOffscreenEnemyCandidate, and BotCanSpawnTourPoint, without claiming source reconstruction for those retail-only bodies.
3. Corrected the BotSetEntityNumForGoal symbol-map comment to document the retail-preserved classname guard that skips exact classname matches and falls through for absent or non-matching classnames.
4. Added a parity regression covering aliases, Ghidra row sizes, symbol-map signatures/comments, source-body anchors, Binary Ninja HLIL entry/flow anchors, and the cvar, userinfo, trace, AAS/BSP, EA command, chat, stack dump, and level-item botlib wiring used by the slice.

Task A323: Pin ZMQ socket-base endpoint and pending compact-map wiring [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_415.md Parity estimate: before 44% -> after 90% for focused ZMQ socket-base endpoint and pending compact-map helper wiring, before 93.2% -> after 93.4% for ZMQ-related source reconstruction confidence, and overall Quake Live source parity 55.78% -> 55.79%.

Completed work:

1. Rechecked socket_base_t::bind, socket_base_t::connect, and socket_base_t::term_endpoint against HLIL around 0x408D90..0x40B360.
2. Added aliases for the shared socket_base_t::add_endpoint helper and the compact endpoint/pending node creation, subtree cleanup, iterator, range erase, and rotation helpers.
3. Preserved the older ctx and socket-base map aliases while distinguishing the socket-base compact layouts by their 0x30/0x31 endpoint and 0x2C/0x2D pending node sentinel offsets.
4. Documented the inferred source reconstruction shape for endpoint registration and unresolved inproc pending-connection storage.

Task A322: Pin qagame DMQ3 activate-goal, obstacle, snapshot, and alternate-route wiring [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_qagame_ai_dmq3_activate_obstacle_parity.py, docs/reverse-engineering/botlib-qagame-ai-dmq3-activate-obstacle-mapping-2026-06-06.md Parity estimate: before 80% -> after 100% for focused qagame ai_dmq3.c activate/obstacle/alternate-route alias coverage, before 92% -> after 98% for focused source-owned activate/obstacle/event parity confidence, before 91% -> after 98% for focused botlib activation, route, movement, avoid, and chat import wiring confidence, and overall botlib plus qagame AI execution wiring confidence 96.0% -> 96.7%.

Completed work:

1. Promoted the remaining retail qagame FUN_... and sub_... names for BotCheckConsoleMessages, BotAlternateRoute, BotGetAlternateRouteGoal, and BotSetupAlternativeRouteGoals.
2. Added a parity regression covering the contiguous 0x1001BCD0..0x1001EE30 qagame band from activate-goal builders through alternate-route setup.
3. Pinned the source bodies against Binary Ninja HLIL entries and flow anchors, Ghidra function sizes, symbol-map signatures, and representative botlib AAS, chat, movement, avoid, and alternate-route import wiring.
4. Documented the observed facts, inferred subsystem role, import wiring, and remaining confidence limits in the reverse-engineering notes.

Task A321: Pin ZMQ ctx endpoint and pending tree helper wiring [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_414.md Parity estimate: before 46% -> after 91% for focused ZMQ ctx endpoint and pending tree helper wiring, before 93.0% -> after 93.2% for ZMQ-related source reconstruction confidence, and overall Quake Live source parity 55.77% -> 55.78%.

Completed work:

1. Rechecked the ctx endpoint and pending inproc connection map helper band against HLIL around 0x404380..0x406820.
2. Added aliases for endpoint-tree find, get-or-insert, lower-bound, insertion, subtree cleanup, iterator, and rotation helpers.
3. Added aliases for pending-connection tree copy, creation, insertion, iterator, subtree cleanup, and rotation helpers, preserving the retail 0x280/0x281 pending-node layout distinction from endpoint nodes.
4. Re-pinned the options_t address-mask vector copy/destruction helpers reached by the ctx endpoint payload copy path.

Task A320: Pin ZMQ ctx runtime and inproc handoff wiring [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_413.md Parity estimate: before 36% -> after 93% for focused ZMQ ctx runtime and inproc-pending wiring, before 92.8% -> after 93.0% for ZMQ-related source reconstruction confidence, and overall Quake Live source parity 55.76% -> 55.77%.

Completed work:

1. Rechecked the ctx runtime helper band around 0x403120..0x404010 against socket destruction, I/O-thread selection, endpoint map cleanup, and inproc pending-connection call sites.
2. Added aliases for socket destruction, I/O-thread selection, endpoint-record destruction, options destruction, endpoint bulk-unregistration, and pending-connection delivery.
3. Corrected two stale compiler-noise aliases: sub_4038B0 is the options_t copy constructor, and sub_403BB0 is the ctx pending inproc connection helper rather than an STL locale helper.

Task A319: Pin qagame DMQ3 aim, attack, and map-script bridge [COMPLETED]

Priority: High Primary areas: src/code/game/ai_dmq3.c, references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_qagame_ai_dmq3_aim_attack_parity.py, docs/reverse-engineering/botlib-qagame-ai-dmq3-aim-attack-mapping-2026-06-06.md Parity estimate: before 0% -> after 96% for focused qagame ai_dmq3.c aim/attack alias coverage, before 88% -> after 97% for focused source-owned aim/attack/map-script parity confidence, before 92% -> after 97% for fire-control and prediction import wiring confidence, and overall botlib plus adjacent qagame AI execution wiring 95% -> 96%.

Completed work:

1. Mapped the retail qagamex86.dll 0x10019E00..0x1001BBE0 ai_dmq3.c bridge from BotAimAtEnemy through BotModelMinsMaxs.
2. Promoted both FUN_... and sub_... aliases for the aim controller, attack gate, map-script helper, movedir helper, and model-bounds helper.
3. Reconstructed the retail BotMapScripts fallback branch from mpq3tourney6 to beyondreality, matching HLIL and symbol-map evidence.
4. Added a parity gate covering aliases, Ghidra row sizes, symbol-map signatures, source-body anchors, Binary Ninja HLIL entry/flow anchors, and the predictive movement, weapon-info, visible-position, characteristic, EA attack, and EA view botlib wiring used by the slice.

Task A318: Pin ZMQ ctx constructor and endpoint-map wiring [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_412.md Parity estimate: before 41% -> after 94% for focused ZMQ ctx constructor, endpoint-map, and map-cleanup wiring, before 92.5% -> after 92.8% for ZMQ-related source reconstruction confidence, and overall Quake Live source parity 55.75% -> 55.76%.

Completed work:

1. Rechecked the ctx_t constructor body against the public zmq_ctx_new wrapper, ctx magic constants, mailbox setup, map sentinels, mutex initialization, and default option writes.
2. Added aliases for the ctx constructor, endpoint registration/removal/lookup, endpoint/pending-connection map constructor-unwind destructors, and the endpoint/pending range-erase helpers.
3. Added a static parity guard tying the constructor and cleanup aliases to ctx.cpp strings, the existing ctx destructor, Ghidra rows, endpoint map errno behavior, map sentinel sizes, and command-mailbox teardown evidence.

Task A317: Pin qagame DMQ3 roaming, visibility, and carrier helpers [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_qagame_ai_dmq3_visibility_parity.py, docs/reverse-engineering/botlib-qagame-ai-dmq3-visibility-mapping-2026-06-06.md Parity estimate: before 0% -> after 95% for focused qagame ai_dmq3.c visibility-band alias coverage, before 90% -> after 97% for focused source-owned visibility-band parity confidence, before 91% -> after 96% for movement/visibility import wiring confidence, and overall botlib plus adjacent qagame AI execution wiring 94% -> 95%.

Completed work:

1. Mapped the retail qagamex86.dll 0x10017D20..0x10019D30 ai_dmq3.c slice from BotRoamGoal through BotEnemyCubeCarrierVisible.
2. Promoted both FUN_... and sub_... aliases for the roam, attack-movement, same-team, field-of-view, entity-visibility, enemy-acquisition, flag-carrier, and cube-carrier helper rows.
3. Added a parity gate covering aliases, Ghidra row sizes, symbol-map signatures, source-body anchors, Binary Ninja HLIL entry/flow anchors, and the AAS contents, characteristic, move-to-goal, and move-in-direction botlib import wiring used by the slice.

Task A316: Pin ZMQ Windows err.cpp helper wiring [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_411.md Parity estimate: before 34% -> after 91% for focused ZMQ Windows err.cpp helper wiring, before 92.4% -> after 92.5% for ZMQ-related source reconstruction confidence, and overall Quake Live source parity 55.74% -> 55.75%.

Completed work:

1. Rechecked the early ZMQ Windows err.cpp helper band against HLIL bodies, Ghidra function rows, imports, and source-file string anchors.
2. Added aliases for the fatal abort helper, WSA last-error string dispatch, Win32 FormatMessageA formatter, and Winsock-to-errno mapper.
3. Added a static parity guard tying the helpers to public zmq_poll, stream-engine send/recv handling, TCP listener setup, and retained err.cpp assertion lines.

Task A315: Pin ZMQ select fd-vector and pollout reset wiring [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_410.md Parity estimate: before 38% -> after 92% for focused ZMQ select-poller fd-vector and pollout reset wiring, before 92.3% -> after 92.4% for ZMQ-related source reconstruction confidence, and overall Quake Live source parity 55.73% -> 55.74%.

Completed work:

1. Rechecked the deferred select_t fd-entry vector helper slab left by Round 370 against select.cpp HLIL, Ghidra rows, and the select vtable/ source-file anchors.
2. Added aliases for the stream-engine-facing pollout reset helper and the fd-entry vector grow/reserve/allocate/copy helpers.
3. Added a static parity guard tying the aliases to add/remove fd behavior, select-loop fd-set copies, retired-slot compaction, and stream-engine output backpressure call sites.

Task A314: Pin qagame AI DMQ3 support-band mapping [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_qagame_ai_dmq3_support_parity.py, docs/reverse-engineering/botlib-qagame-ai-dmq3-support-mapping-2026-06-06.md, src/code/game/ai_dmq3.c, src/code/game/g_public.h, src/code/game/g_syscalls.c, src/code/server/sv_game.c, src/code/server/ql_game_imports.inc Parity estimate: before 0% -> after 95% for focused qagame ai_dmq3.c support-band alias coverage, before 90% -> after 97% for source-owned support-band parity confidence, before 90% -> after 96% for focused botlib support import wiring confidence, and overall botlib plus adjacent qagame AI execution wiring 93% -> 94%.

Completed work:

1. Promoted the retail qagame ai_dmq3.c support-band identities over 0x10015A40..0x10017CD0, covering AAS point-area helpers, client-name helpers, weapon and movement setup, inventory and battle-inventory updates, item use, observer and hazard probes, waypoints, aggression/retreat/chase, rocket-jump, persistent-powerup, camping, avoid-goal, and powerup helpers.
2. Added a parity gate that checks alias promotion against Ghidra rows, symbol-map names/signatures, Binary Ninja HLIL entry and flow anchors, and direct source-body anchors for every promoted helper.
3. Pinned the related AAS, EA, characteristic, avoid-goal, camp-spot, level-item, movement-state, and weapon botlib import wiring from qagame syscall constants through native Quake Live import slots and generated QL_G_trap_* wrappers.

Task A313: Pin ZMQ io_object default callback wiring [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_409.md Parity estimate: before 45% -> after 94% for focused ZMQ io_object_t default callback-surface mapping, before 92.2% -> after 92.3% for ZMQ-related source reconstruction confidence, and overall Quake Live source parity 55.72% -> 55.73%.

Completed work:

1. Rechecked the deferred Round 377 assertion-only io_object_t callbacks against retained io_object.cpp HLIL bodies and the shared i_poll_events/io_object_t vtable order.
2. Added aliases for the default in_event, out_event, and timer_event fail-fast callback handlers.
3. Added a static parity guard tying the names to Ghidra rows, HLIL assert(false) bodies, and reuse by REQ/session, TCP listener, and stream-engine subobject vtables.

Task A312: Pin qagame AI DMQ3 team-goal mapping [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_qagame_ai_dmq3_team_goal_parity.py, docs/reverse-engineering/botlib-qagame-ai-dmq3-team-goal-mapping-2026-06-06.md, src/code/game/ai_dmq3.c, src/code/game/g_public.h, src/code/game/g_syscalls.c, src/code/server/sv_game.c, src/code/server/ql_game_imports.inc Parity estimate: before 0% -> after 95% for focused qagame ai_dmq3.c team-goal alias coverage, before 88% -> after 96% for source-owned team-goal source parity confidence, before 90% -> after 96% for focused botlib/AAS/EA wiring into team-goal selection, and overall botlib plus adjacent qagame AI execution wiring 92% -> 93%.

Completed work:

1. Promoted the retail qagame ai_dmq3.c team-goal helper identities over 0x10013BE0..0x10015960, covering team predicates, ordered-task restore, refusal/status helpers, CTF/1FCTF/Obelisk/Harvester seek and retreat dispatchers, and BotTeamGoals.
2. Added a parity gate that checks alias promotion against Ghidra rows, symbol-map names/signatures, Binary Ninja HLIL entry and cross-call anchors, and direct source-body anchors for every promoted helper.
3. Pinned the related AAS travel-time and EA action botlib import wiring from qagame syscall constants through native Quake Live import slots and generated QL_G_trap_* wrappers.

Task A311: Pin ZMQ message pipe queue tail wiring [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_408.md Parity estimate: before 70% -> after 94% for focused ZMQ message pipe queue tail wiring, before 92.1% -> after 92.2% for ZMQ-related source reconstruction confidence, and overall Quake Live source parity 55.71% -> 55.72%.

Completed work:

1. Rechecked the deferred ypipe_conflate_t<msg_t,256> final vtable callback that Binary Ninja renders as operator new[].
2. Added aliases for the conflating message-pipe probe callback and the non-deleting ypipe_base_t<msg_t,256> destructor.
3. Added a static parity guard tying the aliases to Ghidra rows, HLIL lock/probe body, vtable slot order, and analysis-symbol RTTI/vtable anchors.

Task A310: Pin ZMQ endpoint container lifecycle wiring [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_407.md Parity estimate: before 48% -> after 93% for focused ZMQ endpoint container lifecycle wiring, before 91.9% -> after 92.1% for ZMQ-related source reconstruction confidence, and overall Quake Live source parity 55.70% -> 55.71%.

Completed work:

1. Rechecked the deferred endpoint lifecycle helpers against socket-base endpoint allocation, TCP protocol detection, concrete TCP address ownership, endpoint URI rendering, and session-base destructor cleanup.
2. Added aliases for the endpoint container constructor/destructor at sub_41AAE0 and sub_41AB60.
3. Added a static parity guard tying the aliases to Ghidra rows, retained HLIL constructor/destructor bodies, the tcp literal, renderer dispatch, and session cleanup evidence.

Task A309: Pin qagame AI DMNet tutorial-tail mapping [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, references/symbol-maps/qagame.json, tests/test_botlib_qagame_ai_dmnet_tutorial_tail_parity.py, tests/test_ctf_oneflag_retail_parity.py, docs/reverse-engineering/botlib-qagame-ai-dmnet-tutorial-tail-mapping-2026-06-06.md, src/code/game/ai_dmq3.c, src/code/game/ai_main.c, src/code/game/g_cmds.c, src/code/game/g_public.h, src/code/game/g_syscalls.c, src/code/server/sv_game.c, src/code/server/ql_game_imports.inc Parity estimate: before 0% -> after 94% for focused qagame tutorial-tail alias coverage, before 85% -> after 98% for source-owned BotCTFCarryingFlag retail predicate parity, before 87% -> after 94% for focused training/botlib movement and EA wiring confidence, and overall botlib plus adjacent qagame AI execution wiring 91% -> 92%.

Completed work:

1. Promoted the retail-only qagame ai_dmnet.c tutorial-tail identities over 0x1000ECC0..0x10013B00, including teammate seeking, teammate leading, torment-human, fragbait lead, instagib, BotCTFCarryingFlag, and BotTeam.
2. Reconstructed the safe source-owned BotCTFCarryingFlag predicate so it accepts the retail second CTF-like gametype slot, represented in current source as GT_ATTACK_DEFEND, alongside GT_CTF.
3. Added a parity gate that checks alias promotion against Ghidra rows, symbol-map names/signatures, Binary Ninja HLIL entry/call anchors, Ghidra decompile strings, source-owned helper bodies, training cvars, and botlib movement/EA import wiring.
4. Documented the remaining source gap: the retail tutorial node bodies require a future bot_state_t tail-layout reconstruction before they can be compiled accurately into ai_dmnet.c.

Task A308: Pin ZMQ TCP address endpoint parser wiring [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_406.md Parity estimate: before 57% -> after 91% for focused ZMQ TCP address endpoint parser wiring, before 91.7% -> after 91.9% for ZMQ-related source reconstruction confidence, and overall Quake Live source parity 55.69% -> 55.70%.

Completed work:

1. Rechecked tcp_address.cpp endpoint parsing against HLIL source-file strings, promoted tcp_address_t/tcp_address_mask_t vtable symbols, Ghidra function rows, and existing listener/connecter call-site evidence.
2. Added 5 aliases for NIC-name resolution, default and sockaddr-backed address construction, the address destructor body, and listener-side address mask matching.
3. Re-pinned existing TCP address resolve/stringify and address-mask resolve/stringify aliases so listener getsockname reporting, accept filtering, and options mask parsing remain tied to retail evidence.

Task A307: Pin qagame AI DMNet seek/battle node mapping [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_qagame_ai_dmnet_battle_parity.py, docs/reverse-engineering/botlib-qagame-ai-dmnet-battle-node-mapping-2026-06-06.md, src/code/game/ai_dmnet.c, src/code/game/g_public.h, src/code/game/g_syscalls.c, src/code/server/sv_game.c, src/code/server/ql_game_imports.inc Parity estimate: before 50% -> after 96% for focused qagame ai_dmnet.c seek/battle node alias coverage, before 88% -> after 96% for focused botlib goal-stack/movement import wiring into qagame battle flow, and overall botlib plus adjacent qagame AI execution wiring 90% -> 91%.

Completed work:

1. Completed the qagame ai_dmnet.c seek/battle decision-network alias set over 0x1000C6E0..0x1000E700, filling the missing AIEnter_* aliases around the already-promoted seek and battle node functions.
2. Added a parity gate that checks alias promotion against Ghidra rows, symbol-map names/signatures, Binary Ninja HLIL entry/call anchors, reconstructed source anchors, and botlib goal-stack/movement import wiring.
3. Documented the evidence trail and the observed qagame-to-botlib movement wiring in the dedicated reverse-engineering note for this slice.

Task A306: Pin ZMQ subscription trie/mtrie wiring [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_405.md Parity estimate: before 40% -> after 88% for focused ZMQ subscription trie/mtrie wiring, before 91.5% -> after 91.7% for ZMQ-related source reconstruction confidence, and overall Quake Live source parity 55.68% -> 55.69%.

Completed work:

1. Rechecked XPUB/XSUB subscription-prefix ownership against HLIL call sites, source-file strings, Ghidra rows, and the round-404 DIST routing aliases.
2. Added 8 aliases covering mtrie_t teardown, pipe-set destruction, node deletion, redundancy testing, XPUB mtrie-to-DIST matching, trie_t teardown, trie node deletion, and trie apply/walk behavior.
3. Re-pinned existing mtrie_t add/remove and trie_t add/remove aliases so prefix insertion, recursive pruning, compressed child-range maintenance, and XSUB subscription replay remain tied to retail evidence.

Task A305: Pin qagame AI DMNet goal-prelude mapping [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_qagame_ai_dmnet_goal_parity.py, docs/reverse-engineering/botlib-qagame-ai-dmnet-goal-prelude-mapping-2026-06-06.md, src/code/game/ai_dmnet.c, src/code/game/g_public.h, src/code/game/g_syscalls.c, src/code/server/sv_game.c, src/code/server/ql_game_imports.inc Parity estimate: before 5% -> after 96% for focused qagame ai_dmnet.c LTG/node prelude alias coverage, before 86% -> after 95% for focused botlib goal/movement import wiring into qagame node flow, and overall botlib plus adjacent qagame AI execution wiring 89% -> 90%.

Completed work:

1. Promoted the qagame ai_dmnet.c long-term-goal and node-entry prelude over 0x100083C0..0x1000C640, including air-goal selection, nearby-goal selection, LTG fallback, intermission/stand/respawn entry, path clearing, activate-entity seeking, and seek-NBG entry.
2. Added a parity gate that checks alias promotion against Ghidra rows, symbol-map names/signatures, Binary Ninja HLIL entry/call anchors, reconstructed source anchors, and botlib goal/movement import wiring.
3. Documented the evidence trail and the observed qagame-to-botlib wiring in the dedicated reverse-engineering note for this slice.

Task A304: Pin ZMQ FQ/LB/DIST pipe routing helper wiring [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_404.md Parity estimate: before 42% -> after 90% for focused ZMQ FQ/LB/DIST pipe routing helper wiring, before 91.2% -> after 91.5% for ZMQ-related source reconstruction confidence, and overall Quake Live source parity 55.67% -> 55.68%.

Completed work:

1. Rechecked fq.cpp, lb.cpp, and dist.cpp helper bodies against HLIL entry points, source-file strings, Ghidra function rows, and existing pipe_t/queue aliases.
2. Added 17 aliases covering fair-queue attach/activate/terminate/probe, load-balancer terminate/activate/probe, distribution attach/activate/ terminate/send/write, and the FQ/LB vector push helpers.
3. Re-pinned existing zmq_fq_t_recvpipe and zmq_lb_t_sendpipe names so the recovered routing layer remains tied to read/write vtable dispatch, multipart state handling, flush/termination behavior, and EAGAIN fallbacks.

Task A303: Pin qagame AI command team-order mapping [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_qagame_ai_cmd_parity.py, docs/reverse-engineering/botlib-qagame-ai-cmd-team-order-mapping-2026-06-06.md, src/code/game/ai_cmd.c, src/code/server/sv_game.c, src/code/server/ql_game_imports.inc Parity estimate: before 0% -> after 95% for focused qagame ai_cmd.c team-order alias coverage, before 82% -> after 94% for focused botlib match-parser wiring into qagame command handling, and overall botlib plus adjacent qagame AI/chat command wiring 88% -> 89%.

Completed work:

1. Promoted 39 qagame team-message and team-order identities, with both FUN_... and sub_... aliases, across 0x10004930..0x100080A0.
2. Cross-checked the slice against symbol-map names/signatures, Ghidra function rows, HLIL entry/call anchors, and reconstructed ai_cmd.c source bodies.
3. Added a focused parity gate for team-goal lookup, time parsing, addressed message filtering, patrol waypoint parsing, team-order LTG transitions, subteam/checkpoint/formation/leadership handlers, CTF status handling, and the central BotMatchMessage dispatcher.
4. Pinned the botlib match parser import wiring used by the command layer: BOTLIB_AI_FIND_MATCH and BOTLIB_AI_MATCH_VARIABLE through game syscall mapping, server import initialization, and generated qagame wrappers.

Task A302: Pin ZMQ message pipe backing queue wiring [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_403.md Parity estimate: before 37% -> after 89% for focused ZMQ message pipe backing queue wiring, before 90.9% -> after 91.2% for ZMQ-related source reconstruction confidence, and overall Quake Live source parity 55.66% -> 55.67%.

Completed work:

1. Rechecked pipe.cpp queue allocation and vtable evidence for pipepair, pipe_t, normal ypipe_t<msg_t,256>, conflating ypipe_conflate_t<msg_t,256>, and yqueue<msg_t,256>.
2. Added 28 aliases covering conflating write/read helpers, pipe scalar and array-item destructor thunks, normal and conflating message ypipe constructors/destructors/callbacks, shared ypipe flush publishing, and message yqueue chunk allocation/destruction/push/unpush.
3. Re-pinned 7 existing pipe_t owners so the recovered queue callbacks remain tied to pipepair allocation, pipe construction/destruction, read/write, rollback, and hiccup queue replacement.

Task A301: Pin qagame bot chat prologue mapping [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_chat_parity.py, docs/reverse-engineering/botlib-qagame-chat-prologue-mapping-2026-06-06.md, src/code/game/ai_chat.c Parity estimate: before 0% -> after 96% for focused qagame bot chat/ranking prologue alias coverage, before 90% -> after 93% for focused botlib plus qagame chat wiring confidence, and overall botlib plus adjacent qagame AI/chat wiring 87% -> 88%.

Completed work:

1. Promoted 24 qagame bot chat/ranking prologue identities, with both FUN_... and sub_... aliases, across 0x10001000..0x100033F0.
2. Cross-checked the slice against qagame owner metadata, imports, exports, symbol-map names/signatures, Ghidra function rows, HLIL entry/call anchors, and the reconstructed ai_chat.c source bodies.
3. Added a focused parity gate for the active-player, ranking, opponent-name, map-title, weapon-name, visible-enemy, valid-chat-position, chat-handler, chat-time, and chat-test source/HLIL shapes.

Task A300: Pin ZMQ socket-base default virtual/interface wiring [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_402.md Parity estimate: before 54% -> after 91% for focused socket_base_t default virtual/interface wiring, before 90.7% -> after 90.9% for ZMQ-related source reconstruction confidence, and overall Quake Live source parity 55.65% -> 55.66%.

Completed work:

1. Rechecked the shared socket_base_t primary vtable and the array_item_t<0>, i_poll_events, and i_pipe_events subobject vtables against the retail HLIL and Ghidra function rows.
2. Added 17 aliases covering the socket-base scalar deleting destructor, process-destroy state transition, default unsupported xsetsockopt/xsend/xrecv callbacks, assert-false activation/timer callbacks, poll in-event command pumping, pipe-event forwarding, pipe-terminated cleanup, and subobject destructor thunks.
3. Re-pinned 4 existing socket-base and pending-connection helpers so command-pump, destroyed-flag teardown, complete destruction, and pending connection erase behavior remain tied to the shared callback evidence.

Task A299: Pin ZMQ PAIR/STREAM socket-family wiring [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_401.md, src/code/server/sv_zmq.c Parity estimate: before 35% -> after 88% for focused PAIR/STREAM socket-family wiring, before 90.4% -> after 90.7% for ZMQ-related source reconstruction confidence, and overall Quake Live source parity 55.64% -> 55.65%.

Completed work:

1. Rechecked the Quake Live Steam libzmq factory edges for PAIR (0) and STREAM (11) against HLIL, Ghidra function rows, promoted vtable symbols, existing stream aliases, and the reconstructed server ZMQ service context.
2. Added 22 aliases covering PAIR constructor/destructor, singleton-pipe attach/termination, send/receive/probe callbacks, PAIR interface thunks, STREAM constructor/destructor, pipe attachment, termination, write activation, and STREAM interface thunks.
3. Re-pinned 5 existing shared/STREAM aliases so identity-framed STREAM send/receive/prefetch behavior remains tied to factory-call evidence, source-file strings, vtable slots, and function rows.

Task A298: Pin ROQ cinematic continuation beyond botlib boundary [COMPLETED]

Priority: High Primary areas: tests/test_botlib_structure_tail_cgame_boundary_parity.py, src/code/client/cl_cin.c, references/analysis/quakelive_symbol_aliases.json, docs/reverse-engineering/botlib-roq-cinematic-continuation-boundary-mapping-2026-06-06.md Parity estimate: before 68% -> after 99% for focused ROQ cinematic continuation boundary classification, before 0% -> after 98% for direct decodeCodeBook/RoQInterrupt source and HLIL coverage inside botlib boundary gates, and overall botlib plus adjacent client/cinematic wiring 86% -> 87%.

Completed work:

1. Extended the botlib structure-tail/cgame-boundary gate from ROQ_GenYUVTables through the adjacent retail 0x004B1010..0x004B3510 cinematic decode/control slab.
2. Pinned the promoted adjacent owners from yuv_to_rgb and decodeCodeBook through RoQInterrupt, CIN_RunCinematic, CIN_PlayCinematic, and SCR_RunCinematic.
3. Added direct source anchors for the large ROQ codebook and interrupt dispatch bodies, including sample-depth branches, VQ expansion calls, streamed-read loop handling, audio decode branches, quad-info dispatch, and oversized-frame guard behavior.
4. Rechecked the continuation against committed Ghidra function rows, HLIL call/constant anchors, and cl_cin.c source ownership.
5. Verified that promoted aliases from the botlib support tail through this full cinematic continuation now all have direct test_botlib_*.py mentions.

Task A297: Pin ZMQ publication socket-family wiring [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_400.md, src/code/server/sv_zmq.c Parity estimate: before 46% -> after 89% for focused PUB/SUB/PULL/PUSH and XPUB/XSUB socket-family wiring, before 90.0% -> after 90.4% for ZMQ-related source reconstruction confidence, and overall Quake Live source parity 55.63% -> 55.64%.

Completed work:

1. Rechecked the Quake Live Steam libzmq factory cases for PUB (1), SUB (2), PULL (7), PUSH (8), XPUB (9), and XSUB (10) against HLIL, Ghidra function rows, vtable symbols, existing aliases, and the retained server stats QL_ZMQ_PUB source usage.
2. Added 39 aliases covering PUB/SUB wrappers, PULL fair-queue receive, PUSH load-balancer send, and the XPUB/XSUB scalar/interface destructor thunks that complete their vtable coverage.
3. Re-pinned 23 existing XPUB/XSUB and shared activation/output aliases so the publication pipeline remains tied to factory-call evidence, source-file strings, vtable slots, function rows, and retained stats publication code.

Task A296: Pin post-structure botlib boundary into client timing and ROQ [COMPLETED]

Priority: High Primary areas: tests/test_botlib_structure_tail_cgame_boundary_parity.py, src/code/client/cl_cgame.c, src/code/client/cl_cin.c, references/analysis/quakelive_symbol_aliases.json, docs/reverse-engineering/botlib-post-structure-client-cinematic-boundary-mapping-2026-06-06.md Parity estimate: before 86% -> after 99% for focused post-structure botlib/client ownership-boundary classification, before 72% -> after 99% for focused cgame timing and ROQ cinematic helper coverage inside botlib boundary gates, and overall botlib plus adjacent parser/client wiring 85% -> 86%.

Completed work:

1. Extended the botlib structure-tail/cgame-boundary gate past CL_GetServerCommand into the retail 0x004B0610..0x004B0F60 slab.
2. Pinned the promoted adjacent owners: CL_GameCommand, CL_CGameRendering, CL_AdjustServerTimeDelta, CL_FirstSnapshot, CL_SetCGameTime, RllDecodeStereoToStereo, move8_32, blit8_32, blitVQQuad32fs, and ROQ_GenYUVTables.
3. Explicitly bounded retail 0x004B0A50 as the source-modeled cgame native import integrity guard row, not as a botlib owner or promoted alias.
4. Rechecked source anchors in cl_cgame.c and cl_cin.c against the committed Quake Live HLIL and Ghidra function rows.
5. Verified that promoted aliases from the botlib support tail through this adjacent boundary range now all have direct test_botlib_*.py mentions.

Task A295: Pin ZMQ REP/router socket-family wiring [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_399.md, src/code/server/sv_zmq.c Parity estimate: before 39% -> after 88% for focused REP/router socket-family wiring, before 89.7% -> after 90.0% for ZMQ-related source reconstruction confidence, and overall Quake Live source parity 55.62% -> 55.63%.

Completed work:

1. Rechecked the Quake Live Steam libzmq factory cases for REP (4) and ROUTER (6) against HLIL, Ghidra function rows, vtable symbols, existing aliases, and the retained server RCON QL_ZMQ_ROUTER source usage.
2. Added 20 REP/router aliases and re-pinned 4 existing ROUTER aliases covering constructor/destructor ownership, REP send/receive state gating, ROUTER pipe attachment, option handling, termination, activation, identity lookup, and multipart send/receive callbacks.
3. Added a focused regression guard tying the aliases to factory-call evidence, source-file strings, vtable slots, function rows, and the retained RCON ROUTER socket construction.

Task A294: Reconstruct botlib import callback surface delete identity [COMPLETED]

Priority: High Primary areas: src/code/server/sv_bot.c, src/code/game/botlib.h, tests/test_botlib_import_callback_surface_parity.py, tests/test_botlib_server_game_bridge_parity.py, tests/test_botlib_internal_parity.py, tests/test_engine_cvar_retail_parity.py, docs/reverse-engineering/botlib-import-callback-surface-mapping-2026-06-06.md Parity estimate: before 94% -> after 99% for focused botlib import callback surface mapping, before 85% -> after 99% for focused debug line/polygon callback identity, and overall botlib plus related server/game wiring 97% -> 97%.

Completed work:

1. Rechecked the SV_BotInitBotLib callback table at retail 0x004DD940 against the Quake Live HLIL locals, promoted aliases, Ghidra rows, botlib_import_t field order, and source assignment order.
2. Reconstructed the retail delete callback identity by assigning botlib_import.DebugLineDelete directly to BotImport_DebugPolygonDelete, matching retail's shared sub_4dd430 table entry for both line and polygon delete.
3. Preserved the compatibility BotImport_DebugLineDelete wrapper while keeping the runtime callback table faithful to retail pointer identity.
4. Added focused parity coverage for trace copying, BSP metadata callbacks, memory/hunk callbacks, filesystem callback placement, and debug line/polygon callback behavior.
5. Normalized the botlib parity gates' Ghidra functions.csv references to the committed references/reverse-engineering/ghidra/quakelive_steam/ corpus while keeping alias-map lookups on the existing quakelive_steam_srp key.

Task A293: Pin botlib debug draw native micro-slab [COMPLETED]

Priority: High Primary areas: src/code/game/g_public.h, src/code/game/g_syscalls.c, src/code/game/ai_main.c, src/code/game/botlib.h, src/code/server/sv_game.c, src/code/botlib/be_interface.c, src/code/botlib/be_ai_move.c, tests/test_botlib_debug_draw_native_micro_slab_parity.py, docs/reverse-engineering/botlib-debug-draw-native-micro-slab-mapping-2026-06-06.md Parity estimate: before 72% -> after 99% for focused botlib debug draw native micro-slab mapping, before 90% -> after 97% for focused direct-only botlib native import classification, and overall botlib plus qagame native import wiring 96% -> 97%.

Completed work:

1. Rechecked native slots 83 and 84 against the Quake Live HLIL native import table, Ghidra row sizes, source native enum, qagame direct import wrappers, and the server native import initializer.
2. Pinned the odd retail relationship where these native table slots sit between the AAS tail and EA slab, while their botlib export offsets follow the AI movement/weapon/genetic tail at 0x1e8 and 0x1ec.
3. Verified that the helpers are direct-native only and do not belong in G_MapNativeImport as legacy syscall-compatible IDs.
4. Tied the qagame BotTestAAS caller to the botlib AI export bodies for BotDrawDebugAreas and BotDrawAvoidSpots.

Task A292: Pin botlib AI movement/weapon native tail [COMPLETED]

Priority: High Primary areas: src/code/game/g_public.h, src/code/game/g_syscalls.c, src/code/server/sv_game.c, src/code/server/ql_game_imports.inc, tests/test_botlib_ai_movement_weapon_native_tail_parity.py, docs/reverse-engineering/botlib-ai-movement-weapon-native-tail-mapping-2026-06-06.md Parity estimate: before 78% -> after 99% for focused AI movement/weapon/genetic native tail mapping, before 95% -> after 97% for focused botlib native import table coverage, and overall botlib plus qagame native import wiring 95% -> 96%.

Completed work:

1. Rechecked native slots 166 through 184 against the Quake Live HLIL import table, Ghidra row sizes where present, source native enum, legacy syscall map, server import initializer, and generated qagame wrappers.
2. Pinned the non-linear AddAvoidSpot ordering: wrapper order before BotMoveToGoal, retail native slot 177 after InitMoveState, and retail trampoline address 0x004e23d0.
3. Preserved the raw-thunk classifications for AllocMoveState at 0x4e2500 and AllocWeaponState at 0x4e25b0.
4. Verified the movement float ABI paths for AddAvoidSpot, MoveInDirection, and MovementViewTarget.
5. Added focused parity coverage to keep botlib AI slot 184 from drifting into the adjacent Quake Live service import slab.

Task A291: Reconstruct botlib AI goal native slab float marshaling [COMPLETED]

Priority: High Primary areas: src/code/game/g_syscalls.c, src/code/server/ql_game_imports.inc, tests/test_botlib_ai_goal_native_slab_parity.py, docs/reverse-engineering/botlib-ai-goal-native-slab-mapping-2026-06-06.md Parity estimate: before 76% -> after 99% for focused AI goal/item native slab mapping, before 90% -> after 96% for focused botlib float syscall/native marshaling coverage, and overall botlib plus qagame native import wiring 94% -> 95%.

Completed work:

1. Rechecked native slots 137 through 165 against the Quake Live HLIL import table, Ghidra row sizes where present, source native enum, legacy syscall map, server import initializer, and generated qagame wrappers.
2. Pinned the retail native table-order differences for avoid-goal reset/remove wrappers and camp/map/level goal lookup wrappers.
3. Preserved the non-row classifications for BotInitLevelItems at 0x4e22a0 and raw BotUpdateEntityItems table address 0x4e22b0.
4. Reconstructed BotMutateGoalFuzzyLogic float argument marshaling through PASSFLOAT(range) and QL_G_PASSFLOAT(range).
5. Added focused parity coverage to prevent the goal/item native slab from being reordered by legacy export offset or losing float ABI conversions.

Task A290: Reconstruct botlib AI character/chat native slots [COMPLETED]

Priority: High Primary areas: src/code/game/g_public.h, src/code/game/g_syscalls.c, src/code/server/sv_game.c, src/code/server/ql_game_imports.inc, tests/test_botlib_ai_character_chat_native_slab_parity.py, tests/test_botlib_chat_parity.py, tests/test_game_native_export_helper_parity.py, docs/reverse-engineering/botlib-ai-character-chat-native-slab-mapping-2026-06-06.md Parity estimate: before 74% -> after 99% for focused AI character/chat native slab mapping, before 88% -> after 95% for focused botlib native import direct-vs-compat classification, and overall botlib plus qagame native import wiring 93% -> 94%.

Completed work:

1. Rechecked native slots 110 through 136 against the Quake Live HLIL import table, Ghidra row sizes, source native enum, legacy syscall map, server import initializer, and generated qagame wrappers.
2. Reconstructed G_QL_IMPORT_BOTLIB_AI_CHAT_LENGTH = 126 and mapped BOTLIB_AI_CHAT_LENGTH directly to QL_G_trap_BotChatLength.
3. Reconstructed G_QL_IMPORT_BOTLIB_AI_STRING_CONTAINS = 129 and mapped BOTLIB_AI_STRING_CONTAINS directly to QL_G_trap_StringContains.
4. Preserved the raw-thunk classification for BotAllocChatState at 0x4e1d80 instead of promoting a fake Ghidra function row.
5. Updated stale compatibility-only coverage and added focused parity coverage to prevent the character/chat native import slab from regressing.

Task A289: Reconstruct client SteamUserStats float descriptor lane [COMPLETED]

Priority: High Primary areas: src/code/client/cl_main.c, src/common/platform/platform_steamworks.c, src/common/platform/platform_steamworks.h, tests/steamworks_harness.c, tests/test_steamworks_harness.py, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_383.md Parity estimate: before 86% -> after 99% for focused client STATS descriptor-control-flow parity, before 99% -> after 99.5% for client SteamUserStats readback wrapper coverage, and broader Steamworks parity remains approximately 99% because live backend values, localized achievement metadata, and production service availability are outside this deterministic source/harness pass.

Completed work:

1. Rechecked the retained users.stats callback descriptor walk at HLIL anchors 0x0046006b, 0x00460074, 0x004600ef, 0x00460103, 0x0046010e, 0x0046008d, and 0x00460149.
2. Parsed the committed HLIL data table rooted at data_55da98 and confirmed data_55da8c = 0x58, seven-dword rows, and all 88 shipped descriptor discriminators set to zero.
3. Added QL_Steamworks_GetUserStatFloat for the client SteamUserStats vtable slot 0x44 / 4, plus the disabled output-clearing fallback.
4. Converted the retained client stat catalog into explicit { name, isFloat } descriptors and preserved the current catalog as integer rows.
5. Updated the STATS JSON builder to retain the retail float/int branch while keeping the shipped row behavior unchanged.
6. Extended the executable Steamworks harness and ctypes coverage for the float readback slot, including invalid-input guards, output clearing, forwarding, failure propagation, and disabled fallback behavior.

Task A288: Pin botlib AAS native wrapper slab [COMPLETED]

Priority: High Primary areas: src/code/game/g_public.h, src/code/game/g_syscalls.c, src/code/server/sv_game.c, src/code/server/ql_game_imports.inc, tests/test_botlib_aas_native_wrapper_slab_parity.py, docs/reverse-engineering/botlib-aas-native-wrapper-slab-mapping-2026-06-06.md Parity estimate: before 70% -> after 98% for focused AAS native wrapper slab mapping, before 92% -> after 94% for focused qagame botlib native import table coverage, and overall botlib plus qagame native import wiring 92% -> 93%.

Completed work:

1. Rechecked native slots 61 through 82 against the Quake Live HLIL import table, Ghidra row sizes, source native enum, legacy syscall map, server import initializer, and generated qagame wrappers.
2. Pinned the retail table-order detail where AAS_BBoxAreas and AAS_AreaInfo occupy slots 61 and 62 before the earlier-addressed AAS_EntityInfo trampoline at slot 63.
3. Preserved the raw-thunk classification for AAS_Initialized at 0x4e1840 and AAS_Time at 0x4e1860 instead of promoting fake Ghidra function rows.
4. Verified the wrapper marshaling details for AAS_Time and AAS_PredictClientMovement, including the float bit-cast and QL_G_PASSFLOAT(frametime) handoff.
5. Added focused parity coverage to prevent AAS native imports from being reordered by name/address or disconnected from the legacy syscall bridge.

Task A287: Extend SteamUserStats readback harness coverage [COMPLETED]

Priority: High Primary areas: src/common/platform/platform_steamworks.c, tests/steamworks_harness.c, tests/test_steamworks_harness.py, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_382.md Parity estimate: before 58% -> after 98% for focused client SteamUserStats readback harness parity, before 96% -> after 99% for retained users.stats readback wrapper evidence confidence, and broader Steamworks parity remains approximately 99% because live backend values, localized achievement metadata, and the observed-but-unwrapped float-stat lane are outside this deterministic harness pass.

Completed work:

1. Rechecked the retained users.stats callback JSON builder at HLIL anchors 0x0046008d, 0x0046018e, 0x004601a6, and 0x004601c6.
2. Added deterministic mock state for user stat value, achievement state, unlock time, display attribute text, last SteamID, last stat/achievement name, last display-attribute key, call counts, and result toggles.
3. Added client SteamUserStats harness slots for display attributes, integer stat reads, and achievement reads at 0x30 / 4, 0x48 / 4, and 0x50 / 4.
4. Exported enabled/disabled harness wrappers for QL_Steamworks_GetUserStatInt, QL_Steamworks_GetUserAchievement, and QL_Steamworks_GetAchievementDisplayAttribute.
5. Added ctypes and static parity coverage tying the HLIL evidence, Round 188 source reconstruction note, production wrappers, mock vtable slots, executable test, plan entry, and Round 382 note together.

Task A286: Pin native-only botlib EA action slab [COMPLETED]

Priority: High Primary areas: src/code/game/g_public.h, src/code/game/g_syscalls.c, src/code/game/g_local.h, src/code/game/botlib.h, src/code/botlib/be_ea.c, src/code/botlib/be_interface.c, src/code/server/sv_game.c, src/code/server/ql_game_imports.inc, tests/test_botlib_ea_native_action_slab_parity.py, docs/reverse-engineering/botlib-ea-native-action-slab-mapping-2026-06-06.md Parity estimate: before 78% -> after 99% for focused EA native action slab mapping, before 84% -> after 96% for focused native-only botlib import classification, and overall botlib plus qagame native import wiring 91% -> 92%.

Completed work:

1. Rechecked native slots 85 through 109 against the Quake Live HLIL import table, Ghidra row sizes, source native enum, server import initializer, and generated qagame wrappers.
2. Pinned G_QL_IMPORT_BOTLIB_EA_WALK = 89 as a direct-native-only import.
3. Verified that no legacy BOTLIB_EA_WALK enum member or trap_EA_Walk wrapper should be reconstructed in g_syscalls.c / g_local.h.
4. Cross-checked the ea_export_t source chain from ACTION_WALK through EA_Walk, Init_EA_Export, and QL_G_trap_EA_Walk.
5. Added focused parity coverage to prevent the native EA slab from being reordered by name or backfilled with a fake legacy syscall.

Task A285: Extend SteamUserStats clear/reset harness coverage [COMPLETED]

Priority: High Primary areas: src/common/platform/platform_steamworks.c, tests/steamworks_harness.c, tests/test_steamworks_harness.py, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_375.md Parity estimate: before 45% -> after 98% for focused SteamUserStats clear/reset harness parity, before 96% -> after 99% for retained stats_clear wrapper evidence confidence, and broader Steamworks parity remains approximately 99% because live backend behavior is outside this opt-in reconstruction pass.

Completed work:

1. Rechecked the retained stats_clear command owner at 0x00460520 and its SteamUserStats() slot 0x54 call at 0x00460531.
2. Added deterministic mock state for reset-call count, last achievementsToo value, and configurable reset result.
3. Added QLR_SteamUserStats_ResetAllStats to the mock SteamUserStats vtable at 0x54 / 4 and exported enabled/disabled harness wrappers for QL_Steamworks_ClearStats.
4. Added ctypes coverage for the pre-initialisation guard, boolean forwarding, failure propagation, and disabled-build fallback behavior.
5. Added static parity coverage and docs/reverse-engineering/quakelive_steam_mapping_round_375.md to pin the HLIL evidence, production offset, mock vtable, executable test, and plan note together.

Task A284: Correct botlib-adjacent qagame world import slots [COMPLETED]

Priority: High Primary areas: src/code/game/g_public.h, src/code/game/g_syscalls.c, src/code/server/sv_game.c, src/code/server/ql_game_imports.inc, tests/test_botlib_qagame_world_import_slab_parity.py, docs/reverse-engineering/botlib-qagame-world-import-slab-mapping-2026-06-06.md Parity estimate: before 76% -> after 98% for focused qagame world/import slab native slot parity, before 90% -> after 91% for focused botlib-adjacent qagame import environment coverage, and overall botlib plus related server/game wiring 90% -> 91%.

Completed work:

1. Rechecked the qagame world/import native table at 0x0056CFFC..0x0056D02C against Quake Live HLIL and promoted aliases.
2. Reconstructed direct native import slots for G_IN_PVS_IGNORE_PORTALS, G_AREAS_CONNECTED, and G_ENTITY_CONTACT.
3. Corrected the native enum/server import order so slot 39 is G_QL_IMPORT_LINKENTITY and slot 40 is G_QL_IMPORT_UNLINK_ENTITY.
4. Bound the recovered slots in G_MapNativeImport and SV_InitGameImports while leaving EntityContactCapsule on the legacy compatibility path.
5. Added focused parity coverage for aliases, Ghidra row sizes, HLIL wrapper shapes, native table order, syscall remap, qagame wrappers, and source initializer assignments.

Task A283: Extend client Steam identity and SteamUtils harness coverage [COMPLETED]

Priority: High Primary areas: src/common/platform/platform_steamworks.c, tests/steamworks_harness.c, tests/test_steamworks_harness.py, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_373.md Parity estimate: before 68% -> after 98% for focused client Steam identity/SteamUtils harness parity, before 96% -> after 99% for bootstrap identity/app-id wrapper evidence confidence, and broader Steamworks parity remains approximately 99% because live backend behavior is outside this opt-in reconstruction pass.

Completed work:

1. Rechecked the retained local SteamID helper at 0x00460550, persona cvar sync helper at 0x00460610, IP-country helper at 0x00460690, and SteamUtils()->GetAppID callsites at 0x00431c48 and 0x00460dd6.
2. Added deterministic mock call counters for SteamFriends persona name, SteamUser identity, SteamUtils IP country, and SteamUtils app ID.
3. Exported enabled/disabled harness wrappers for persona-name, IP-country, app-id, and user-SteamID lookup.
4. Added ctypes coverage for persona/country copying, app-id forwarding, SteamID low/high projection, unavailable-user failure behavior, and disabled-build output clearing.
5. Added static parity coverage and docs/reverse-engineering/quakelive_steam_mapping_round_373.md to pin the HLIL evidence, production offsets, mock vtable, executable test, and plan note together.

Task A282: Bind and pin server/qagame botlib bridge slots [COMPLETED]

Priority: High Primary areas: src/code/game/g_public.h, src/code/game/g_syscalls.c, src/code/server/sv_game.c, src/code/server/sv_bot.c, tests/test_botlib_server_game_bridge_parity.py, docs/reverse-engineering/botlib-server-game-bridge-mapping-2026-06-06.md Parity estimate: before 72% -> after 100% for focused server/qagame botlib bridge alias coverage, before 82% -> after 98% for focused native qagame botlib import-slot source parity, and focused botlib plus server/game wiring 89% -> 90%.

Completed work:

1. Cross-checked the promoted SV_Bot*, BotImport_*, and QL_G_trap_Bot* aliases against the committed Ghidra row sizes and Quake Live HLIL.
2. Reconstructed native qagame import slots 44, 47, and 48 as G_QL_IMPORT_BOT_FREE_CLIENT, G_QL_IMPORT_DEBUG_POLYGON_CREATE, and G_QL_IMPORT_DEBUG_POLYGON_DELETE.
3. Routed G_BOT_FREE_CLIENT, G_DEBUG_POLYGON_CREATE, and G_DEBUG_POLYGON_DELETE through G_MapNativeImport and SV_InitGameProgs.
4. Added focused source/HLIL/table coverage for the server bot lifecycle, botlib setup/shutdown, reliable command consumption, snapshot entity lookup, and qagame native import table.
5. Added a dedicated mapping note and a range scan requiring direct test_botlib_*.py mentions for the promoted bridge aliases.

Task A281: Extend SteamFriends enumeration harness coverage [COMPLETED]

Priority: High Primary areas: src/common/platform/platform_steamworks.c, tests/steamworks_harness.c, tests/test_steamworks_harness.py, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_372.md Parity estimate: before 72% -> after 98% for focused SteamFriends friend-list harness parity, before 98% -> after 99% for retained browser friend-summary wrapper evidence confidence, and broader Steamworks parity remains approximately 99% because live backend behavior is outside this opt-in reconstruction pass.

Completed work:

1. Rechecked the retained GetFriendList HLIL block at 0x0043355d through 0x00433a00, confirming SteamFriends slots 0x0c, 0x10, 0x1c, 0x18, 0x14, 0x2c, 0xb4, and 0x20.
2. Added deterministic SteamFriends mock enumeration state plus GetFriendCount and GetFriendByIndex vtable slots at 0x0c / 4 and 0x10 / 4.
3. Exported enabled/disabled harness wrappers for friend count, by-index identity lookup, friend summary, and persona-name lookup.
4. Added ctypes coverage for flag forwarding, SteamID projection, invalid-index clearing, persona-name copying, full friend-summary projection, and offline fallback behavior.
5. Added static parity coverage and docs/reverse-engineering/quakelive_steam_mapping_round_372.md to pin the HLIL evidence, production offsets, mock vtable, executable test, and plan note together.

Task A280: Split Freeze winning-team thaw respawn path [COMPLETED]

Priority: High Primary areas: src/code/game/g_active.c, tests/test_game_active_pmove_wiring_parity.py, tests/test_game_round_controller_helper_parity.py, docs/reverse-engineering/freeze-tag-winning-thaw-respawn-mapping-2026-06-06.md, docs/reverse-engineering/freeze-tag-mapping-2026-06-06.md, docs/reverse-engineering/qagame-mapping.md Parity estimate: before 55% -> after 96% for the focused configured winning-team thaw respawn lane. The broader Freeze thaw-respawn tail moves approximately 88% -> 94%, while broader Freeze Tag gametype wiring moves approximately 95% -> 96%.

Completed work:

1. Rechecked FUN_1004C1B0 / Freeze_RoundStateTransition in the committed qagame Ghidra corpus and pinned the pm_type == PM_FREEZE branch that emits EV_THAW_PLAYER, writes PM_NORMAL, and tailcalls ClientSpawn.
2. Added G_FreezeRespawnThawedWinner in g_active.c to mirror that direct round-controller thaw visual and respawn handoff.
3. Routed G_FreezeThawWinningPlayers through the new helper instead of G_FreezeThawClient, keeping winning-team thaw distinct from assisted, admin, and auto-thaw paths.
4. Added focused source/evidence tests for the direct helper and updated Freeze round-controller parity coverage.
5. Added a dedicated mapping note and updated the umbrella Freeze/qagame maps.

Task A279: Pin botlib-adjacent native cgame import slab [COMPLETED]

Priority: High Primary areas: src/code/client/cl_cgame.c, src/code/cgame/cg_public.h, references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_cgame_native_import_slab_parity.py, docs/reverse-engineering/botlib-cgame-native-import-slab-mapping-2026-06-06.md Parity estimate: before 45% -> after 100% for focused native cgame import-slab alias coverage, before 78% -> after 97% for focused import-table source initializer coverage, and focused botlib plus cgame-boundary related wiring 88% -> 89%.

Completed work:

1. Widened the botlib-related alias/test scan to the native cgame slab at 0x004AF820..0x004B0500.
2. Pinned all promoted aliases in that range against quakelive_symbol_aliases.json and Ghidra row sizes where rows exist.
3. Added HLIL wrapper-shape anchors for command, collision-model, sound, renderer, advertisement bridge, snapshot/input, parser, cinematic, social, avatar, and adjacent lifecycle owners.
4. Cross-checked CG_QL_IMPORT_COUNT = 128 and the reconstructed CL_InitCGameImports source initializer assignments.
5. Added a final scan gate confirming no promoted names in this slab remain without direct test_botlib_*.py coverage; reserved retail-only rows remain fail-closed rather than overpromoted.

Task A278: Close botlib parser-tail promoted alias coverage [COMPLETED]

Priority: High Primary areas: src/code/botlib/l_precomp.c, src/code/botlib/l_script.c, src/code/botlib/l_script.h, references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_parser_tail_coverage_parity.py, docs/reverse-engineering/botlib-parser-tail-coverage-mapping-2026-06-06.md Parity estimate: before 82% -> after 100% for focused parser-tail promoted-alias test coverage, before 78% -> after 97% for focused PC_EvaluateTokens source-shape coverage, before 74% -> after 96% for focused l_script.c escape/number/expect-type source-shape coverage, and overall botlib parser/support static mapping 88% -> 89%.

Completed work:

1. Widened the botlib alias/test scan to 0x004A83C0..0x004AF820 and isolated the four remaining promoted parser-tail aliases without direct botlib test mentions: PC_EvaluateTokens, PS_ReadEscapeCharacter, PS_ReadNumber, and PS_ExpectTokenType.
2. Pinned those helpers against quakelive_symbol_aliases.json, functions.csv, HLIL function headers, and retail callsites in the quakelive_steam.exe part03/part04 HLIL split.
3. Cross-checked reconstructed source shape for the precompiler expression evaluator, script escape decoder, numeric token scanner, and token-type validator.
4. Added a final support-tail scan gate confirming no aliases in this botlib parser/support band remain without a direct test_botlib_*.py mention.
5. Added a focused mapping note; no C source body changes or alias changes were justified.

Task A277: Extend SteamFriends voice-speaking wrapper coverage [COMPLETED]

Priority: High Primary areas: tests/steamworks_harness.c, tests/test_steamworks_harness.py, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_368.md, docs/plans/steamworks-parity-plan.md Parity estimate: before 60% -> after 97% for focused SteamFriends voice-speaking wrapper harness parity, before 98% -> after 99% for retained client voice command wrapper evidence confidence, and broader Steamworks parity remains approximately 99% because live Steam backend behavior and modern networking replacement work remain outside this opt-in reconstruction pass.

Completed work:

1. Rechecked the retained SteamFriends voice-speaking evidence at 0x00460441 and 0x004604dc.
2. Added mock state and a SteamFriends vtable slot at 0x6c / 4 for SetInGameVoiceSpeaking.
3. Exported stable enabled and disabled harness wrappers for QL_Steamworks_SetInGameVoiceSpeaking.
4. Added ctypes coverage for SteamID word-combination, speaking-on/off forwarding, call counts, and the offline fallback contract.
5. Added static parity coverage tying the source wrapper slot, mock vtable, executable coverage, HLIL evidence, and round 368 mapping note together.

Task A276: Close residual promoted botlib alias test coverage [COMPLETED]

Priority: High Primary areas: src/code/botlib/be_aas_cluster.c, src/code/botlib/be_aas_debug.c, src/code/botlib/be_ai_chat.c, references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_residual_promoted_alias_coverage.py, docs/reverse-engineering/botlib-residual-promoted-alias-coverage-2026-06-06.md Parity estimate: before 70% -> after 100% for focused residual promoted-alias test coverage, before 82% -> after 96% for focused AAS cluster/debug residual source-shape coverage, before 84% -> after 97% for focused chat match-piece and integrity-tail source-shape coverage, and overall botlib static mapping/test coverage 87% -> 88%.

Completed work:

1. Rescanned the promoted core botlib alias band from 0x004829C0 through 0x004A83C0 and isolated the five remaining direct test-mention gaps: AAS_CheckAreaForPossiblePortals, AAS_ShowFacePolygon, BotFreeMatchPieces, BotCheckInitialChatIntegrety, and BotCheckReplyChatIntegrety.
2. Pinned those helpers against quakelive_symbol_aliases.json, functions.csv, HLIL function headers, and their retail callsites in quakelive_steam.exe_hlil_part03.txt.
3. Cross-checked the reconstructed source shape for cluster portal candidate detection, debug face polygon drawing, chat match-piece cleanup, and the initial/reply chat integrity walkers.
4. Added a final promoted-core botlib scan gate confirming no aliases in this band remain without a direct test_botlib_*.py mention.
5. Added a focused mapping note; no C source body changes or alias changes were justified.

Task A275: Extend SteamUser voice wrapper harness coverage [COMPLETED]

Priority: High Primary areas: tests/steamworks_harness.c, tests/test_steamworks_harness.py, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_367.md, docs/plans/steamworks-parity-plan.md Parity estimate: before 72% -> after 97% for focused SteamUser voice wrapper harness parity, before 96% -> after 98% for client Steam voice transport evidence confidence, and broader Steamworks parity remains approximately 99% because live Steam microphone/backend behavior and modern networking replacement work remain outside this opt-in reconstruction pass.

Completed work:

1. Rechecked the retained SteamUser voice slot evidence at 0x0046044c, 0x00460459, 0x004604b1, 0x00460d4b, and 0x00461b07.
2. Added deterministic harness mock state and SteamUser vtable slots for start, stop, compressed voice, decompression, and optimal sample rate.
3. Exported stable enabled and disabled harness wrappers for the production Steam voice API family.
4. Added ctypes coverage for payload projection, output-size failure zeroing, compressed/uncompressed flags, sample-rate forwarding, call counts, and the offline fallback contract.
5. Added static parity coverage tying the source wrapper slots, mock vtable, harness coverage, HLIL evidence, and round 367 mapping note together.

Task A274: Pin late botlib movement support mapping [COMPLETED]

Priority: High Primary areas: src/code/botlib/be_aas_move.c, src/code/botlib/be_aas_reach.c, src/code/botlib/be_ai_move.c, references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_movement_late_support_parity.py, docs/reverse-engineering/botlib-movement-late-support-mapping-2026-06-06.md Parity estimate: before 68% -> after 96% for focused late AAS movement/support mapping, before 72% -> after 96% for focused late travel dispatch coverage, and overall botlib plus movement/reachability wiring 86% -> 87%.

Completed work:

1. Recomputed the botlib row coverage and confirmed the direct core remains closed except for the documented libjpeg false leads and folded 0x00486F40 weapon-jump thunk.
2. Identified promoted movement/reachability aliases that lacked focused parity-test coverage.
3. Pinned AAS_DropToFloor, AAS_AgainstLadder, AAS_OnGround, AAS_ApplyFriction, AAS_ClipToBBox, AAS_GetJumpPadInfo, AAS_Reachability_EqualFloorHeight, and AAS_FindFaceReachabilities against aliases, Ghidra rows, HLIL anchors, and source shape.
4. Pinned the late travel handlers BotTravel_Elevator, BotTravel_FuncBobbing, BotFinishTravel_WeaponJump, BotTravel_JumpPad, BotFinishTravel_JumpPad, BotReachabilityTime, and BotMoveInGoalArea, plus their BotMoveToGoal dispatch wiring.
5. Added a focused parity gate and mapping note; no C source body changes or alias changes were justified.

Task A273: Extend legacy Steam P2P read-boundary coverage [COMPLETED]

Priority: High Primary areas: tests/test_steamworks_harness.py, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_366.md, docs/plans/steamworks-parity-plan.md Parity estimate: before 96% -> after 98% for focused legacy P2P read-boundary harness parity, before 98% -> after 98.5% for legacy Steam P2P wrapper evidence confidence, and broader Steamworks parity remains approximately 99% with live backend behavior and modern networking replacement design still outside this opt-in reconstruction pass.

Completed work:

1. Rechecked the retained SteamNetworking and SteamGameServerNetworking availability/read slot evidence in the retail HLIL and existing Steamworks mapping rounds.
2. Extended the client P2P harness path so a staged packet rejected by a too-small read buffer clears stale output size and remote SteamID values.
3. Extended the server P2P harness path with the matching too-small buffer rejection coverage.
4. Added static parity coverage for import names, client/server vtable slots, mock read-output clearing, and the round 366 evidence note.
5. Left production Steamworks wrappers unchanged because this pass proved local harness boundary behavior rather than a new retail slot or live SDK edge.

Task A272: Close botlib-adjacent cgame shutdown import boundary [COMPLETED]

Priority: High Primary areas: src/code/client/cl_cgame.c, src/code/client/cl_main.c, src/code/client/client.h, src/code/qcommon/common.c, src/code/cgame/cg_public.h, references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_cgame_shutdown_import_boundary_parity.py, docs/reverse-engineering/botlib-cgame-shutdown-import-boundary-mapping-2026-06-06.md Parity estimate: before 70% -> after 98% for focused cgame shutdown owner mapping, before 80% -> after 92% for adjacent cgame native residual import classification, and overall botlib plus related parser/import/lifecycle wiring 85% -> 86%.

Completed work:

1. Rechecked 0x004AF820..0x004B0500 across HLIL, Ghidra functions.csv, source lifecycle owners, and native cgame import wiring.
2. Promoted 0x004AFBF0 -> CL_ShutdownCGame, matching the source KEYCATCH_CGAME, CG_SHUTDOWN, VM_Free, and cgvm = NULL lifecycle sequence.
3. Bounded the remaining unaliased cgame-native import rows at 0x004AFFC0, 0x004B00C0, and 0x004B0370 as residual table trampolines whose exact public meanings remain weaker than their slot neighborhoods.
4. Confirmed the current source keeps the unresolved retail-only slots behind QL_CG_trap_RetailReservedImport, preserving fail-closed behavior instead of inventing live service or renderer behavior.
5. Added a focused parity gate and mapping note for the botlib-adjacent cgame shutdown/import boundary; no C source body changes were justified.

Task A271: Extend Steam UGC call-result failure projection coverage [COMPLETED]

Priority: High Primary areas: tests/steamworks_harness.c, tests/test_steamworks_harness.py, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_364.md, docs/plans/steamworks-parity-plan.md Parity estimate: before 58% -> after 96% for focused UGC call-result failure projection coverage, before 94% -> after 98% for combined UGC query callback pump projection, and broader Steamworks parity remains approximately 99% because live backend timing, workshop availability, and the raw filter semantic remain intentionally bounded.

Completed work:

1. Rechecked the retained SteamUGC create/send query slots and the SteamUGCQueryCompleted_t call-result owner.
2. Added a harness queue helper for UGC call-result payloads with explicit ioFailure and payload-presence controls.
3. Added executable coverage for raw failed results, ioFailure, no-payload failure, and no-payload non-IO failure projections.
4. Pinned call-handle preservation and zeroed no-payload event projection.
5. Added static dispatcher assertions for the mapped call-result fields and failure branches.

Task A270: Close botlib structure-tail and cgame boundary mapping [COMPLETED]

Priority: High Primary areas: src/code/botlib/l_struct.c, src/code/client/cl_cgame.c, src/code/client/ql_cgame_imports.inc, src/code/cgame/cg_syscalls.c, references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_structure_tail_cgame_boundary_parity.py, docs/reverse-engineering/botlib-structure-tail-cgame-boundary-mapping-2026-06-06.md Parity estimate: before 78% -> after 98% for focused botlib structure-tail boundary classification, before 72% -> after 97% for adjacent cgame snapshot/configstring owner naming, and overall botlib plus related parser/import wiring 84% -> 85%.

Completed work:

1. Rechecked 0x004AE830..0x004AF820 across HLIL, Ghidra functions.csv, botlib source, and client/cgame import wiring.
2. Promoted 0x004AF570 -> CL_GetSnapshot and 0x004AF690 -> CL_ConfigstringModified, separating true client owners from the existing native cgame import wrappers.
3. Confirmed ReadStructure remains the final promoted botlib structure-reader owner in this seam.
4. Classified ReadChar, ReadString, WriteIndent, WriteFloat, WriteStructWithIndent, and WriteStructure as source-visible but unpromoted in this retail address band.
5. Added a focused parity gate and mapping note for the structure-tail to cgame-import boundary; no C source body changes were justified.

Task A269: Extend Steam UGC query result readback coverage [COMPLETED]

Priority: High Primary areas: tests/steamworks_harness.c, tests/test_steamworks_harness.py, docs/reverse-engineering/quakelive_steam_mapping_round_363.md, docs/plans/steamworks-parity-plan.md Parity estimate: before 55% -> after 94% for focused UGC query result/preview/release harness coverage, before 96% -> after 98% for combined UGC query wrapper and call-result projection, and broader Steamworks parity remains approximately 99% because live backend timing, workshop availability, and the raw filter semantic remain intentionally bounded.

Completed work:

1. Rechecked the retained SteamUGC query/readback HLIL slots at 0x00460DF3, 0x00460E04, 0x0045FD88, and 0x0045FDAA.
2. Extended the Steamworks harness with mocked GetQueryUGCResult, GetQueryUGCPreviewURL, and ReleaseQueryUGCRequest vtable slots.
3. Added mock state for published-file ID, title, description, preview URL, result/preview success flags, query handles, and row indexes.
4. Exported enabled and disabled harness wrappers for the retained UGC query readback/release surface.
5. Added ctypes coverage for success, release, failure zeroing, and disabled offline fallback behavior.

Task A268: Promote botlib CRC and libvar runtime owners [COMPLETED]

Priority: High Primary areas: src/code/botlib/l_crc.c, src/code/botlib/l_libvar.c, src/code/botlib/be_aas_route.c, references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_crc_libvar_alias_closure_parity.py, docs/reverse-engineering/botlib-crc-libvar-alias-closure-2026-06-06.md Parity estimate: before 76% -> after 99% for focused CRC/libvar alias ownership, before 88% -> after 98% for route-cache CRC owner mapping, and overall botlib support/runtime mapping 83% -> 84%.

Completed work:

1. Rechecked the unpromoted 0x004A84B0..0x004A8790 support-runtime block immediately after GetBotLibAPI.
2. Promoted CRC_ProcessString and nine libvar runtime helpers in the retail symbol-alias map.
3. Captured LibVarGet and LibVarDeAlloc as source-visible readability helpers that retail folded into emitted callers, so they remain intentionally unpromoted.
4. Tied the CRC helper to AAS route-cache area and cluster CRC write/read wiring.
5. Added a focused regression gate and mapping note; no C source body changes were justified for this slice.

Task A267: Close WebUI server-browser retained refresh parity [COMPLETED]

Priority: High Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, tests/test_netcode_parity_manifest.py, docs/reverse-engineering/steam-retail-server-browser-appid-interop-2026-06-06.md Parity estimate: before 96% -> after 100% for the focused WebUI server-browser list/refresh support lane. Broader Steam server-browser parity remains approximately 99% because live Steam backend timing and server-side Steam GameServer public-app publication still require opt-in runtime validation.

Completed work:

1. Rechecked the retained WebUI RequestServers, RefreshList, list-response, list-failure, and refresh-complete HLIL against quakelive_steam.exe.
2. Confirmed retail RefreshList only refreshes the retained ISteamMatchmakingServers query handle and does not publish a new servers.refresh.start event or rearm browser active/timeout state.
3. Removed the SRP-only active-refresh gate from native list response/failure callbacks so retained-handle refresh rows publish like retail rows.
4. Kept native timeout protection for SRP's frame-driven initial request path while allowing callback-driven RefreshComplete to publish the retail servers.refresh.end event after retained-handle refreshes.
5. Re-audited the SRP-to-retail publication side and confirmed the existing Steam GameServer bootstrap/state path mirrors retail product, game-dir, identity, metadata, and heartbeat publication; public-retail app identity remains a runtime Steam context requirement because retail has no source-side app-id argument in SteamGameServer_Init.
6. Added static parity gates and mapping notes for the retained refresh callback semantics.

Task A266: Extend Steam GameServer callback pump coverage [COMPLETED]

Priority: High Primary areas: tests/steamworks_harness.c, tests/test_steamworks_harness.py, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_360.md, docs/plans/steamworks-parity-plan.md Parity estimate: before 63% -> after 97% for the focused Steam GameServer callback pump coverage lane. Steam GameServer callback registration plus payload projection moves approximately 92% -> 98%, while broader Steamworks parity remains approximately 99% because live backend connection timing, auth policy, VAC behavior, and transport behavior remain intentionally bounded behind opt-in online services.

Completed work:

1. Rechecked SteamServerCallbacks_Init (0x00466DB0) against the mapping report, promoted server callback aliases, and imported SteamServerCallbacks vtables in the Ghidra symbols corpus.
2. Confirmed the platform layer already prepares retained callback objects for SteamServersConnected_t, SteamServerConnectFailure_t, SteamServersDisconnected_t, ValidateAuthTicketResponse_t, server-side P2PSessionRequest_t, GSStatsReceived_t, and GSStatsStored_t.
3. Extended the Steamworks harness with retained targets for the four server callbacks that were not previously executable-test-backed.
4. Added queue helpers and ctypes coverage proving those payloads cross QL_Steamworks_RunServerCallbacks, including result, retry, and SteamID projection.
5. Tightened static parity coverage for all seven server callback IDs, object preparations, object registrations, and unregister calls.

Task A265: Reconstruct Freeze Tag thaw respawn tail [COMPLETED]

Priority: High Primary areas: src/code/game/g_freeze.c, src/code/game/g_combat.c, src/code/game/g_local.h, tests/test_game_active_pmove_wiring_parity.py, tests/test_cgame_event_transport_parity.py, docs/reverse-engineering/freeze-tag-thaw-respawn-tail-mapping-2026-06-06.md Parity estimate: before 38% -> after 88% for the focused Freeze EV_THAW_PLAYER / PM_NORMAL / ClientSpawn respawn tail. The broader Freeze thaw-completion event lane moves approximately 86% -> 93%, while broader Freeze Tag gametype wiring moves approximately 94% -> 95%.

Completed work:

1. Rechecked FUN_1004CD40 / G_FreezeClientEndFrame and confirmed assisted thaw falls into FUN_10046d80 after helper reward, obituary, and team-sound publication.
2. Rechecked FUN_1004C1B0 / Freeze_RoundStateTransition and pinned the independent winning-team thaw sequence: EV_THAW_PLAYER, PM_NORMAL, then ClientSpawn.
3. Reconstructed the Freeze marker branch in GibEntity: the normal EV_GIB_PLAYER event remains first, then marked frozen clients emit EV_THAW_PLAYER, restore PM_NORMAL, and rerun ClientSpawn.
4. Split helper assist credit into G_FreezeAwardThawAssist so score, assist medal, and centerprint side effects survive before the thawed client state is rebuilt through ClientSpawn.
5. Added focused parity tests and mapping notes for the retail thaw respawn tail, while retaining the old direct thaw visual only as a stale-marker fallback.

Task A264: Close botlib residual address and native import wiring audit [COMPLETED]

Priority: High Primary areas: references/analysis/quakelive_symbol_aliases.json, references/reverse-engineering/ghidra/quakelive_steam/functions.csv, src/code/botlib/be_interface.c, src/code/game/g_syscalls.c, src/code/server/sv_game.c, src/code/server/ql_game_imports.inc, tests/test_botlib_residual_gap_and_wiring_parity.py, docs/reverse-engineering/botlib-residual-gap-and-wiring-audit-2026-06-06.md Parity estimate: before 70% -> after 98% for focused residual botlib address classification and before 90% -> after 98% for native botlib import bridge closure. Overall botlib plus related wiring moves approximately 82% -> 83% because this round locks down classification and bridge consistency without changing source behavior.

Completed work:

1. Recomputed the 0x00482000..0x004A8400 botlib-neighborhood function scan against the nested quakelive_steam symbol alias map.
2. Captured the exact residual set as fifteen pre-AAS_Trace libjpeg memory-manager false-lead rows plus the compiler-folded sub_486F40 weapon-jump thunk.
3. Added a regression gate proving the neighborhood has 417 Ghidra rows, 401 row-backed promoted aliases, 407 promoted aliases in the band overall, and only the documented residual addresses left.
4. Added a bridge-wide native import check proving all 134 G_QL_IMPORT_BOTLIB_* slots are registered and backed by generated/local QL_G_trap_* wrappers, with the three intentional direct-native slots called out explicitly.
5. Documented why no new source reconstruction or alias promotion is justified for this residual pass.

Task A263: Extend main Steam client callback pump coverage [COMPLETED]

Priority: High Primary areas: tests/steamworks_harness.c, tests/test_steamworks_harness.py, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_359.md, docs/plans/steamworks-parity-plan.md Parity estimate: before 62% -> after 97% for the focused main client callback pump coverage lane. Steam client callback registration plus payload projection moves approximately 91% -> 98%, while broader Steamworks parity remains approximately 99% because live Steam backend timing, stats validity, presence cadence, and P2P networking behavior remain intentionally bounded behind opt-in online services.

Completed work:

1. Rechecked SteamCallbacks_Init (0x004613A0) against the mapping report, promoted callback handler aliases, and imported SteamCallbacks callback/call-result vtables in the Ghidra symbols corpus.
2. Confirmed the platform layer already prepares the retained client callback objects for rich presence, user stats, persona state, client P2P session, game-server change, friend rich presence, and UGC query completion.
3. Extended the Steamworks harness with retained callback targets for the five normal client callbacks that were not previously executable-test-backed.
4. Added queue helpers and ctypes coverage proving those payloads cross QL_Steamworks_RunCallbacks, including friend-summary enrichment and server/password string projection.
5. Tightened static parity coverage for all client callback IDs, normal callback registrations, and the UGC call-result preparation.

Task A262: Restore public retail Steam server-browser AppID interop [COMPLETED]

Priority: High Primary areas: src/common/platform/platform_steamworks.c, src/common/platform/platform_steamworks.h, src/code/client/cl_main.c, src/code/client/cl_cgame.c, tests/steamworks_harness.c, tests/test_steamworks_harness.py, docs/reverse-engineering/steam-retail-server-browser-appid-interop-2026-06-06.md Parity estimate: before 72% -> after 96% for the focused public-retail Steam server-browser AppID interop lane. Broader WebUI server-browser parity remains approximately 99% because live Steam backend timing and server-side Steam GameServer app-context publication still require opt-in runtime validation.

Completed work:

1. Traced the empty WebUI list to the native Steam browser path filtering list requests and returned rows by the running Steam app id while the legacy Quake III master fallback is policy-disabled.
2. Added explicit ForApp Steam browser and favorite helpers so the recovered current-app wrappers remain intact while the WebUI browser can target public retail Quake Live discovery.
3. Added cl_steamServerBrowserAppId, defaulting to public retail app id 282440, and stored the selected app id on native list/detail state for callback-time validation.
4. Routed WebUI favorite add/remove through the same discovery app id used by Favorites/History list requests.
5. Added executable harness coverage and source-parity assertions for current-app behavior, public-retail request app id, public-retail row validation, ping/detail projection, and favorites.

Task A261: Map botlib edge math and grapple helper owners [COMPLETED]

Priority: High Primary areas: src/code/botlib/be_aas_reach.c, src/code/botlib/be_aas_move.c, src/code/botlib/be_ai_move.c, references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_move_edge_grapple_helper_parity.py, docs/reverse-engineering/botlib-move-edge-grapple-helper-mapping-2026-06-06.md Parity estimate: before 70% -> after 97% for focused AAS edge helper mapping and before 78% -> after 97% for focused move-AI angle/grapple helper mapping. Overall botlib plus movement/reachability helper wiring moves approximately 81% -> 82% because this pass closes static helper ownership without changing live-map movement behavior.

Completed work:

1. Rechecked the remaining real botlib-neighborhood gaps after the route prelude pass and separated them from the known libjpeg false lead.
2. Promoted VectorDistance, VectorBetweenVectors, AngleDiff, GrappleState, and BotResetGrapple from HLIL/Ghidra evidence.
3. Pinned their source consumers in AAS_ClosestEdgePoints, BotTravel_Grapple, rocket/BFG jump travel, and BotMoveToGoal.
4. Preserved sub_486F40 as an explicit non-promotion because retail folds the identical rocket/BFG AAS_WeaponJumpZVelocity(origin, 120) wrappers into a single thunk.

Task A260: Reconstruct Freeze Tag assisted-thaw obituary and team sound [COMPLETED]

Priority: High Primary areas: src/code/game/g_freeze.c, src/code/cgame/cg_event.c, tests/test_game_active_pmove_wiring_parity.py, tests/test_cgame_event_transport_parity.py, docs/reverse-engineering/freeze-tag-thaw-obituary-sound-mapping-2026-06-06.md, docs/reverse-engineering/freeze-tag-mapping-2026-06-06.md, docs/reverse-engineering/qagame-mapping.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 45% -> after 91% for the focused assisted-thaw obituary/global-team-sound corridor. The broader Freeze thaw-completion event lane moves approximately 72% -> 86%, while broader Freeze Tag gametype wiring moves approximately 93% -> 94% because the deeper retail GibEntity/ClientSpawn thaw visual tail remains a separate source reconstruction target.

Completed work:

1. Rechecked FUN_1004CD40 in the committed qagame Ghidra/HLIL corpus and confirmed assisted thaw emits EV_OBITUARY with MOD_THAW, thawed client, helper client, and SVF_BROADCAST immediately after assist credit.
2. Mapped the following EV_GLOBAL_TEAM_SOUND payload to the existing G_BroadcastGlobalTeamSound bridge, using GTS_RED_RETURN for blue-team clients and GTS_BLUE_RETURN otherwise.
3. Added G_FreezeEmitThawCompletionEvents and routed non-auto thaw completion through it before the existing direct EV_THAW_PLAYER visual handoff.
4. Rechecked cgame CG_Obituary strings in the committed cgame HLIL and symbol map, then restored MOD_THAW text for world auto-thaw, local helper thaw, and remote helper thaw.
5. Added static parity coverage and a dedicated mapping note that records the reconstructed assisted-thaw layer and leaves the retail GibEntity respawn tail as the next explicit gap.

Task A259: Map AAS route prelude helpers [COMPLETED]

Priority: High Primary areas: src/code/botlib/be_aas_route.c, src/code/botlib/be_aas_route.h, src/code/botlib/be_aas_main.c, src/code/botlib/be_ai_move.c, references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_route_prelude_parity.py, docs/reverse-engineering/botlib-route-prelude-mapping-2026-06-06.md Parity estimate: before 55% -> after 96% for focused AAS route-prelude helper mapping. Route-cache invalidation and travel-flag wiring moves approximately 72% -> 97%, while overall botlib plus related AAS route/import wiring moves approximately 80% -> 81% because later route runtime and alternative-route work was already mapped in the prior tranche.

Completed work:

1. Rechecked the retail quakelive_steam.exe HLIL range 0x004925F0..0x00492F20 and separated actual AAS route helpers from the adjacent libjpeg memory-manager false lead at 0x00482150..0x004829A0.
2. Promoted eleven missing route-prelude aliases for routing debug info, cluster-local area numbering, travel-flag table initialization/access, routing-cache invalidation, area-content flag table access, reversed reachability creation, area travel time, and portal max travel-time initialization.
3. Documented which source helpers remain unpromoted because retail inlines or fuses them into larger route-cache functions.
4. Added focused parity coverage tying aliases to Ghidra rows, HLIL anchors, debug strings, source structures, AAS init order, and move-AI consumers.

Task A258: Extend Steam lobby callback pump coverage [COMPLETED]

Priority: High Primary areas: tests/steamworks_harness.c, tests/test_steamworks_harness.py, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_358.md, docs/plans/steamworks-parity-plan.md Parity estimate: before 58% -> after 96% for the focused lobby callback pump coverage lane. Steam lobby callback registration plus payload projection moves approximately 90% -> 98%, while broader Steamworks parity remains approximately 99% because live Steam backend timing and browser event ordering remain intentionally bounded behind opt-in online services.

Completed work:

1. Rechecked the retail SteamLobbyCallbacks_Init HLIL constructor at 0x004656A0, the Ghidra imported SteamLobbyCallbacks callback vtables, and the mapping report row for the eight lobby callback payload types.
2. Confirmed the platform layer already prepares and registers callback objects for LobbyCreated_t, LobbyEnter_t, LobbyChatUpdate_t, LobbyChatMsg_t, LobbyDataUpdate_t, LobbyGameCreated_t, LobbyKicked_t, and GameLobbyJoinRequested_t.
3. Extended the Steamworks harness with retained callback targets and raw queue helpers for the seven lobby callback payloads that were not previously executable-test-backed.
4. Expanded ctypes callback-pump coverage so all eight lobby payloads cross QL_Steamworks_RunCallbacks, including the chat-message body retrieval through the mocked GetLobbyChatEntry slot.
5. Tightened static parity coverage for all eight callback IDs, object preparations, and object registrations.

Task A257: Reconstruct Freeze Tag frozen powerup marker producer [COMPLETED]

Priority: High Primary areas: src/code/game/g_freeze.c, src/code/cgame/cg_players.c, src/code/cgame/cg_draw.c, tests/test_game_active_pmove_wiring_parity.py, docs/reverse-engineering/freeze-tag-frozen-marker-mapping-2026-06-06.md, docs/reverse-engineering/freeze-tag-mapping-2026-06-06.md, docs/reverse-engineering/qagame-mapping.md Parity estimate: before 35% -> after 94% for the focused qagame frozen-marker producer lane. The combined Freeze Tag qagame/cgame presentation bridge moves approximately 88% -> 94%, while broader Freeze Tag gametype wiring moves approximately 92% -> 93% because the larger FUN_1004BC80 thaw obituary/team-sound/GibEntity split remains open.

Completed work:

1. Rechecked the retail qagame death/freeze path against committed Ghidra and confirmed client + 0x17c = 0x7fffffff plus gentity + 0xc4 = 0x8000 when Freeze Tag records a frozen player.
2. Cross-checked FUN_1004BC80 HLIL and confirmed the arg2 == 1 thaw/respawn branch clears the same client marker and entitystate bit.
3. Added G_FreezeSetClientFrozenPowerupMarker so qagame now publishes ps.powerups[PW_NUM_POWERUPS] = INT_MAX and ent->s.powerups = ( 1 << PW_NUM_POWERUPS ) on freeze entry, and clears both surfaces during init/thaw.
4. Anchored the bridge against existing cgame consumers in CG_IsFrozenPlayer, CG_PlayerSprites, and the team overlay frozen bit.
5. Added a dedicated mapping note and static parity coverage for the retail offsets, source helper, and remaining thaw-event reconstruction gap.

Task A256: Map botlib chat private helper owners [COMPLETED]

Priority: Medium Primary areas: src/code/botlib/be_ai_chat.c, src/code/botlib/be_interface.c, src/code/game/botlib.h, references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_chat_private_helper_parity.py, docs/reverse-engineering/botlib-chat-private-helper-mapping-2026-06-06.md Parity estimate: before 82% -> after 96% for the focused chat private-helper corridor and before 90% -> after 91% for overall botlib chat mapping plus public chat export/import wiring. Remaining uncertainty is live chat-data behavior across retail syn.c, rnd.c, match.c, rchat.c, and per-bot initial chat files rather than private helper ownership.

Completed work:

1. Rechecked the be_ai_chat.c private-helper band against committed quakelive_steam.exe HLIL and Ghidra function rows.
2. Promoted InitConsoleMessageHeap, BotRemoveTildes, BotReplaceWeightedSynonyms, BotReplaceReplySynonyms, and StringsMatch.
3. Preserved source bodies unchanged; the current reconstruction already matches the retail static shape for heap setup, tilde scrubbing, synonym replacement, and match-piece variable capture.
4. Documented why BotChatStateFromHandle, AllocConsoleMessage, FreeConsoleMessage, and StringReplaceWords remain source-visible but unpromoted: the observed retail rows fold those operations into callers or do not expose standalone owners.
5. Added focused parity coverage for aliases, Ghidra sizes, HLIL helper call sites, source helper bodies, setup/shutdown/matching/expansion wiring, and the private/public AI export boundary.

Task A255: Normalize Steam GameServer and P2P vtable ABI wrappers [COMPLETED]

Priority: High Primary areas: src/common/platform/platform_steamworks.c, tests/steamworks_harness.c, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_356.md, docs/plans/steamworks-parity-plan.md Parity estimate: before 70% -> after 96% for the focused GameServer/P2P vtable ABI wrapper family. The combined Steam GameServer metadata, identity, packet bridge, and legacy P2P wrapper surfaces move approximately 95% -> 98%, while broader Steamworks parity remains approximately 99% because live backend validation remains intentionally bounded behind opt-in online services.

Completed work:

1. Rechecked the committed HLIL around 0x00465A40, 0x00465A60, 0x00465B00, 0x00465B70, and 0x00465D50 and confirmed the affected lanes are Steam C++ interface vtable dispatches.
2. Added QL_STEAMWORKS_FASTCALL and converted the retained ISteamNetworking, ISteamGameServer, and ISteamGameServerNetworking typedefs to the same self, unused, ... ABI shape used elsewhere in the Steam wrapper layer.
3. Updated GameServer metadata, identity, logged-on, public-IP, incoming/outgoing packet, and legacy P2P call sites to pass the explicit unused second argument before mapped retail parameters.
4. Updated Steamworks harness vtable mocks to the corrected ABI shape so argument ordering remains executable-test-backed.
5. Tightened static parity assertions for the ABI macro, representative typedefs, and mapped call sites.

Task A254: Map botlib goal item heap and avoid-goal wiring [COMPLETED]

Priority: Medium Primary areas: src/code/botlib/be_ai_goal.c, src/code/botlib/be_interface.c, src/code/game/botlib.h, src/code/game/g_local.h, src/code/game/g_public.h, src/code/game/g_syscalls.c, src/code/game/ai_dmq3.c, src/code/server/sv_game.c, src/code/server/ql_game_imports.inc, references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_goal_item_parity.py, docs/reverse-engineering/botlib-goal-item-mapping-2026-06-06.md Parity estimate: before 79% -> after 95% for the focused goal item heap/info-entity/avoid-selection corridor. Overall botlib plus goal item export/import wiring moves approximately 88% -> 89% because this pass closes static ownership and routing gaps without claiming live map-specific AAS item behavior, dynamic dropped-item behavior, or tactical item-selection quality.

Completed work:

1. Rechecked the retail be_ai_goal.c item config, fuzzy-weight index, level-item heap, info-entity, avoid-goal, dynamic entity-item update, and LTG/NBG selection band against the committed HLIL and Ghidra function rows.
2. Promoted the missing private-owner aliases ItemWeightIndex, InitLevelItemHeap, BotFreeInfoEntities, and BotAddToAvoidGoals.
3. Confirmed no C source rewrite is currently justified; small source-visible level-item helpers remain unaliased because the evidence supports them as inlined operations inside larger retail owners.
4. Added focused parity coverage for the source shape, botlib export block, qagame legacy/native syscall bridge, server VM dispatch, direct native import wrappers, and qagame level-item/camp-spot consumers.
5. Added a dedicated mapping note for the goal-item heap and avoid-goal wiring tranche.

Task A253: Reconstruct Steam GameServer P2P active-client gate [COMPLETED]

Priority: High Primary areas: src/code/server/sv_client.c, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_355.md, docs/plans/steamworks-parity-plan.md, docs/reverse-engineering/quakelive_steam_mapping_round_01.md, docs/mapping-ref/quakelive_steam_mapping_report.md Parity estimate: before 72% -> after 96% for the focused server-side P2PSessionRequest_t admission gate. The broader Steam GameServer callback bundle moves approximately 98% -> 99%, while broader Steamworks parity remains approximately 99% because live backend callback cadence and NAT behavior stay intentionally bounded behind opt-in online services.

Completed work:

1. Rechecked retail SteamServerCallbacks_OnP2PSessionRequest (0x00465B70) in the committed HLIL and confirmed the callback scans client slots for *client == 4, matching CS_ACTIVE, plus a two-word SteamID match before dispatching SteamGameServerNetworking slot 0x0c.
2. Added SV_FindActiveClientBySteamId beside the broader SteamID finder so auth-ticket responses can still locate pre-active clients while P2P session admission mirrors the retail active-slot gate.
3. Updated SV_SteamServerP2PSessionRequestCallback to remove the extra platformAuthSucceeded condition, reject missing active SteamID matches, log accepted matches, and then call the retained server P2P accept wrapper.
4. Tightened static parity coverage for the active-client finder, callback dispatch, removed auth-success gate, accepted diagnostic, and missing-player rejection path.
5. Corrected stale Steamworks mapping language that described the callback as authenticated-client gated rather than active-client gated.

Task A252: Reconstruct Freeze Tag thaw-progress eFlag buckets [COMPLETED]

Priority: High Primary areas: src/code/game/g_client.c, src/code/game/g_freeze.c, tests/test_game_active_pmove_wiring_parity.py, docs/reverse-engineering/freeze-tag-thaw-progress-eflags-mapping-2026-06-06.md, docs/reverse-engineering/freeze-tag-mapping-2026-06-06.md, docs/reverse-engineering/qagame-mapping.md, docs/gameplay/cvars.md Parity estimate: before 40% -> after 92% for the focused Freeze thaw-progress low-bit producer lane. The broader Freeze thaw/end-frame lane moves approximately 94% -> 96%, while broader Freeze Tag gametype wiring moves approximately 91% -> 92% because this pass closes a static producer-side bucket mismatch without claiming live client visual validation.

Completed work:

1. Rechecked FUN_1004CD40 / G_FreezeClientEndFrame in the committed Ghidra corpus and pinned the top-of-function comparisons against DAT_105a472c / 3 and (DAT_105a472c / 3) * 2.
2. Mapped the retail gclient + 0x1c0 bit writes to the shared EF_DEAD (0x1) and EF_TICKING (0x2) constants, preserving the retail clear-high-bucket and OR-only middle/final bucket behavior.
3. Reconstructed G_FreezeUpdateThawProgressFlags and called it at the top of G_FreezeClientEndFrame, before helper lookup and remaining-time mutation, matching the recovered retail ordering.
4. Cleared the reused low Freeze progress bits on freeze entry and thaw exit so stale progress state does not leak across reconstructed source transitions.
5. Added static parity coverage and mapping notes for the Ghidra anchors, source bucket helper, transition cleanup, and related EV_THAW_TICK producer wiring.

Task A251: Map botlib character cache and characteristic export wiring [COMPLETED]

Priority: Medium Primary areas: src/code/botlib/be_ai_char.c, src/code/botlib/be_interface.c, src/code/game/be_ai_char.h, src/code/game/botlib.h, src/code/game/g_local.h, src/code/game/g_public.h, src/code/game/g_syscalls.c, src/code/server/sv_game.c, src/code/server/ql_game_imports.inc, references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_character_cache_parity.py, docs/reverse-engineering/botlib-character-cache-mapping-2026-06-06.md Parity estimate: before 80% -> after 96% for the focused character cache/default/interpolation/accessor helper corridor and before 87% -> after 88% for overall botlib plus related character export/import wiring. Remaining uncertainty is live bot-character data behavior across maps and character files, not the static helper ownership or API routing covered here.

Completed work:

1. Rechecked quakelive_steam.exe HLIL and Ghidra rows for the be_ai_char.c helper band from BotDumpCharacter through BotShutdownCharacters.
2. Promoted missing private helper aliases for character dumping, string cleanup, unconditional cleanup, default merge, cache lookup, interpolation, and characteristic index validation.
3. Verified that no C source rewrite is currently justified; the existing source already preserves the retail cache fallback ladder, default-character merge, float-only interpolation, characteristic conversion/clamping, and parser/runtime index-check split.
4. Added tests/test_botlib_character_cache_parity.py to pin aliases, Ghidra sizes, HLIL anchors, source shape, botlib export order, legacy syscall dispatch, native qagame import wrappers, and bot AI consumer wiring.
5. Added a focused mapping note documenting observed evidence, inferred helper names, confidence, and residual live-data risk.

Task A250: Recheck demo record/playback and browser demo-list wiring [COMPLETED]

Priority: Medium Primary areas: src/code/client/cl_main.c, src/code/client/cl_cgame.c, src/code/client/cl_keys.c, src/code/cgame/cg_draw.c, src/code/cgame/cg_view.c, tests/test_demo_record_playback_mapping_parity.py, docs/reverse-engineering/demo-record-playback-mapping-2026-06-06.md Parity estimate: before 96% -> after 99% for the focused protocol-91 demo recording/playback, browser GetDemoList, console completion, and cgame demo HUD/control wiring surface. Remaining uncertainty is live .dm_91 replay freshness across third-party captures, not a known static source mismatch.

Completed work:

1. Rechecked the quakelive_steam.exe HLIL and Ghidra rows for CL_WriteDemoMessage, CL_StopRecord_f, CL_DemoFilename, CL_Record_f, CL_WalkDemoExt, CL_NextDemo, CL_DemoCompleted, CL_ReadDemoMessage, and CL_PlayDemo_f.
2. Verified that the core packet envelope and playback reader already match retail: little-endian server sequence and payload length, post-header payload writes, MAX_MSGLEN rejection, truncation handling, protocol-list fallback, and the -1/-1 terminator.
3. Reconstructed the browser-facing demo-list return shape by preserving raw .dm_91 file-list entries from CL_WebHost_BuildDemoListJson instead of stripping the protocol suffix before returning the JSON array.
4. Added static parity coverage that pins the retail aliases, HLIL anchors, protocol-91 extension path, browser dm_91 listing, retained console .dm_73 completion quirk, cgame demoPlayback draw handoff, demo HUD controls, and the new mapping note.

Task A249: Reconstruct Freeze Tag helper entity-query topology [COMPLETED]

Priority: High Primary areas: src/code/game/g_freeze.c, tests/test_game_active_pmove_wiring_parity.py, tests/test_game_round_controller_helper_parity.py, tests/test_game_helper_seam_parity.py, docs/reverse-engineering/freeze-tag-helper-query-mapping-2026-06-06.md, docs/reverse-engineering/freeze-tag-mapping-2026-06-06.md, docs/reverse-engineering/qagame-mapping.md, docs/gameplay/cvars.md Parity estimate: before 72% -> after 94% for the focused Freeze thaw-helper search topology. The broader Freeze thaw/end-frame lane moves approximately 91% -> 94%, while broader Freeze Tag gametype wiring moves approximately 90% -> 91% because this pass closes helper enumeration topology without claiming live match validation.

Completed work:

1. Rechecked FUN_1004CD40 / G_FreezeClientEndFrame in Ghidra and HLIL, including the thaw-radius mins/maxs build and qagame import call at DAT_104b13ac + 0xa4.
2. Identified the import as the recovered trap_EntitiesInBox / QL_G_trap_EntitiesInBox bridge with a 0x400 entity-list limit.
3. Reconstructed G_FreezeCountThawHelpers to enumerate the thaw-radius AABB through trap_EntitiesInBox before applying the shared thaw-helper predicate, replacing the previous source-only flat level.maxclients scan.
4. Kept the public count-returning helper API stable while selecting the first valid entity-list helper, matching retail's per-frame helper choice more closely than the previous nearest-client selection.
5. Updated Freeze mapping notes, gameplay cvar notes, and static parity tests to pin the retail entity-query topology.

Task A248: Map botlib genetic-selection wiring [COMPLETED]

Priority: Medium Primary areas: src/code/botlib/be_ai_gen.c, src/code/botlib/be_interface.c, src/code/game/be_ai_gen.h, src/code/game/botlib.h, src/code/game/ai_main.c, src/code/game/g_local.h, src/code/game/g_public.h, src/code/game/g_syscalls.c, src/code/server/sv_game.c, src/code/server/ql_game_imports.inc, references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_genetic_selection_parity.py, docs/reverse-engineering/botlib-genetic-selection-mapping-2026-06-06.md Parity estimate: before 72% -> after 96% for the focused genetic-selection helper/export corridor and before 86% -> after 87% for overall botlib plus fuzzy/genetic interbreed wiring. Remaining uncertainty is live interbreed behavior quality and map/bot-data effects rather than the static selector ownership or import/export routing covered here.

Completed work:

1. Rechecked quakelive_steam.exe HLIL and Ghidra function rows for the be_ai_gen.c helper pair around GeneticSelection and GeneticParentsAndChildSelection.
2. Promoted the missing sub_49C6C0 -> GeneticSelection alias and rechecked the existing sub_49C810 -> GeneticParentsAndChildSelection export owner.
3. Verified no C body change is currently justified; the checked-in genetic selector source already matches the retail static shape for weighted selection, fallback selection, parent exclusion, rank reversal, child selection, and failure-output zeroing.
4. Added tests/test_botlib_genetic_selection_parity.py to pin aliases, Ghidra sizes, HLIL anchors, source shape, AI export order, qagame consumer behavior, legacy syscall dispatch, server VM dispatch, and Quake Live native import wrapper wiring.

Task A247: Reconstruct Steam GameServerStats logged-on request gate [COMPLETED]

Priority: Medium Primary areas: src/common/platform/platform_steamworks.c, src/common/platform/platform_steamworks.h, tests/steamworks_harness.c, tests/test_steamworks_harness.py, tests/test_platform_services.py, docs/plans/steamworks-parity-plan.md, docs/reverse-engineering/quakelive_steam_mapping_round_354.md Parity estimate: before 40% -> after 94% for the focused GameServerStats logged-on request gate, before 97% -> after 98% for the combined idSteamStats callback/value/descriptor/request-bootstrap lane, and approximately 99% for broader Steamworks parity. Remaining uncertainty is live backend timing and callback cadence, which stay intentionally bounded behind QL_BUILD_ONLINE_SERVICES / QL_BUILD_STEAMWORKS opt-in behavior.

Completed work:

1. Rechecked the retail quakelive_steam.exe HLIL for SteamStats_OnServersConnected (0x00467190) and the repeated constructor/bootstrap gate at 0x004679d9.
2. Reconstructed QL_Steamworks_ServerIsLoggedOn through the observed SteamGameServer vtable slot 0x20, with the disabled-build stub returning qfalse.
3. Gated QL_Steamworks_ServerRequestUserStats behind the logged-on check before issuing SteamGameServerStats::RequestUserStats through slot 0x00.
4. Extended the Steamworks harness with a mocked SteamGameServer::BLoggedOn slot, result control, call counting, and enabled/disabled exports.
5. Added regression coverage proving the logged-off path suppresses the SteamGameServerStats request while the logged-on path reaches the retail request slot.

Task A246: Map botlib AAS lifecycle wiring [COMPLETED]

Priority: Medium Primary areas: src/code/botlib/be_aas_bspq3.c, src/code/botlib/be_aas_cluster.c, src/code/botlib/be_aas_entity.c, src/code/botlib/be_aas_file.c, src/code/botlib/be_aas_main.c, src/code/botlib/be_interface.c, src/code/game/botlib.h, src/code/game/g_syscalls.c, src/code/server/sv_game.c, src/code/server/ql_game_imports.inc, references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_aas_lifecycle_parity.py, docs/reverse-engineering/botlib-aas-lifecycle-mapping-2026-06-06.md Parity estimate: before 78% -> after 94% for the focused AAS BSP/file/entity lifecycle corridor and before 85% -> after 86% for overall botlib plus AAS lifecycle/import wiring. Remaining uncertainty is live-map .aas content, map-dependent clustering/reachability quality, and runtime bot behavior rather than the static lifecycle ownership or public wiring covered here.

Completed work:

1. Rechecked quakelive_steam.exe HLIL and Ghidra function rows for the AAS BSP wrapper, BSP entity parser, cluster initialization, entity query, AAS file/lump, and main setup/start-frame/load-map/shutdown bands.
2. Promoted missing aliases for AAS_Trace, AAS_EntityCollision, AAS_inPVS, AAS_BSPModelMinsMaxsOrigin, AAS_FreeBSPEntities, AAS_EntityModelindex, AAS_EntityType, AAS_EntityModelNum, AAS_OriginOfMoverWithModelNum, AAS_NextEntity, AAS_SwapAASData, AAS_LoadAASLump, and AAS_WriteAASLump.
3. Verified no C body change is currently justified; the checked-in AAS lifecycle source already matches the retail static shape for the mapped owners.
4. Added tests/test_botlib_aas_lifecycle_parity.py to pin aliases, Ghidra sizes, HLIL anchors, lifecycle source shape, BSP/AAS export order, server VM syscall dispatch, Quake Live native import slab entries, and qagame syscall wrappers.

Task A245: Reconstruct Grappling Hook impact media and consolidate hook wiring [COMPLETED]

Priority: Medium Primary areas: src/code/cgame/cg_local.h, src/code/cgame/cg_main.c, src/code/cgame/cg_weapons.c, references/analysis/quakelive_symbol_aliases.json, references/symbol-maps/cgame.json, tests/test_grappling_hook_mapping_parity.py, docs/reverse-engineering/grappling-hook-wiring-mapping-2026-06-06.md Parity estimate: before 93% -> after 98% for the focused Grappling Hook wiring and cgame-impact lane. Broader repo-wide parity remains approximately 99%; the remaining grapple uncertainty is the qagame hook temp-event weapon-field question and any untested live-map bot grapple path quality.

Completed work:

1. Rechecked quakelive_steam.exe, qagamex86.dll, and cgamex86.dll Ghidra/HLIL evidence for bot grapple reachability/travel, qagame hook projectile lifecycle, shared grapple pmove pull, cgame hook rendering, and cgame impact media.
2. Reconstructed the missing cgame grapple-impact media branch: registered the retail gfx/damage/cracked_mrk shader and sound/weapons/grapple/grhit.ogg sound, then routed CG_MissileHitWall WP_GRAPPLING_HOOK impacts through that sound, cracked mark, and the retail 16-unit mark radius.
3. Promoted qagame hook lifecycle aliases for fire_grapple, G_MissileImpact, Weapon_GrapplingHook_Fire, Weapon_HookFree, Weapon_HookThink, FireWeapon, and TeleportPlayer, while retaining the existing cgame and botlib hook aliases.
4. Added a focused mapping document and regression test covering source, symbol-map comments, alias promotion, and Ghidra/HLIL anchors across the Grappling Hook corridor.

Task A244: Reconstruct Freeze Tag thaw-helper timer and wiring [COMPLETED]

Priority: High Primary areas: src/code/game/g_freeze.c, src/code/game/g_client.c, src/code/game/g_local.h, src/code/game/g_main.c, tools/tests/match_sim/harness.py, tests/test_game_active_pmove_wiring_parity.py, tests/test_game_round_controller_helper_parity.py, tests/test_game_runframe_parity.py, docs/reverse-engineering/freeze-tag-mapping-2026-06-06.md, docs/reverse-engineering/qagame-mapping.md, docs/gameplay/cvars.md Parity estimate: before 71% -> after 91% for the focused Freeze thaw-helper/end-frame lane. Broader Freeze Tag gametype wiring moves approximately 88% -> 90% because round-state, scoreboard, last-alive, and death wiring were already mostly reconstructed. The follow-up A249 pass closes the exact helper entity-query topology; live retail DLL validation remains open.

Completed work:

1. Rechecked qagamex86.dll Ghidra and HLIL evidence for the retail Freeze helper band, especially FUN_1004CD40 / G_FreezeClientEndFrame and the small retained-helper lookup at FUN_1004CD00.
2. Replaced the source-only accumulated thaw/tick-deadline model with the retail-shaped remaining-time counter and helper-active latch.
3. Reworked per-frame thaw handling to decrement by level.msec while a valid helper remains in range, refill toward g_freezeThawTime when abandoned, retain helper credit through the recovered lookup boundary, and treat g_freezeThawTick as the non-zero EV_THAW_TICK event gate.
4. Centralized helper validity checks for same-team connected live clients, PM_NORMAL state, radius, and line of sight, then shared that predicate between counting and retained-helper lookup.
5. Refreshed the lightweight match-simulation harness, cvar notes, qagame mapping notes, and static parity tests to pin the corrected retail interpretation.

Task A243: Reconstruct CG_Missile renderer wiring [COMPLETED]

Priority: Medium Primary areas: src/code/cgame/cg_ents.c, src/code/cgame/cg_weapons.c, src/code/cgame/cg_local.h, tests/test_cgame_displaycontext_parity.py, tests/test_game_weapon_parity.py, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 86% -> after 96% for the focused CG_Missile renderer and related color/axis wiring. Overall cgame mapping coverage remains 854 / 854 committed anchors (100%); repo-wide parity remains approximately 98% because this pass tightens an already mapped renderer rather than opening new corpus coverage.

Completed work:

1. Rechecked 0x100175F0 -> CG_Missile in cgame HLIL, the Ghidra FUN_100175f0 function row, and the adjacent weapon-color helper evidence.
2. Reconstructed the retail invalid-weapon guard as CG_Error without mutating s1->weapon, matching the >= 16 retail table boundary.
3. Removed the source-only cent->lerpAngles rewrite from CG_Missile and aligned stationary proximity-mine axis setup to currentState.angles.
4. Exposed the shared weapon-color override helper and routed grenade missile entity colors through forced team/enemy weapon color before falling back to cg_weaponColor_grenade.

Task A242: Reconstruct Steam GameServer incoming UDP packet bridge [COMPLETED]

Priority: High Primary areas: src/common/platform/platform_steamworks.c, src/common/platform/platform_steamworks.h, src/code/server/sv_main.c, tests/steamworks_harness.c, tests/test_steamworks_harness.py, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_353.md, docs/plans/steamworks-parity-plan.md Parity estimate: before 35% -> after 92% for the focused incoming Steam GameServer UDP handoff lane. The combined incoming/outgoing GameServer UDP bridge moves approximately 82% -> 96%, while broader Steamworks parity remains approximately 99% because live backend validation and online service replacement remain intentionally bounded.

Completed work:

1. Rechecked SteamServer_HandleIncomingPacket (0x00465d50) in HLIL, the Ghidra FUN_00465d50 row, the Round 04 alias promotion, and the adjacent outgoing packet drain at SteamGameServer slot 0x98.
2. Added QL_Steamworks_ServerHandleIncomingPacket for slot 0x94, plus the matching disabled inline stub.
3. Wired SV_PacketEvent through SV_SteamServerHandleIncomingPacket, packing IPv4 bytes in the retail order and forwarding the original packet payload and source port into the Steam GameServer wrapper.
4. Extended the Steamworks harness and source-bound tests to cover incoming packet payload capture, endpoint capture, result propagation, and disabled fallback behavior.

Task A241: Map botlib AAS sample/query wiring [COMPLETED]

Priority: Medium Primary areas: src/code/botlib/be_aas_sample.c, src/code/botlib/be_interface.c, src/code/game/botlib.h, src/code/game/g_public.h, src/code/game/g_syscalls.c, src/code/server/sv_game.c, src/code/server/ql_game_imports.inc, references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_aas_sample_parity.py, docs/reverse-engineering/botlib-aas-sample-mapping-2026-06-06.md Parity estimate: before 79% -> after 95% for the focused AAS sample/query helper corridor and before 84% -> after 85% for overall botlib plus AAS sample/import wiring. Remaining uncertainty is live-map .aas content and map-dependent bot behavior, not the static sample/query ownership or qagame-facing import surface covered here.

Completed work:

1. Rechecked quakelive_steam.exe HLIL and Ghidra function rows for the AAS sample/query band from AAS_PresenceTypeBoundingBox through AAS_PlaneFromNum, including link heap lifecycle, point/area presence queries, client bbox tracing, trace-area collection, face/plane helpers, entity linking, bbox area enumeration, and area info copy-out.
2. Promoted missing aliases for AAS_AreaPresenceType, AAS_PointPresenceType, AAS_AreaEntityCollision, AAS_InsideFace, AAS_BoxOnPlaneSide2, AAS_UnlinkFromAreas, AAS_AASLinkEntity, AAS_LinkEntityClientBBox, and AAS_PlaneFromNum.
3. Verified no be_aas_sample.c C body change is currently justified; the checked-in source already matches the retail static shape for the mapped sample/query corridor.
4. Added tests/test_botlib_aas_sample_parity.py to pin aliases, Ghidra sizes, HLIL anchors, source helper shape, public AAS export order, server VM syscall dispatch, legacy qagame imports, Quake Live native import slab, qagame direct wrappers, and qagame syscall wrappers.

Task A240: Reconstruct Steam GameServerStats descriptor replay owner [COMPLETED]

Priority: High Primary areas: src/code/server/sv_client.c, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_352.md, docs/plans/steamworks-parity-plan.md Parity estimate: before 54% -> after 86% for the focused descriptor-table replay lane. The combined idSteamStats callback/value/descriptor lane moves approximately 98% -> 99%, while broader Steamworks parity remains approximately 99% because online services stay opt-in and exact non-int descriptor promotion plus live backend validation remain bounded.

Completed work:

1. Rechecked SteamStats_FlushPendingValues (0x004670c0), SteamStats_OnStatsReceived (0x004671d0), data_561060, data_561c00, the promoted symbol aliases, and the Ghidra rows for the retained idSteamStats descriptor owner.
2. Reconstructed the server stats descriptor shape: count at owner +0x10, pointer at +0x18, stride 0x1c, type at descriptor +0x04, name at +0x08, int value at +0x0c, float/current value at +0x10, and average-rate count/session fields at +0x14/+0x18.
3. Converted the server stats owner from a bare name table to typed descriptor records and added type-aware load/flush dispatch for int, float, and average-rate Steam GameServerStats slots.
4. Kept all 88 currently recovered qagame-facing stat names as int descriptors and made the qagame AddSteamStat import decline non-int descriptors until stronger retail evidence identifies the public float/average-rate update owner.

Task A239: Reconstruct botlib AAS reachability-generation support [COMPLETED]

Priority: Medium Primary areas: src/code/botlib/be_aas_move.c, src/code/botlib/be_aas_optimize.c, src/code/botlib/be_aas_reach.c, references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_reachability_generation_parity.py, docs/reverse-engineering/botlib-reachability-generation-mapping-2026-06-06.md Parity estimate: before 82% -> after 96% for the focused reachability-generation support corridor and before 82% -> after 84% for overall botlib plus related movement/route wiring. Remaining uncertainty is live-map .aas content and generated reachability quality, not the static runtime helper ownership covered here.

Completed work:

1. Rechecked quakelive_steam.exe HLIL and Ghidra function rows for the reachability-generation utility band around jump velocity solving, jump reach run-start probing, geometry helpers, area predicates, reachability heap allocation, store/flattening, best-link selection, weapon-jump area flags, and reachability initialization.
2. Promoted missing aliases from AAS_ClientMovementHitBBox through AAS_BestReachableLinkArea, including AAS_HorizontalVelocityForJump, AAS_JumpReachRunStart, AAS_FaceArea, AAS_AreaVolume, AAS_SetupReachabilityHeap, AAS_AllocReachability, AAS_FaceCenter, fall/jump-distance math, liquid/damage area predicates, AAS_ReachabilityExists, and AAS_StoreReachability.
3. Reconstructed the retail non-BSPC AAS_Optimize body as the observed no-op diagnostic stub while retaining the previous GPL optimizer under the BSPC map-compiler/tooling path.
4. Added tests/test_botlib_reachability_generation_parity.py to pin aliases, Ghidra sizes, HLIL anchors, source shape, the runtime optimizer stub, and retained BSPC optimizer branch.

Task A238: Reconstruct Steam GameServerStats value wrappers [COMPLETED]

Priority: High Primary areas: src/common/platform/platform_steamworks.[ch], tests/steamworks_harness.c, tests/test_steamworks_harness.py, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_351.md, docs/plans/steamworks-parity-plan.md Parity estimate: before 72% -> after 94% for the focused SteamGameServerStats int/float/average-rate value-wrapper lane. The combined idSteamStats callback plus value-wrapper reconstruction moves approximately 97% -> 98% for the focused lane, while broader Steamworks parity remains approximately 99% because live backend behavior and full descriptor-table replay remain explicit validation boundaries.

Completed work:

1. Rechecked SteamStats_FlushPendingValues (0x004670c0), SteamStats_OnStatsReceived (0x004671d0), the SteamGameServerStats import row, and promoted symbol aliases for the server stats owner.
2. Reconstructed the observed SteamGameServerStats float get slot 0x04, float set slot 0x10, and average-rate update slot 0x18, alongside the already reconstructed request, int, achievement, and store slots.
3. Extended disabled stubs so the online-services-off build keeps a stable API while returning clean fallbacks and zeroed outputs.
4. Built a full SteamGameServerStats harness mock vtable and executable coverage for request, int/float read, int/float write, average-rate update, achievement, store, and failure zeroing.

Task A237: Map botlib movement travel runtime parity [COMPLETED]

Priority: Medium Primary areas: src/code/botlib/be_ai_move.c, src/code/botlib/be_interface.c, src/code/game/botlib.h, src/code/game/g_syscalls.c, src/code/server/sv_game.c, src/code/server/ql_game_imports.inc, references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_move_travel_parity.py, docs/reverse-engineering/botlib-move-travel-mapping-2026-06-06.md Parity estimate: before 73% -> after 94% for the focused be_ai_move.c movement/travel helper corridor and before 80% -> after 82% for overall botlib plus movement/import wiring. Remaining uncertainty is live-map .aas behavior and bot tactical movement quality, not the static runtime owner mapping covered here.

Completed work:

1. Rechecked quakelive_steam.exe HLIL and Ghidra function rows for mover helpers, reachability selection, direct movement, travel methods, finish handlers, and BotMoveToGoal dispatch.
2. Promoted missing movement/travel aliases from BotOnMover through BotFuncBobStartEnd, including low-level gap/barrier/blocking helpers and jump, ladder, teleport, elevator, func_bobbing, waterjump, and walk-off-ledge owners.
3. Corrected the stale sub_4A2620 alias from BotFinishTravel_WalkOffLedge to BotTravel_Jump; the walk-off-ledge pair is now mapped as sub_4A21A0 -> BotTravel_WalkOffLedge and sub_4A24E0 -> BotFinishTravel_WalkOffLedge.
4. Verified no be_ai_move.c C body change is currently justified; the current source already matches the retail travel dispatch and helper shape for this static slice.
5. Added tests/test_botlib_move_travel_parity.py to pin aliases, Ghidra sizes, HLIL anchors, source shape, public botlib export order, server VM syscall dispatch, Quake Live native import slab, qagame direct wrappers, and qagame syscall wrappers.

Task A236: Reconstruct Steam GameServer stats callback wiring [COMPLETED]

Priority: High Primary areas: src/common/platform/platform_steamworks.[ch], src/code/server/sv_client.c, tests/steamworks_harness.c, tests/test_steamworks_harness.py, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_350.md, docs/plans/steamworks-parity-plan.md Parity estimate: before 84% -> after 97% for the focused idSteamStats GameServer stats callback/bootstrap lane. Broader Steamworks parity remains approximately 99% because online services remain opt-in and live Steam backend behavior plus full descriptor-table callback replay remain explicit validation boundaries.

Completed work:

1. Rechecked SteamGameServerUtils import evidence, idSteamStats GSStatsReceived_t / GSStatsStored_t callback vtables, and HLIL registrations at callback IDs 0x708 and 0x709.
2. Added public GS stats callback event structs, binding slots, raw payload dispatchers, retained registration/unregistration, and the optional QL_Steamworks_ServerGetAppID wrapper through server-utils slot 0x24.
3. Wired the server stats owner to retain the server app id, observe received/stored callback results, and re-request values for the retail partial-validation result 8.
4. Extended the Steamworks harness with mock SteamGameServerUtils, queueable GS stats payloads, and full enabled/disabled regression coverage for the callback pump and server app-id wrapper.

Task A235: Map botlib AAS route runtime parity [COMPLETED]

Priority: Medium Primary areas: src/code/botlib/be_aas_route.c, src/code/botlib/be_aas_routealt.c, src/code/botlib/be_interface.c, src/code/game/botlib.h, src/code/game/g_syscalls.c, src/code/server/sv_game.c, src/code/server/ql_game_imports.inc, references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_route_runtime_parity.py, docs/reverse-engineering/botlib-route-runtime-mapping-2026-06-06.md Parity estimate: before 70% -> after 96% for the focused AAS route-cache/readback helper mapping and before 82% -> after 97% for the focused route query, prediction, and alternative-route-goal source mapping. Overall botlib plus related AAS route/import wiring moves approximately 78% -> 80% because this pass promotes a cohesive route-runtime tranche without claiming exhaustive live-map .aas graph behavior.

Completed work:

1. Rechecked quakelive_steam.exe HLIL, Ghidra function rows, and existing mapping rounds for the AAS route-cache, route-query, prediction, and alternative-route helper corridor.
2. Promoted missing route-cache helper aliases for AAS_FreeOldestCache, AAS_FreeAllClusterAreaCache, AAS_InitClusterAreaCache, AAS_FreeAllPortalCache, AAS_InitRoutingUpdate, AAS_ReadCache, AAS_ReadRouteCache, and AAS_InitReachabilityAreas.
3. Verified the current source shape already matches retail for .rcd route-cache persistence, readback validation, initialization/shutdown ordering, area/portal route lookup, AAS_PredictRoute, reachability copy/iteration, model reachability, and AAS_AlternativeRouteGoals.
4. Added tests/test_botlib_route_runtime_parity.py to pin the new aliases, Ghidra sizes, Binary Ninja HLIL anchors, source body shape, public AAS export order, server import wrappers, qagame syscall wrappers, and qagame alternative-route consumers.
5. Documented the negative check that no route algorithm C body change is currently justified; the remaining uncertainty is live-map route data and broader bot behavior, not the mapped route-runtime source owner.

Task A234: Pin legacy Steam P2P networking vtables [COMPLETED]

Priority: High Primary areas: tests/steamworks_harness.c, tests/test_steamworks_harness.py, docs/reverse-engineering/quakelive_steam_mapping_round_349.md, docs/plans/steamworks-parity-plan.md Parity estimate: before 82% -> after 96% for the focused legacy ISteamNetworking client/server wrapper and harness lane. Broader Steamworks parity remains approximately 99% because the retained retail-era P2P path is now executable evidence while live Steam runtime behavior and a documented modern ISteamNetworkingSockets / ISteamNetworkingMessages adapter remain explicitly open.

Completed work:

1. Rechecked the SteamNetworking and SteamGameServerNetworking import evidence plus HLIL availability/read/send call sites for the legacy P2P wrapper slots.
2. Added a mock SteamAPI_SteamNetworking provider with send, availability, read, and accept slots matching the reconstructed wrapper offsets.
3. Replaced the null game-server networking mock with equivalent server-side P2P slots and added a one-shot SteamGameServer::GetNextOutgoingPacket harness path.
4. Added enabled/disabled ctypes wrappers and focused tests proving payload, SteamID, channel, send-type, failure, read, accept, and outgoing UDP packet behavior.

Task A233: Pin Steam favorite-server matchmaking wiring [COMPLETED]

Priority: High Primary areas: src/code/client/cl_cgame.c, tests/steamworks_harness.c, tests/test_steamworks_harness.py, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_348.md, docs/plans/steamworks-parity-plan.md Parity estimate: before 88% -> after 98% for the focused SetFavoriteServer SteamMatchmaking add/remove wrapper and client fallback lane. Broader Steam server-browser and WebUI bridge parity remains approximately 99% because live friends/history/recent-mode behavior and live Steam favorite persistence remain validation uncertainties.

Completed work:

1. Rechecked the retail qz_instance method row at 0x0055C188, the 0x00432681 dispatcher case, and the Ghidra import evidence for SteamMatchmaking plus SteamUtils.
2. Extended the Steamworks harness with mock SteamMatchmaking favorite-game add/remove slots at vtable offsets 0x08 and 0x0c, capture getters, result controls, and disabled-build stubs.
3. Added executable coverage proving app ID, IP, shared connection/query port, favorite flag, add timestamp, remove timestamp absence, and provider failure propagation.
4. Made the client local favorites cache mirror an explicit compatibility fallback when an opted-in Steam provider cannot update favorites.

Task A232: Close Attack and Defend retail parity wiring [COMPLETED]

Priority: High Primary areas: src/code/game/g_team.c, src/code/game/g_items.c, src/code/cgame/cg_main.c, tests/test_attack_defend_retail_parity.py Parity estimate: before 97% -> after 100% for the focused Attack and Defend gametype rules, flag-status transport, map item validation, score history, round controller, spawn/reset, media registration, POI, scoreboard, and HUD wiring surface. Broader gameplay/gametype parity remains approximately 99% repo-wide because this pass closes AD-specific drift while leaving unrelated gametype and validation-freshness boundaries unchanged.

Completed work:

1. Rechecked qagamex86.dll and cgamex86.dll symbol maps plus Binary Ninja HLIL and Ghidra evidence for Team_SetFlagStatus, G_CheckTeamItems, Team_TouchOurFlag, Team_TouchEnemyFlag, G_ADHandleDamageScore, AD_RoundStateTransition, G_ADUpdateScoreHistory, CG_ParseADScores, CG_DrawADRoundScoreboard, CG_DrawObjectiveStatus, CG_PlayerObjectiveSprite, and CG_RegisterGraphics.
2. Restored the retail CTF-family AD flag-status path so Team_InitGame initializes red/blue status for GT_ATTACK_DEFEND and Team_SetFlagStatus publishes the two-character CS_FLAGSTATUS payload for both CTF and Attack and Defend.
3. Restored AD map-item validation in G_CheckTeamItems, matching the retail red/blue flag warning branch for gametype 0xb.
4. Restored AD client media registration for the red/blue flag model bank and neutral/dropped flag model bank in CG_RegisterGraphics, matching the retail cgame evidence that includes gametype 0xb in those assets.
5. Added tests/test_attack_defend_retail_parity.py to pin the AD symbol-map owners, retail evidence snippets, server flag/status and item-validation gates, round controller, damage/bonus flow, spawn/frame hooks, scores_ad/adscores transport, cgame flag media, POI, and HUD wiring.

Task A231: Label invalid Steam browser request modes as internet default [COMPLETED]

Priority: High Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/reverse-engineering/quakelive_steam_mapping_round_347.md, docs/plans/steamworks-parity-plan.md Parity estimate: before 85% -> after 99% for the focused invalid/default WebUI server-browser request-mode labeling path. Broader Steam server-browser parity remains approximately 99% because native list/detail owners are already reconstructed and the remaining uncertainty is live friends/history result parity plus any distinct recent-mode UI evidence, not the observed invalid/default RequestServers dispatch.

Completed work:

1. Rechecked the retail JSBrowser_RequestServers HLIL branch arg2 - 1 u> 3, which routes mode 0 and out-of-range retained values to the internet ISteamMatchmakingServers request slot with the gamedir=baseq3 filter.
2. Confirmed the source dispatch helpers already agreed with retail: invalid/default modes enter QL_STEAM_SERVER_BROWSER_INTERNET natively and AS_GLOBAL in the source-browser fallback path.
3. Updated CL_SteamBrowser_RequestModeLabel so the retained default case now reports internet instead of unknown, aligning diagnostics and fallback telemetry with the actual dispatch path.
4. Added a platform-service parity assertion that the request-mode label block has the explicit mode-0 and default internet returns and no longer contains an unknown request-mode label.

Task A230: Lock CTF and One Flag retail parity wiring [COMPLETED]

Priority: High Primary areas: src/code/game/g_team.c, src/code/game/g_combat.c, src/code/game/g_cmds.c, src/code/game/ai_dmq3.c, src/code/game/ai_team.c, src/code/cgame/cg_servercmds.c, src/code/cgame/cg_main.c, src/code/cgame/cg_newdraw.c, tests/test_ctf_oneflag_retail_parity.py Parity estimate: before 99% -> after 100% for the focused CTF and One Flag gametype rules, scoring, status, bot, scoreboard, and HUD wiring surface. Broader gameplay/gametype parity remains approximately 99% repo-wide because this pass found the source behavior already aligned with the committed retail references and converted the audit into focused regression coverage without changing unrelated gametype lanes.

Completed work:

1. Rechecked qagamex86.dll and cgamex86.dll symbol maps plus Binary Ninja HLIL evidence for Team_SetFlagStatus, Team_TouchOurFlag, Team_TouchEnemyFlag, Pickup_Team, Team_FragBonuses, Team_CheckHurtCarrier, CheckAlmostCapture, CG_ParseCtfScores, CG_ParseCTFStats, CG_OneFlagStatus, CG_OtherTeamHasFlag, CG_YourTeamHasFlag, and CG_DrawPlayerHasFlag.
2. Verified qagame CTF and One Flag initialization/status configstrings, dropped-flag return/reset handling, neutral-flag pickup/capture flow, capture assist accounting, near-capture event selection, and the retail carrier-bonus split where Team_CheckHurtCarrier only stamps red/blue flag or Harvester skull carrier damage while Team_FragBonuses handles One Flag neutral-carrier frags.
3. Verified score transports and client parsing stay on the retail scores_ctf/ctfstats family path, including the two-character CTF flag status payload versus the single-character One Flag payload.
4. Verified client HUD/objective wiring for One Flag neutral status icons, CTF/One Flag ownerdraw visibility predicates, local carried-flag rendering, and CTF-family feeder rows.
5. Verified bot inventory, flag-state event handling, and team-order dispatch keep CTF red/blue flag state separate from One Flag neutral-flag state.
6. Added tests/test_ctf_oneflag_retail_parity.py to pin the audited retail surface across qagame, cgame, and bot wiring.

Task A229: Tighten native Steam detail query-handle fallback [COMPLETED]

Priority: High Primary areas: src/common/platform/platform_steamworks.c, tests/steamworks_harness.c, tests/test_steamworks_harness.py, tests/test_netcode_parity_manifest.py, docs/reverse-engineering/quakelive_steam_mapping_round_346.md Parity estimate: before 86% -> after 98% for the focused native server-browser detail query-handle failure path. Broader WebUI server-browser ownership remains approximately 99% because the list and detail native owners were already wired; this pass closes the partial ping/rules/players query allocation hole and preserves the documented source status-query fallback. Repo-wide parity remains 99% because the QL_BUILD_ONLINE_SERVICES default-disabled policy boundary is unchanged.

Completed work:

1. Rechecked the JSBrowserDetails_RequestServerDetails evidence for the retained PingServer -> ServerRules -> PlayerDetails dispatch order and the adjacent CancelServerQuery vtable slot.
2. Updated QL_Steamworks_RequestServerDetails so the three-detail-probe bundle is all-or-nothing: any non-positive query handle cancels the valid partial handles, leaves output handles zeroed, and returns false.
3. Verified QL_Steamworks_BeginServerBrowserDetailRequest stays inactive on partial native detail allocation failure, allowing the client's existing UDP status-query fallback to run.
4. Added a Steamworks harness detail-query result setter plus enabled-build regression coverage for the low-level wrapper and retained detail owner.
5. Pinned the new lifecycle invariant in the netcode parity manifest and documented the source-side inference boundary in round 346.

Task A228: Close Race gametype retail parity [COMPLETED]

Priority: High Primary areas: src/code/game/g_race.c, tests/test_racepoint_commands.py, tests/test_game_nonteam_scoreboard_helper_parity.py, tests/test_game_helper_seam_parity.py Parity estimate: before 98% -> after 100% for the focused Race gametype server command, checkpoint, timing, score transport, admin-point, and scoreboard wiring surface. Broader gameplay/gametype parity remains approximately 99% repo-wide because this pass closes the remaining Race-specific qagame scoreboard/start-predicate drift without changing the known compatibility-only lanes.

Completed work:

1. Rechecked qagame and cgame retail evidence for race_info, race_init, admin_race_point_N, Race touch events, finish ranking, and scores_race row transport.
2. Aligned G_RaceBuildScoreString with retail qagame ownership by consuming level.numPlayingClients and level.sortedClients instead of rebuilding a local Race order with a private qsort comparator.
3. Tightened the Race start predicate to the retail targetname-free authored map behavior while preserving the command-spawned arp* admin chain for immediate testing/editing.
4. Extended the compiled Race command probe and static parity guards to pin the retail score row shape: client number, personal best/PERS_SCORE, clamped ping, and session seconds in the sorted-client order.

Task A227: Close Red Rover retail gametype parity [COMPLETED]

Priority: High Primary areas: src/code/game/g_client.c, src/code/game/g_cmds.c, src/code/game/g_combat.c, src/code/game/g_local.h, src/code/game/g_rr.c, tests/test_game_round_controller_helper_parity.py, tests/test_game_active_pmove_wiring_parity.py, tests/test_game_helper_seam_parity.py, docs/reverse-engineering/qagame-setteam-reconstruction-2026-05-25.md Parity estimate: before 90% -> after 100% for the focused Red Rover gametype server surface: damage-score/active-round gating, infected spawn/loadout ownership, and SetTeam warmup/pre-release wiring. Broader Red Rover gametype wiring moves approximately 95% -> 100% because the round controller, infection seeding/spread, survivor bonus, autojoin, scoreboard, cgame HUD, and last-alive paths were already source-owned while the retail boolean damage helper dispatch and direct Red Rover SetTeam spawn branch still needed to be closed. Repo-wide parity remains 99%.

Completed work:

1. Rechecked qagamex86.dll symbols, Ghidra function inventory, Binary Ninja HLIL 0x100643E0, and the G_Damage gametype-12 call site.
2. Converted G_RRHandleDamageScore to the retail helper shape: same-team suppression, active-round gating, health/armor damage caps, and 100-damage score buckets for the non-infected lane.
3. Preserved the already recovered infected-survivor score policy inside the active-round helper after retail damage capping.
4. Wired G_Damage so Red Rover calls the helper at the retail post-armor split point and aborts inactive-round damage when the helper returns false.
5. Split the Red Rover spawn/loadout finalizer so infected clients consume the zombie gauntlet/health path before the generic loadout finalizer, matching the retail survivor-vs-infected branch.
6. Restored the Red Rover seed/spread forced-conversion shortcut so RR team mutations rerun G_RRResetClientForRound directly, matching the retail internal team-change latch without broadening external SetTeam.
7. Retired the stale non-owning g_rr.c stub so the source tree no longer carries a second, divergent Red Rover implementation.
8. Extended the Red Rover parity tests to pin the source helper, public prototypes, combat/SetTeam dispatch, loadout split, build-owner boundary, and HLIL offsets used for reconstruction.

Task A226: Close Domination gametype retail parity [COMPLETED]

Priority: High Primary areas: src/code/game/g_team.c, src/code/game/g_local.h, src/code/game/g_main.c, tests/test_game_domination_parity.py Parity estimate: before 93% -> after 100% for the focused Domination gametype server and client-wiring surface. Broader gameplay/gametype parity remains approximately 99% repo-wide because this pass closes the known Domination-specific qagame reward, point activation, and defense routing gaps without changing unrelated gametype lanes.

Completed work:

1. Rechecked qagamex86.dll Binary Ninja HLIL and symbol-map evidence for Team_SelectDominationSpawnPoint, G_DominationRewardCaptureParticipants, G_DominationSelectPrimaryCapturer, G_DominationPointActivate, G_UpdateDominationPointCountConfigstrings, G_DominationCheckDefenseBonus, SP_team_dom_point, and the Team_FragBonuses Domination branch.
2. Restored the retail qagame five-point registration cap and delayed team_dom_point floor-settling activation trace while preserving delayed spawn-target retry wiring.
3. Added current-frame Domination participant lists, primary-capturer retention, and retail capture/assist reward fan-out with the 25/15 score bonuses, award sprites, persistent counters, and rank medal events.
4. Routed Domination frags through the retail owned-point defense check before CTF flag logic and accepted the retail DEFENSE medal token in the rank resolver.
5. Added focused Domination parity tests and validated them with the existing Domination count and gametype lifecycle fixtures plus a Debug Win32 native solution build.

Task A225: Tighten native Steam browser request-handle fallback [COMPLETED]

Priority: High Primary areas: src/common/platform/platform_steamworks.c, src/code/client/cl_main.c, tests/steamworks_harness.c, tests/test_steamworks_harness.py, tests/test_platform_services.py, tests/test_netcode_parity_manifest.py, docs/reverse-engineering/quakelive_steam_mapping_round_345.md Parity estimate: before 88% -> after 98% for the focused native server-browser request-owner failure path. Broader WebUI server-browser ownership remains approximately 99% because the list and detail native owners were already wired; this pass closes the null-handle fallback hole and relabels friends/history compatibility telemetry as fallback-only. Repo-wide parity remains 99% because the QL_BUILD_ONLINE_SERVICES default-disabled policy boundary is unchanged.

Completed work:

1. Rechecked quakelive_steam.exe HLIL and alias evidence for JSBrowser_RequestServers, SteamBrowser_RefreshList, JSBrowser_OnRefreshComplete, and the retained ISteamMatchmakingServers request-mode slots.
2. Confirmed the current client list path is native-first for internet, LAN, friends, favorites, and history modes when an opted-in Steamworks provider exposes the native server-browser interface.
3. Updated QL_Steamworks_BeginServerBrowserOwnerRequest so a zero native request handle returns false, leaves the owner idle, and lets the existing source-browser fallback run instead of timing out an active-but-handleless native refresh.
4. Added a Steamworks harness request-result setter and lifecycle coverage for the zero-handle path.
5. Relabeled the browser compatibility diagnostics so friends/history now read as source fallback mappings after native request-handle failure, not as permanently source-owned modes.

Task A224: Wire Clan Arena retail damage-score helper [COMPLETED]

Priority: High Primary areas: src/code/game/g_team.c, src/code/game/g_combat.c, src/code/game/g_local.h, tests/test_game_round_controller_helper_parity.py Parity estimate: before 92% -> after 98% for the focused Clan Arena damage-score and active-round damage gate lane. Broader Clan Arena gametype wiring moves approximately 96% -> 98% because the round controller, spectator reset/release, compact scoreboard, castats, and last-alive event surfaces were already reconstructed while the dedicated retail CA damage helper was still absent. Repo-wide parity remains 99%.

Completed work:

1. Rechecked qagamex86.dll Ghidra metadata/function inventory and Binary Ninja HLIL for 0x10037F80, 0x10037FD0, 0x10038230, and the sibling Attack & Defend damage helper.
2. Added G_CAHandleDamageScore to mirror the retail CA post-armor damage helper: same-team suppression, active-round gating, health/armor damage caps, and 100-damage score bucket conversion through direct PERS_SCORE increments plus CalculateRanks.
3. Wired G_Damage so Clan Arena calls the CA helper at the retail post-armor split point and aborts inactive-round damage when the helper returns false.
4. Extended the round-controller parity test to pin the source helper, G_Damage dispatch, and the HLIL evidence offsets used for reconstruction.

Task A223: Name the SteamDataSource non-avatar fallback owner [COMPLETED]

Priority: High Primary areas: src/code/client/cl_steam_resources.c, src/code/client/cl_main.c, tests/test_platform_services.py, tests/test_awesomium_browser_parity.py, tools/ci/verify-awesomium-browser-host-parity.ps1, docs/reverse-engineering/quakelive_steam_mapping_round_344.md Parity estimate: before 82% -> after 96% for the focused SteamDataSource fallback-owner diagnostic surface. Broader SteamDataSource resource-bridge source visibility moves approximately 93% -> 95% because the fallback owner is now named and cvar-published, while exact live Awesomium delayed-response ownership remains bounded. Repo-wide parity remains 99% because the QL_BUILD_ONLINE_SERVICES default-disabled policy boundary is unchanged.

Completed work:

1. Rechecked quakelive_steam.exe symbols and HLIL for SteamDataSource, QLResourceInterceptor, the Awesomium::DataSource::SendResponse boundary, and prior mapping rounds 91 and 289.
2. Added a source-visible QLResourceInterceptor launcher/web fallback owner label for non-avatar steam:// requests outside the avatar-native SteamDataSource subset.
3. Published SteamDataSource subset, native-gap, and fallback-owner labels through the retained resource-bridge ROM cvar surface.
4. Updated non-avatar Steam resource diagnostics to say they are routed to the launcher/web fallback owner rather than leaving the route as a generic missing-owner message.
5. Extended platform-service, Awesomium parity, and verifier-script anchors to pin the label and cvars.

Task A222: Wire WebUI server browser native Steam detail owner [COMPLETED]

Priority: High Primary areas: src/code/client/cl_main.c, src/common/platform/platform_steamworks.[ch], tests/test_steamworks_harness.py, tests/steamworks_harness.c, tests/test_platform_services.py, tests/test_netcode_parity_manifest.py, docs/reverse-engineering/quakelive_steam_mapping_round_343.md Parity estimate: before 74% -> after 98% for the focused WebUI server-detail request dispatch lane. Broader server-browser/details parity moves approximately 98% -> 99% because native detail callbacks are now wired while provider-disabled and failed native requests still retain explicit source-owned fallbacks. Repo-wide parity remains 99% because the QL_BUILD_ONLINE_SERVICES default-disabled policy boundary is unchanged.

Completed work:

1. Rechecked quakelive_steam.exe HLIL, alias, and prior mapping evidence for SteamBrowser_RequestServerDetails, JSBrowserDetails_RequestServerDetails, the 0x58-byte callback object, the base/base+4/base+8 callback views, and the PingServer -> ServerRules -> PlayerDetails query order.
2. Added QL_Steamworks_ReadServerBrowserPingResponse so raw Steam gameserveritem_t ping payloads project into the retained servers.details.<ip>_<port>.response WebUI shape.
3. Added a client-owned native detail object with ping, rules, and players callbacks that publish the retained detail response/failed/end event families.
4. Made CL_Steam_RequestServerDetails native-first for opted-in Steamworks builds, keeping the existing UDP status query as an explicit fallback.
5. Extended harness, platform-service, and netcode parity tests to pin the new native detail owner, projection helper, and source fallback boundary.

Task A221: Allow retail LAN clients through server auth fallback [COMPLETED]

Priority: High Primary areas: src/code/server/sv_game.c, src/code/server/sv_client.c, tests/test_platform_services.py Parity estimate: before 92% -> after 100% for the focused retail-client LAN dedicated-server auth slice. Repo-wide parity remains 99% because this closes a narrow policy mismatch without changing the broader default-disabled online-service boundary.

Completed work:

1. Aligned the qagame-facing trap_VerifySteamAuth fallback with the existing Sys_IsLANAddress challenge fast path so retail clients on a private LAN no longer fail with Failed to verify Steam auth token.
2. Kept LAN joins out of the live Steam GameServer auth-session bootstrap even if userinfo happens to carry an auth token.
3. Updated platform-service coverage so the LAN fallback and auth-session skip remain pinned.

Task A220: Wire WebUI server browser native Steam list owner [COMPLETED]

Priority: High Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, tests/test_netcode_parity_manifest.py, docs/reverse-engineering/webui-server-browser-native-steamworks-2026-06-05.md Parity estimate: before 82% -> after 100% for the focused WebUI server-list dispatch slice. Broader server-browser/details parity moves approximately 96% -> 98% because native list callbacks are now wired while detail requests still retain the existing UDP status fallback. Repo-wide parity remains 99% because the QL_BUILD_ONLINE_SERVICES default-disabled policy boundary is unchanged.

Completed work:

1. Rechecked the quakelive_steam.exe HLIL and alias evidence for JSBrowser_RequestServers, SteamBrowser_RequestServers, JSBrowser_OnServerResponded, JSBrowser_OnServerFailedToRespond, JSBrowser_OnRefreshComplete, and the SteamMatchmakingServers import.
2. Confirmed the empty WebUI Internet list was caused by routing through the policy-disabled legacy UDP master query rather than by missing Steam auth.
3. Added a native ISteamMatchmakingServerListResponse owner in cl_main.c so opted-in Steamworks builds publish servers.details.*.response, servers.details.*.failed, and servers.refresh.end events into the WebUI.
4. Kept default builds and runtime-disabled profiles on the explicit fallback path; Steam and other Quake Live online services remain opt-in behind QL_BUILD_ONLINE_SERVICES.
5. Updated focused platform and netcode parity guards so the browser request surface is pinned as native-first with source-owned fallback.

Task A219: Map botlib precompiler source-handle API parity [COMPLETED]

Priority: Medium Primary areas: src/code/botlib/l_precomp.c, src/code/botlib/be_interface.c, src/code/game/botlib.h, src/code/game/g_public.h, src/code/game/g_syscalls.c, src/code/server/sv_game.c, src/code/server/ql_game_imports.inc, tests/test_botlib_internal_parity.py, docs/reverse-engineering/botlib-precompiler-source-handle-mapping-2026-06-05.md Parity estimate: before 76% -> after 96% for the focused precompiler source-handle API slice. Overall botlib plus related parser/API wiring moves approximately 68% -> 69% because this pass closes a compact ABI boundary without changing broader macro expansion, script lexing, AAS, or gameplay heuristics.

Completed work:

1. Selected l_precomp.c public source-handle wrappers as the next botlib slice after the structure/character/weight parser corridor.
2. Rechecked retail quakelive_steam.exe HLIL and Ghidra functions.csv evidence for PC_LoadSourceHandle, PC_FreeSourceHandle, PC_ReadTokenHandle, PC_SourceFileAndLine, PC_SetBaseFolder, and PC_CheckOpenSourceHandles.
3. Pinned source and retail behavior for MAX_SOURCEFILES == 64, valid handle range 1..63, token field copying, TT_STRING quote stripping, source file/line reporting, base-folder forwarding, and open-source leak warnings.
4. Confirmed related wiring through botlib_export_t, GetBotLibAPI, legacy SV_GameSystemCallsImpl dispatch, qagame compatibility import table wrappers, and classic trap_PC_* syscall stubs.
5. Documented that BOTLIB_PC_ADD_GLOBAL_DEFINE has a named direct native qagame import slot while the four PC source-handle calls intentionally remain behind the retained compatibility import surface.

Task A218: Extend botlib parser-resource sibling mapping [COMPLETED]

Priority: Medium Primary areas: src/code/botlib/be_ai_char.c, src/code/botlib/be_ai_weight.c, tests/test_botlib_internal_parity.py, docs/reverse-engineering/botlib-structure-resource-reader-mapping-2026-06-05.md Parity estimate: before 97% -> after 98% for the focused botlib parser-resource corridor selected around l_struct.c and its adjacent resource loaders. Overall botlib plus related parser wiring moves approximately 67% -> 68% because this pass adds characteristic and fuzzy-weight parser evidence without changing broader AAS, movement, or combat heuristics.

Completed work:

1. Extended the selected botlib parser section from structure-table resources into the closest source siblings: be_ai_char.c character resources and be_ai_weight.c fuzzy-weight resources.
2. Rechecked retail quakelive_steam.exe HLIL and Ghidra evidence for BotLoadCharacterFromFile, BotLoadCachedCharacter, BotLoadCharacterSkill, ReadValue, ReadFuzzyWeight, ReadFuzzySeperators_r, ReadWeightConfig, and FindFuzzyWeight.
3. Pinned retail allocation/layout evidence: character load allocation 0x2cc, weight config allocation 0x444, MAX_CHARACTERISTICS == 80, MAX_WEIGHT_FILES == 128, and the weight filename copy at offset 0x404.
4. Preserved parser quirks as retail parity facts, including the characteristic upper-bound check shape and fuzzy-switch missing-default fallback.
5. Documented that be_ai_weight.c writer helpers remain inside #if 0, reinforcing that active retail evidence belongs to the read-side parser surface for this round.

Task A217: Map botlib structure resource reader parity [COMPLETED]

Priority: Medium Primary areas: src/code/botlib/l_struct.c, src/code/botlib/be_ai_goal.c, src/code/botlib/be_ai_weap.c, references/analysis/quakelive_symbol_aliases.json, tests/test_botlib_internal_parity.py, docs/reverse-engineering/botlib-structure-resource-reader-mapping-2026-06-05.md Parity estimate: before 72% -> after 97% for the focused l_struct.c resource-reader plus item/weapon/projectile consumer-table mapping slice. Overall botlib plus related parser wiring moves approximately 65% -> 67% because this pass tightens a narrow resource-parser proof surface without changing broader bot movement, AAS, or gameplay heuristics.

Completed work:

1. Selected botlib structure resource parsing as the next productive forward section because it gates bot item, weapon, and projectile resource data.
2. Rechecked retail quakelive_steam.exe HLIL and Ghidra evidence for FindField, ReadNumber, and ReadStructure, including diagnostic strings, numeric bounds behavior, nested-structure handling, and consumer call sites.
3. Confirmed the checked-in C source already matches the observed retail shape, including the source-local ReadChar/ReadString helpers that retail folds into ReadStructure.
4. Added focused regression coverage that pins the structure-reader source contract, bot item/weapon consumers, alias map entries, HLIL anchors, and Ghidra call-site evidence.
5. Extended the pass to the concrete iteminfo, weaponinfo, and projectileinfo field tables, retail allocation sizes, table data addresses, top-level resource tokens, and weapon/projectile fix-up wiring. Documented the adjacent writer helpers as source-real but not yet safely retail-aliased because the visible WriteStructure caller is debug-gated.

Task A216: Verify server snapshot related wiring parity [COMPLETED]

Priority: High Primary areas: src/code/server/sv_snapshot.c, src/code/server/sv_bot.c, src/code/server/sv_game.c, src/code/game/g_syscalls.c, src/code/game/ai_main.c, src/code/client/cl_parse.c, src/code/client/cl_cgame.c, src/code/cgame/cg_snapshot.c, tests/test_netcode_parity_manifest.py Parity estimate: before 100% -> after 100% for the focused server snapshot related-wiring slice. Repo-wide parity remains 99% because this pass confirmed adjacent retail wiring and added a missing guard without changing the broader compatibility-only or validation-freshness boundaries.

Completed work:

1. Rechecked the server snapshot construction/send path, client parse/readback path, cgame snapshot pump, qagame visibility exports, and bot snapshot entity bridge against the committed HLIL aliases and existing parity tests.
2. Confirmed no additional source behavior drift after the SV_SendClientSnapshot wrapper closure: snapshot body writing still flows directly to overflow handling/send, client/cgame readback remains ordered, and bot clients query the built snapshot ring without packet emission.
3. Added a focused parity guard that pins SV_BotGetSnapshotEntity and QL_G_trap_BotGetSnapshotEntity to retail sub_4DDAC0/sub_4E17E0, including the legacy syscall, native import slot, qagame syscall mapping, and BotAI_GetSnapshotEntity handoff.

Task A215: Verify client connection adjacent wiring parity [COMPLETED]

Priority: High Primary areas: src/code/client/cl_main.c, src/code/client/cl_net_chan.c, src/code/server/sv_client.c, src/code/server/sv_main.c, src/code/cgame/cg_newdraw.c, tests/test_engine_netcode_parity.py Parity estimate: before 96% -> after 100% for the adjacent client connection wiring slice. Repo-wide parity remains 99% because this pass confirmed the focused wiring and corrected a stale guard rather than changing the broader online-service or validation-freshness boundaries.

Completed work:

1. Rechecked the related connection corridor around client connect/retry, challenge and connect responses, packet dispatch, client netchan setup, and server challenge/direct-connect acceptance.
2. Confirmed the cgame serverinfo handoff still pulls the retail map-name ownerdraw label from CS_SERVERINFO through SERVERINFO_KEY_MAPNAME.
3. Corrected the broad engine-netcode parity guard so it matches the stricter retail-backed cgame ownerdraw tests for the CG_MapNameText helper.
4. Reran the focused and adjacent parity tests covering netcode manifest, engine netcode, client full-parity gate, client command/netchan wrappers, command registration/run-loop mapping, platform challenge/direct-connect wiring, and cgame map-name display wiring.

Task A214: Close client connection-response protocol parity [COMPLETED]

Priority: High Primary areas: src/code/client/client.h, src/code/client/cl_main.c, tests/test_netcode_parity_manifest.py, docs/reverse-engineering/netcode-parity-manifest.json Parity estimate: before 94% -> after 100% for the focused client connection-response protocol slice. Repo-wide parity remains 99% because the online-service policy and broader external validation freshness boundaries are unchanged.

Completed work:

1. Rechecked CL_Connect_f, CL_CheckForResend, and CL_ConnectionlessPacket against the committed quakelive_steam.exe HLIL and Ghidra corpus.
2. Restored the retail bad-address path by publishing CL_WebView_PublishGameError( "Bad server address" ) after the failed address parse, matching retail sub_4F3570.
3. Added the missing 0x400-byte challengeResponse sideband text slot before clc.netchan and copied Cmd_Argv(2) into it before entering CA_CHALLENGING, matching the retail 0x016177fc -> 0x01617bfc layout/order.
4. Tightened the netcode parity guard so the retail sideband copy, client connection layout, and connectionless state transition order remain pinned.

Task A213: Close server snapshot retail send parity [COMPLETED]

Priority: High Primary areas: src/code/server/sv_snapshot.c, src/code/server/sv_client.c, tests/test_netcode_parity_manifest.py, docs/reverse-engineering/netcode-parity-manifest.json Parity estimate: before 95% -> after 100% for the focused SV_SendClientSnapshot retail send-wrapper slice. Repo-wide parity remains 99% because broader compatibility-only and online-service boundaries are unchanged.

Completed work:

1. Rechecked the quakelive_steam.exe HLIL for 0x004E5AC0 and the Ghidra functions.csv row for SV_SendClientSnapshot.
2. Removed the classic SV_WriteDownloadToClient injection from SV_SendClientSnapshot, matching the retail order from SV_WriteSnapshotToClient directly to overflow handling and SV_SendMessageToClient.
3. Kept legacy UDP download requests from arming stale downloadName state now that QL snapshots no longer carry svc_download blocks.
4. Updated the netcode manifest and parity guard so reintroducing snapshot-side classic download payloads is treated as a regression.

Task A212: Recheck client-side prediction retail parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_predict.c, src/code/cgame/cg_view.c, src/code/cgame/cg_ents.c, src/code/cgame/cg_snapshot.c, tests/test_cgame_snapshot_parity.py, references/symbol-maps/cgame.json, docs/reverse-engineering/cgame-mapping.md, docs/reverse-engineering/cgame-client-prediction-parity-recheck-2026-06-05.md Parity estimate: before 100% -> after 100% for the focused cgame client-side prediction and related wiring slice. Repo-wide parity remains 99% because this pass strengthened validation for an already-closed strict-retail corridor rather than closing a new whole-repo gap.

Completed work:

1. Rechecked the cgamex86.dll Ghidra metadata, symbol map, and cgame mapping notes for the retail prediction corridor.
2. Confirmed no gameplay source patch was needed: cg_predict.c already keeps the retail replay shape for interpolation fallbacks, local pmove_t setup, command replay, item/trigger prediction, mover adjustment, step smoothing, projectile nudge, predicted rail replay, and transition handoff.
3. Tightened tests/test_cgame_snapshot_parity.py to pin the mapped solid-list, box/capsule trace, point-contents, interpolation, red/blue-flag predicate, high-value item skip, and complete prediction symbol-map evidence set.
4. Rechecked and pinned related wiring around prediction: CG_DrawActiveFrame caller order, CG_AddPacketEntities predicted-player proxy, non-predicted local-player refresh, mover compensation, initial/next/ transition snapshot handoffs, snapshot pump, native draw wrapper, snapshot import, usercmd import, pmove settings, and the mapped retail symbols for those edges.

Task A211: Close server browser protocol and master heartbeat parity [COMPLETED]

Priority: High Primary areas: src/code/client/cl_main.c, src/code/client/cl_cgame.c, src/code/server/sv_main.c, src/code/server/sv_init.c, src/code/server/sv_client.c, src/code/qcommon/common.c, src/common/platform/platform_steamworks.c, tests/test_netcode_parity_manifest.py, docs/reverse-engineering/network-server-browser-master-heartbeat-parity-2026-06-05.md Parity estimate: before 96% -> after 100% for the focused server browser protocol and master heartbeat slice. Repo-wide parity remains 99% because the online-service boundary and broader native Steam product-integration policy are unchanged.

Completed work:

1. Rechecked the quakelive_steam.exe HLIL and Ghidra corpus for getserversResponse, getinfo xxx, getstatus, sv_master, protocol 91, and the Steam GameServer / SteamMatchmakingServers imports.
2. Confirmed the active profile owns the Quake Live browser/status tokens and info keys, while legacy UDP master traffic remains policy-gated.
3. Fixed SV_MasterHeartbeat so explicit configured host:port master entries keep their port instead of being overwritten with PORT_MASTER.
4. Added a focused regression guard for the QL browser parser, request grammar, heartbeat token, Steam heartbeat owner, and committed retail evidence.
5. Rechecked the related qz_instance browser dispatch, browser event publication, native ISteamMatchmakingServers wrapper, server info/status key publication, Steam GameServer published state, spawn/frame/reconnect/ shutdown heartbeat lifecycle, and default-disabled legacy UDP service gate.
6. Added a wiring-level parity guard so the related browser/heartbeat source graph remains pinned at 100% for this focused source-owned and policy-gated slice.

Task A210: Close selected 335-339 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_BLUE_TEAM_TIMEHELD_HASTE, CG_BLUE_TEAM_TIMEHELD_INVIS, CG_FLAG_STATUS, CG_HEALTH_COLORIZED, and CG_MATCH_STATE ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher for raw ownerdraw cases 0x14F through 0x153.
2. Confirmed blue Haste and Invisibility time-held ownerdraws normalize through 0x10039AE0 selector cases 4 and 5, select the blue scorestats_team row, accumulate qagame hold-stat seconds through the Haste/Invisibility powerup mappings, and format through the recovered %i:%i%i minute-second helper.
3. Confirmed CG_FLAG_STATUS routes to the recovered retail objective-status graphic strip, with source parity covering the CTF-family, one-flag, Harvester, and Domination strip paths plus the preserved fallback label route for unsupported states.
4. Confirmed CG_HEALTH_COLORIZED resolves local or followed-client health/armor color, preserves the competitive HUD scorebox_follow shader fallback, and keeps both shader-tint and filled-rect draw paths.
5. Confirmed CG_MATCH_STATE paints only the compact uppercase phase banner (MATCH SUMMARY, MATCH WARMUP, or MATCH IN PROGRESS) and leaves timeout/overtime text to their dedicated ownerdraws.
6. Added focused structural coverage and refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A209: Close selected 330-334 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_BLUE_TEAM_PICKUPS_REGEN, CG_BLUE_TEAM_PICKUPS_HASTE, CG_BLUE_TEAM_PICKUPS_INVIS, CG_BLUE_TEAM_TIMEHELD_FLAG, and CG_BLUE_TEAM_TIMEHELD_REGEN ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, 0x10037F80 sparse pickup-count helper, and 0x10039AE0 time-held helper for raw ownerdraw cases 0x14A through 0x14E.
2. Confirmed blue Regeneration, Haste, and Invisibility pickup ownerdraws select the blue team-stat row, normalize through retail pickup selector cases 9, 10, and 11, and paint direct integer counts from the parsed scorestats_team payload.
3. Confirmed blue flag and Regeneration time-held ownerdraws select the blue row, accumulate qagame hold-stat seconds through the flag/Regen powerup mappings, and format through the recovered %i:%i%i minute-second helper.
4. Confirmed the selected rows keep retail TDM/CTF payload membership, route through the correct pickup versus time-held draw helpers, refresh competitive scorestats through CG_IsTeamPickupOwnerDraw, and have no value, width, or key callback routes.
5. Added focused structural coverage and refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A208: Close selected 325-329 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_BLUE_TEAM_PICKUPS_BS, CG_BLUE_TEAM_TIMEHELD_QUAD, CG_BLUE_TEAM_TIMEHELD_BS, CG_BLUE_TEAM_PICKUPS_FLAG, and CG_BLUE_TEAM_PICKUPS_MEDKIT ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, 0x10037F80 pickup-count helper, and 0x10039AE0 time-held helper for raw ownerdraw cases 0x145 through 0x149.
2. Confirmed blue Battle Suit, flag, and Medkit pickup ownerdraws select the blue team-stat row, normalize through the retail pickup selector, and paint direct integer counts from the parsed scorestats_team payload.
3. Confirmed blue Quad and Battle Suit time-held ownerdraws select the blue row, accumulate qagame hold-stat seconds, and format through the recovered %i:%i%i minute-second helper.
4. Confirmed the selected rows keep retail TDM/CTF payload membership, route through the correct pickup versus time-held draw helpers, refresh competitive scorestats through CG_IsTeamPickupOwnerDraw, and have no value, width, or key callback routes.
5. Added focused structural coverage and refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A207: Close selected 308-312 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_RED_TEAM_TIMEHELD_HASTE, CG_RED_TEAM_TIMEHELD_INVIS, CG_BLUE_FLAGSTATUS, CG_BLUE_NAME, and CG_BLUE_SCORE ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher for raw ownerdraw cases 0x134 through 0x138.
2. Confirmed red Haste and Invisibility time-held ownerdraws normalize through 0x10039AE0 selector cases 4 and 5, read the retail scorestats_team caches, and format with the minute-second helper.
3. Confirmed the blue flag-status, team-name, and score ownerdraws route to retail leaves 0x10030A80, 0x10030090, and 0x1002FF50 with matching team selection, fallback, alignment, and score value callback behavior.
4. Confirmed only the time-held pair refreshes competitive scorestats through the team-pickup ownerdraw predicate and that all selected width/key callback boundaries match retail.
5. Added focused structural coverage and refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A206: Close selected 315-319 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_BLUE_BASESTATUS, CG_BLUE_PLAYER_COUNT, CG_BLUE_CLAN_PLYRS, CG_BLUE_TIMEOUT_COUNT, and CG_BLUE_TEAM_MAP_PICKUPS ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher for raw ownerdraw cases 0x13B through 0x13F.
2. Confirmed CG_BLUE_BASESTATUS stays on the shared 0x10030A80 flag/base-status helper with blue/base semantics and the expected competitive refresh boundary.
3. Confirmed CG_BLUE_PLAYER_COUNT, CG_BLUE_CLAN_PLYRS, and CG_BLUE_TIMEOUT_COUNT preserve the retail player-count formatter, round-family players-remaining gate, positive-team-size timeout gate, blue cached team-count source, blue timeout source, and alignment paths.
4. Confirmed CG_BLUE_TEAM_MAP_PICKUPS enters the zero-offset blue aggregate pickup strip, consumes the parsed blue scorestats_team row, and refreshes competitive scorestats through CG_IsTeamPickupOwnerDraw.
5. Added focused structural coverage and refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A205: Close selected 320-324 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_BLUE_TEAM_PICKUPS_RA, CG_BLUE_TEAM_PICKUPS_YA, CG_BLUE_TEAM_PICKUPS_GA, CG_BLUE_TEAM_PICKUPS_MH, and CG_BLUE_TEAM_PICKUPS_QUAD ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and 0x10037F80 team-pickup helper for raw ownerdraw cases 0x140 through 0x144.
2. Confirmed the selected blue pickup ownerdraws normalize from CG_BLUE_TEAM_MAP_PICKUPS to offsets 1 through 5, consume the blue row of parsed scorestats_team fields, and paint RA/YA/GA/MH/Quad counts through the retail direct integer text path.
3. Confirmed qagame publishes red then blue team-stat rows, cgame parses both rows into cg.teamScoreStats.values, and the retail TDM/CTF field order backs the selected RA/YA/GA/MH/Quad cache slots.
4. Confirmed the selected rows route through CG_DrawTeamPickupOwnerDraw, refresh competitive scorestats through CG_IsTeamPickupOwnerDraw, and have no value, width, or key callback routes.
5. Added focused structural coverage and refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A204: Close selected 303-307 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_servercmds.c, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 96% -> after 100% for the selected CG_RED_TEAM_PICKUPS_REGEN, CG_RED_TEAM_PICKUPS_HASTE, CG_RED_TEAM_PICKUPS_INVIS, CG_RED_TEAM_TIMEHELD_FLAG, and CG_RED_TEAM_TIMEHELD_REGEN ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, pickup helper 0x10037F80, time-held helper 0x10039AE0, retail selector tables, and team scoreboard parsers for raw ownerdraw cases 0x12f through 0x133.
2. Confirmed red Regen/Haste/Invis pickup ownerdraws normalize through the sparse retail selector to caches data_10aa687c, data_10aa6880, and data_10aa6884, then paint through the direct integer text path.
3. Confirmed red flag and Regeneration time-held ownerdraws normalize through the retail time-held selector to data_10aa68a4 and data_10aa6898, then use the shared %i:%i%i minute-second formatter.
4. Corrected the source TDM/CTF scoreboard aggregate header order so retail scoreboard payloads populate the same sparse cache slots; the synthetic CG_TEAMSTAT_MAP_PICKUPS slot remains limited to the compact scorestats_team path.
5. Added focused structural coverage and refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A203: Close selected 282 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for CG_2ND_PLYR_TIER. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll ownerdraw target set for raw ownerdraw 0x11A and confirmed CG_2ND_PLYR_TIER has no retail draw helper route.
2. Confirmed source keeps CG_2ND_PLYR_TIER as an explicit direct no-op with the neighboring CG_2ND_PLYR_PR boundary while leaving the broad second-player resolver range intact for adjacent ids.
3. Confirmed compact scorestats may still parse progression tier for data parity, but the tier ownerdraw has no placement guard, metric dispatcher, competitive refresh, value, width, key, or direct draw route.
4. Added focused structural coverage and refreshed the cgame ownerdraw and mapping ledgers for the selected id.

Task A202: Close selected 277-281 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_2ND_PLYR_AVG_PICKUP_TIME_MH, CG_2ND_PLYR_EXCELLENT, CG_2ND_PLYR_IMPRESSIVE, CG_2ND_PLYR_HUMILIATION, and CG_2ND_PLYR_PR ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, pickup-average helper 0x10037730, award-count helper 0x10037990, and raw no-route boundary for ownerdraw cases 0x115 through 0x119.
2. Confirmed CG_2ND_PLYR_AVG_PICKUP_TIME_MH selects slot 1, consumes the compact MH pickup-average scorestats payload, paints through retail %3.2f, and preserves the "-" fallback when the MH pickup count is zero.
3. Confirmed CG_2ND_PLYR_EXCELLENT, CG_2ND_PLYR_IMPRESSIVE, and CG_2ND_PLYR_HUMILIATION select slot 1, consume parsed duel-score award counts, and paint through the retail %d formatter.
4. Confirmed CG_2ND_PLYR_PR remains an explicit no-op ownerdraw boundary: progression PR is still parsed from compact scorestats, but retail has no raw 0x119 draw helper, placement guard, competitive refresh, value, width, or key callback route.
5. Added focused structural coverage and refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A201: Close selected 237-241 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_2ND_PLYR_SHOTS_RG, CG_2ND_PLYR_SHOTS_PG, CG_2ND_PLYR_SHOTS_BFG, CG_2ND_PLYR_SHOTS_CG, and CG_2ND_PLYR_SHOTS_NG ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, 0x10036B00 weapon-shot leaf, and 0x1002E2B0 weapon resolver for raw ownerdraw cases 0xed through 0xf1.
2. Confirmed the selected second-player shot ownerdraws select slot 1, resolve RG/PG/BFG/CG/NG to weapon indexes 7, 8, 9, 0xd, and 0xb, read data_10A6038C, and paint through the retail %d formatter.
3. Confirmed source normalization, metric dispatch, qagame scorestats emission, cgame scorestats parsing, competitive placement refresh, and ownerdraw/value/width/key callback absence match the retail wiring.
4. Added focused structural coverage and refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A200: Close selected 272-276 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_2ND_PLYR_PICKUPS_GA, CG_2ND_PLYR_PICKUPS_MH, CG_2ND_PLYR_AVG_PICKUP_TIME_RA, CG_2ND_PLYR_AVG_PICKUP_TIME_YA, and CG_2ND_PLYR_AVG_PICKUP_TIME_GA ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, pickup-count helper 0x10036EB0, and pickup-average helper 0x10037730 for raw ownerdraw cases 0x110 through 0x114.
2. Confirmed CG_2ND_PLYR_PICKUPS_GA and CG_2ND_PLYR_PICKUPS_MH select slot 1, consume compact GA/MH pickup-count scorestats payloads, and paint through the retail %d formatter.
3. Confirmed CG_2ND_PLYR_AVG_PICKUP_TIME_RA, CG_2ND_PLYR_AVG_PICKUP_TIME_YA, and CG_2ND_PLYR_AVG_PICKUP_TIME_GA select slot 1, consume compact pickup average payloads, paint through retail %3.2f, and preserve the retail "-" fallback when the associated pickup count is zero.
4. Confirmed source normalization, placement metric dispatch, qagame scorestats emission, cgame scorestats parsing, competitive placement refresh, and ownerdraw/value/width/key callback absence match the retail wiring.
5. Added focused structural coverage and refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A198: Close selected 232-236 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_2ND_PLYR_SHOTS_MG, CG_2ND_PLYR_SHOTS_SG, CG_2ND_PLYR_SHOTS_GL, CG_2ND_PLYR_SHOTS_RL, and CG_2ND_PLYR_SHOTS_LG ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, 0x10036B00 weapon-shot leaf, and 0x1002E2B0 weapon resolver for raw ownerdraw cases 0xe8 through 0xec.
2. Confirmed the selected second-player shot ownerdraws select slot 1, resolve MG/SG/GL/RL/LG to weapon indexes 2, 3, 4, 5, and 6, read data_10A6038C, and paint through the retail %d formatter.
3. Confirmed source normalization, metric dispatch, qagame scorestats emission, cgame scorestats parsing, competitive placement refresh, and ownerdraw/value/width/key callback absence match the retail wiring.
4. Added focused structural coverage and refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A199: Close selected 267-271 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_2ND_PLYR_ACC_PL, CG_2ND_PLYR_ACC_HMG, CG_2ND_PLYR_PICKUPS, CG_2ND_PLYR_PICKUPS_RA, and CG_2ND_PLYR_PICKUPS_YA ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, HLIL weapon-accuracy leaf, pickup-count/summary leaf, and shared placement weapon resolver for raw ownerdraw cases 0x10B through 0x10F.
2. Confirmed CG_2ND_PLYR_ACC_PL and CG_2ND_PLYR_ACC_HMG stay on the pre-switch 0x10036D60 accuracy helper, normalize through the retail 0x1002E2B0 weapon resolver, preserve the Quake Live enum gaps for PL/HMG, and consume hit/shot-derived scorestats accuracy payloads.
3. Confirmed CG_2ND_PLYR_PICKUPS, CG_2ND_PLYR_PICKUPS_RA, and CG_2ND_PLYR_PICKUPS_YA stay on the pre-switch 0x10036EB0 pickup helper, select slot 1, and consume compact RA/YA/GA/MH pickup count and average scorestats payloads.
4. Confirmed the selected pickup summary/count ownerdraws preserve retail placement summary icon/text layout, direct %d count formatting for RA/YA, competitive refresh, and absence from direct CG_OwnerDraw, value, width, and key callback routes.
5. Added focused structural coverage for constants, retail target-group membership, HLIL resolver/cache/format/layout signals, server/client scorestats transport, weapon and pickup mapping, dispatcher routing, callback absence, and competitive refresh wiring.
6. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A197: Close selected 298-302 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_RED_TEAM_PICKUPS_BS, CG_RED_TEAM_TIMEHELD_QUAD, CG_RED_TEAM_TIMEHELD_BS, CG_RED_TEAM_PICKUPS_FLAG, and CG_RED_TEAM_PICKUPS_MEDKIT ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, 0x10037F80 pickup-count leaf, and 0x10039AE0 time-held leaf for raw ownerdraw cases 0x12a through 0x12e.
2. Confirmed CG_RED_TEAM_PICKUPS_BS, CG_RED_TEAM_PICKUPS_FLAG, and CG_RED_TEAM_PICKUPS_MEDKIT stay in the arg5 - 0x124 retail pickup range and paint the direct red team integer fields for offsets 6, 9, and 0xa.
3. Confirmed CG_RED_TEAM_TIMEHELD_QUAD and CG_RED_TEAM_TIMEHELD_BS route through 0x10039AE0, resolve offsets 0 and 1, and format seconds through the recovered %i:%i%i minute-second helper.
4. Confirmed source metadata, team-scorestats transports, pickup/time-held dispatcher split, competitive refresh, and value/width/key callback absence match retail wiring.
5. Added focused structural coverage and refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A196: Close selected 227-231 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_2ND_PLYR_HITS_BFG, CG_2ND_PLYR_HITS_CG, CG_2ND_PLYR_HITS_NG, CG_2ND_PLYR_HITS_PL, and CG_2ND_PLYR_HITS_HMG ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, 0x100369D0 weapon-hit leaf, and 0x1002E2B0 weapon resolver for raw ownerdraw cases 0xe3 through 0xe7.
2. Confirmed the selected second-player hit ownerdraws select slot 1, resolve BFG/CG/NG/PL/HMG to weapon indexes 9, 0xd, 0xb, 0xc, and 0xe, read data_10A603CC, and paint through the retail %d formatter.
3. Confirmed source normalization, metric dispatch, qagame scorestats emission, cgame scorestats parsing, competitive placement refresh, and ownerdraw/value/width/key callback absence match the retail wiring.
4. Added focused structural coverage and refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A195: Close selected 293-297 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_RED_TEAM_PICKUPS_RA, CG_RED_TEAM_PICKUPS_YA, CG_RED_TEAM_PICKUPS_GA, CG_RED_TEAM_PICKUPS_MH, and CG_RED_TEAM_PICKUPS_QUAD ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and 0x10037F80 team-pickup leaf for raw ownerdraw cases 0x125 through 0x129.
2. Confirmed all five selected ownerdraws stay in the arg5 - 0x124 retail range, resolve to offsets 1 through 5, and paint the direct red team pickup-count fields through the common integer text path.
3. Confirmed source metadata maps the selected rows to CG_TEAMSTAT_PICKUPS_RA, CG_TEAMSTAT_PICKUPS_YA, CG_TEAMSTAT_PICKUPS_GA, CG_TEAMSTAT_PICKUPS_MH, and CG_TEAMSTAT_PICKUPS_QUAD, backed by the retail team-scorestats header and scorestats_team transports.
4. Confirmed the selected rows route through CG_DrawTeamPickupOwnerDraw, refresh competitive scorestats through CG_IsTeamPickupOwnerDraw, and have no value, width, key, or time-held callback routes.
5. Added focused structural coverage and refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A194: Close selected 262-266 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_2ND_PLYR_ACC_RG, CG_2ND_PLYR_ACC_PG, CG_2ND_PLYR_ACC_BFG, CG_2ND_PLYR_ACC_CG, and CG_2ND_PLYR_ACC_NG ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and HLIL weapon-accuracy leaf for raw ownerdraw cases 0x106 through 0x10A.
2. Confirmed all five selected ownerdraws stay on the pre-switch 0x10036D60 accuracy helper, normalize through the retail 0x1002E2B0 weapon resolver, and consume hit/shot-derived scorestats accuracy payloads.
3. Confirmed the selected accuracy ownerdraws resolve to Railgun, Plasmagun, BFG, Chaingun, and Nailgun, preserving the Quake Live enum gaps for CG and NG.
4. Confirmed all five selected ownerdraws preserve the retail best-row white 0.8 alpha override, refresh the competitive score cache, draw through the placement metric dispatcher instead of direct CG_OwnerDraw cases, and have no value, width, or key callback routes.
5. Added focused structural coverage for constants, retail target-group membership, HLIL resolver/cache/format/color/paint signals, server/client scorestats transport, weapon mapping, dispatcher routing, callback absence, and competitive refresh wiring.
6. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A193: Close selected 222-226 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_2ND_PLYR_HITS_GL, CG_2ND_PLYR_HITS_RL, CG_2ND_PLYR_HITS_LG, CG_2ND_PLYR_HITS_RG, and CG_2ND_PLYR_HITS_PG ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, 0x100369D0 weapon-hit leaf, and 0x1002E2B0 weapon resolver for raw ownerdraw cases 0xde through 0xe2.
2. Confirmed the selected second-player hit ownerdraws select slot 1, resolve GL/RL/LG/RG/PG to weapon indexes 4, 5, 6, 7, and 8, read data_10A603CC, and paint through the retail %d formatter.
3. Confirmed source normalization, metric dispatch, qagame scorestats emission, cgame scorestats parsing, competitive placement refresh, and ownerdraw/value/width/key callback absence match the retail wiring.
4. Added focused structural coverage and refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A192: Close selected 257-261 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_2ND_PLYR_ACC_MG, CG_2ND_PLYR_ACC_SG, CG_2ND_PLYR_ACC_GL, CG_2ND_PLYR_ACC_RL, and CG_2ND_PLYR_ACC_LG ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and HLIL weapon-accuracy leaf for raw ownerdraw cases 0x101 through 0x105.
2. Confirmed all five selected ownerdraws stay on the pre-switch 0x10036D60 accuracy helper, normalize through the retail 0x1002E2B0 weapon resolver, and consume hit/shot-derived scorestats accuracy payloads.
3. Confirmed the selected accuracy ownerdraws resolve to Machinegun, Shotgun, Grenade Launcher, Rocket Launcher, and Lightning Gun.
4. Confirmed all five selected ownerdraws preserve the retail best-row white 0.8 alpha override, refresh the competitive score cache, draw through the placement metric dispatcher instead of direct CG_OwnerDraw cases, and have no value, width, or key callback routes.
5. Added focused structural coverage for constants, retail target-group membership, HLIL resolver/cache/format/color/paint signals, server/client scorestats transport, weapon mapping, dispatcher routing, callback absence, and competitive refresh wiring.
6. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A191: Close selected 288-292 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_RED_BASESTATUS, CG_RED_PLAYER_COUNT, CG_RED_CLAN_PLYRS, CG_RED_TIMEOUT_COUNT, and CG_RED_TEAM_MAP_PICKUPS ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher paths for raw ownerdraw cases 0x120 through 0x124.
2. Confirmed red base status shares 0x10030A80 with flag status, red player count routes to 0x100333C0, red clan-player count routes to 0x100335F0 behind the players-remaining gate, and red timeout count routes to 0x10033530 behind the positive team-size gate.
3. Confirmed red map pickups enter the 0x10037F80 aggregate pickup strip as the zero-offset red range case, preserving parsed scorestats_team fields, icon selection, visible positive-count filtering, and the retail 25/28 slot strides.
4. Added focused structural coverage and refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A190: Close selected 217-221 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_2ND_PLYR_FRAGS_NG, CG_2ND_PLYR_FRAGS_PL, CG_2ND_PLYR_FRAGS_HMG, CG_2ND_PLYR_HITS_MG, and CG_2ND_PLYR_HITS_SG ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, 0x100368A0 weapon-frag leaf, 0x100369D0 weapon-hit leaf, and 0x1002E2B0 weapon resolver for raw ownerdraw cases 0xd9 through 0xdd.
2. Confirmed CG_2ND_PLYR_FRAGS_NG, CG_2ND_PLYR_FRAGS_PL, and CG_2ND_PLYR_FRAGS_HMG select slot 1, resolve to weapon indexes 0xb, 0xc, and 0xe, read data_10A6030C, and paint through the retail %d formatter.
3. Confirmed CG_2ND_PLYR_HITS_MG and CG_2ND_PLYR_HITS_SG select slot 1, resolve to weapon indexes 2 and 3, read data_10A603CC, and paint through the retail %d formatter.
4. Confirmed source normalization, metric dispatch, qagame scorestats emission, cgame scorestats parsing, competitive placement refresh, and ownerdraw/value/width/key callback absence match the retail wiring across the frag-order and accuracy-order payload split.
5. Added focused structural coverage and refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A189: Close selected 252-256 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_2ND_PLYR_DMG_BFG, CG_2ND_PLYR_DMG_CG, CG_2ND_PLYR_DMG_NG, CG_2ND_PLYR_DMG_PL, and CG_2ND_PLYR_DMG_HMG ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and HLIL damage leaf for raw ownerdraw cases 0xfc through 0x100.
2. Confirmed all five selected ownerdraws stay on the pre-switch 0x10036C30 damage helper, normalize through the retail 0x1002E2B0 weapon resolver, and consume the parsed scorestats damage payload.
3. Confirmed the selected damage ownerdraws resolve to BFG, Chaingun, Nailgun, Proximity Launcher, and Heavy Machinegun, preserving the Quake Live enum gaps for CG/NG/PL/HMG.
4. Confirmed all five selected ownerdraws refresh the competitive score cache, draw through the placement metric dispatcher instead of direct CG_OwnerDraw cases, and have no value, width, or key callback routes.
5. Added focused structural coverage for constants, retail target-group membership, HLIL resolver/cache/format/paint signals, server/client scorestats transport, weapon mapping, dispatcher routing, callback absence, and competitive refresh wiring.
6. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A188: Close selected 212-216 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_2ND_PLYR_FRAGS_LG, CG_2ND_PLYR_FRAGS_RG, CG_2ND_PLYR_FRAGS_PG, CG_2ND_PLYR_FRAGS_BFG, and CG_2ND_PLYR_FRAGS_CG ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, 0x100368A0 weapon-frag leaf, and 0x1002E2B0 weapon resolver for raw ownerdraw cases 0xd4 through 0xd8.
2. Confirmed the selected second-player frag ownerdraws select slot 1, resolve LG/RG/PG/BFG/CG to weapon indexes 6, 7, 8, 9, and 0xd, read data_10A6030C, and paint through the retail %d formatter.
3. Confirmed source normalization, metric dispatch, qagame scorestats emission, cgame scorestats parsing, competitive placement refresh, and ownerdraw/value/width/key callback absence match the retail wiring.
4. Added focused structural coverage and refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A187: Close selected 247-251 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_2ND_PLYR_DMG_GL, CG_2ND_PLYR_DMG_RL, CG_2ND_PLYR_DMG_LG, CG_2ND_PLYR_DMG_RG, and CG_2ND_PLYR_DMG_PG ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and HLIL damage leaf for raw ownerdraw cases 0xf7 through 0xfb.
2. Confirmed all five selected ownerdraws stay on the pre-switch 0x10036C30 damage helper, normalize through the retail 0x1002E2B0 weapon resolver, and consume the parsed scorestats damage payload.
3. Confirmed the selected damage ownerdraws resolve to Grenade Launcher, Rocket Launcher, Lightning Gun, Railgun, and Plasmagun in the same order as the qagame producer and cgame parser.
4. Confirmed all five selected ownerdraws refresh the competitive score cache, draw through the placement metric dispatcher instead of direct CG_OwnerDraw cases, and have no value, width, or key callback routes.
5. Added focused structural coverage for constants, retail target-group membership, HLIL resolver/cache/format/paint signals, server/client scorestats transport, weapon mapping, dispatcher routing, callback absence, and competitive refresh wiring.
6. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A186: Close selected 283-287 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_RED_FLAGSTATUS, CG_RED_SCORE, CG_RED_NAME, CG_RED_OWNED_FLAGS, and CG_RED_AVG_PING ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and HLIL leaves for raw ownerdraw cases 0x11b through 0x11f.
2. Confirmed CG_RED_FLAGSTATUS stays on the shared team flag/base-status icon leaf, including red status selection and shader fallback behavior.
3. Confirmed CG_RED_SCORE stays on the aligned team-score leaf and remains the only selected ownerdraw with a CG_GetValue callback route.
4. Confirmed CG_RED_NAME stays on the aligned team-name leaf with the configured-name and default-name fallback path.
5. Confirmed CG_RED_OWNED_FLAGS stays behind the retail Domination / Attack-Defend gate before drawing the red owned-point count.
6. Restored the shared CG_DrawTeamAveragePing leaf to retail behavior: matching score rows are averaged without a negative-ping skip, no matching rows still paint (0), thresholds are >40 and >80, alpha is 0.8, and text formats as (%d).
7. Added focused structural coverage for constants, retail target-group membership, HLIL branch/format/color signals, local dispatcher routes, value/width/key callback membership, competitive-refresh absence, and the selected ownerdraw ledger rows.
8. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A185: Close selected 242-246 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_2ND_PLYR_SHOTS_PL, CG_2ND_PLYR_SHOTS_HMG, CG_2ND_PLYR_DMG_G, CG_2ND_PLYR_DMG_MG, and CG_2ND_PLYR_DMG_SG ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and HLIL leaves for raw ownerdraw cases 0xf2 through 0xf6.
2. Confirmed CG_2ND_PLYR_SHOTS_PL and CG_2ND_PLYR_SHOTS_HMG stay on the pre-switch 0x10036B00 shot-count helper, normalize through the retail weapon resolver, and consume the parsed scorestats shot-count payload.
3. Confirmed CG_2ND_PLYR_DMG_G, CG_2ND_PLYR_DMG_MG, and CG_2ND_PLYR_DMG_SG stay on the pre-switch 0x10036C30 damage helper, normalize to Gauntlet/Machinegun/Shotgun, and consume the parsed scorestats damage payload.
4. Confirmed all five selected ownerdraws refresh the competitive score cache, draw through the placement metric dispatcher instead of direct CG_OwnerDraw cases, and have no value, width, or key callback routes.
5. Added focused structural coverage for constants, retail target-group membership, HLIL helper-field signals, server/client scorestats transport, weapon mapping, dispatcher routing, callback absence, and competitive refresh wiring.
6. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A184: Close selected 207-211 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_2ND_PLYR_FRAGS_G, CG_2ND_PLYR_FRAGS_MG, CG_2ND_PLYR_FRAGS_SG, CG_2ND_PLYR_FRAGS_GL, and CG_2ND_PLYR_FRAGS_RL ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, 0x100368A0 weapon-frag leaf, and 0x1002E2B0 placement weapon resolver for raw ownerdraw cases 0xcf through 0xd3.
2. Confirmed the selected second-player frag head resolves to G/MG/SG/GL/RL weapon indexes 1 through 5, reads the slot-1 weapon-frag cells, and formats through the retail %d literal.
3. Confirmed the source placement resolver normalizes the selected second-slot ownerdraws back to the first-player family before weapon lookup and text construction.
4. Confirmed the qagame-to-cgame compact scorestats transport uses the same G/MG/SG/GL/RL-leading weapon-frag order before the ownerdraws read cg.scoreStats[clientNum].weaponFrags.
5. Added focused structural coverage for constants, retail target-group membership, HLIL resolver/cache/format/paint signals, source weapon mapping, server send path, cgame parse path, placement dispatch, competitive refresh wiring, and callback/direct-switch absence.
6. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A183: Close selected 202-206 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_2ND_PLYR_ACC, CG_2ND_PLYR_FLAG, CG_2ND_PLYR_AVATAR, CG_2ND_PLYR_TIMEOUT_COUNT, and CG_2ND_PLYR_HEALTH_ARMOR ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and HLIL leaves for raw ownerdraw cases 0xca through 0xce.
2. Confirmed CG_2ND_PLYR_ACC stays on the pre-switch placement accuracy helper and uses the retail %d%% total-accuracy path.
3. Confirmed CG_2ND_PLYR_FLAG stays on the pre-switch placement flag helper and draws the cached country-flag shader or CG_RegisterCountryFlag fallback without team-flag powerup logic.
4. Confirmed CG_2ND_PLYR_AVATAR directly calls the slot-1 avatar/model-icon path and preserves the avatar-handle-before-model-icon fallback.
5. Confirmed CG_2ND_PLYR_TIMEOUT_COUNT remains an explicit retail no-op, with no placement guard, competitive refresh, value, width, or key callback route.
6. Confirmed CG_2ND_PLYR_HEALTH_ARMOR directly calls the secondary health/armor comparison bars, including local-client stat fallback, damage-loss segment, left-anchored secondary bars, and no text paint.
7. Added focused structural coverage for constants, retail target-group membership, HLIL offset/format/fallback signals, local helper routes, competitive refresh wiring, callback absence, and direct no-op behavior.
8. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A182: Feed retail arena catalog into live Awesomium start-server map list [COMPLETED]

Priority: High Primary areas: src/code/client/cl_cgame.c, src/code/client/cl_awesomium_win32.cpp, tests/test_awesomium_browser_parity.py Parity estimate: before 99.1% -> after 99.2% for the focused Awesomium WebUI arena-listing lane. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail GetMapList HLIL, the map-pool loader notes, and the shipped WebUI bundle: Start Match synchronously imports GetMapList() and GetFactoryList(), then filters maps by the selected factory's numeric basegt.
2. Switched the native WebUI map builder to prefer Quake Live's retail mappool.txt rotation source, merging duplicate map|factory rows into one descriptor per unique map and deriving the 13-slot gametypes table from parsed factory basegt definitions; arena files now provide display metadata and fallback descriptors.
3. Added startup-time map/factory JSON payloads plus post-document-ready map and factory catalog sync. The map catalog still streams in bounded batches, and the factory catalog sync keeps the WebUI factory filter from staying on the one-entry fallback.
4. Hardened the injected bridge so factory basegt tokens map to retail gametype_t ids, the full bridge reinstalls after Awesomium document context resets, and Browserify/CommonJS mapdb/factories module exports are refreshed in place when the native catalog changes.
5. Added runtime diagnostics for qz/module map and factory counts, updated focused Awesomium parity coverage, and revalidated the windowed live client: the retail map pool loaded 173 rotation entries and the WebUI exposed 93 unique maps plus 22 factories.

Task A181: Close selected 197-201 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_2ND_PLYR_DEATHS, CG_2ND_PLYR_DMG, CG_2ND_PLYR_TIME, CG_2ND_PLYR_PING, and CG_2ND_PLYR_WINS ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and HLIL leaves for raw ownerdraw cases 0xc5 through 0xc9.
2. Confirmed CG_2ND_PLYR_DEATHS and CG_2ND_PLYR_DMG stay on the second-slot placement deaths/damage leaves and now use the retail %d formatter literal.
3. Confirmed CG_2ND_PLYR_TIME remains an inert ownerdraw boundary: retail exposes no raw 0xc7 switch case, placement guard, value, width, or key callback route.
4. Confirmed CG_2ND_PLYR_PING stays on the retail ping leaf, including %d text formatting, >40/>80 color thresholds, zero blue component, and 0.8 alpha.
5. Confirmed CG_2ND_PLYR_WINS stays on the slot-1 wins/losses string path fed by the retail %d/%d scorestats payload.
6. Added focused structural coverage for constants, retail target-group membership, HLIL cache/format/color signals, local dispatcher routes, competitive score refresh wiring, callback absence, and the inert time boundary.
7. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A180: Close selected 192-196 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, tests/test_cgame_displaycontext_parity.py, tests/test_cgame_ownerdraw_text_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_1ST_PLYR_TIER, CG_2ND_PLYR, CG_2ND_PLYR_READY, CG_2ND_PLYR_SCORE, and CG_2ND_PLYR_FRAGS ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and HLIL leaves for raw ownerdraw cases 0xc0 through 0xc4.
2. Confirmed CG_1ST_PLYR_TIER remains an inert boundary: compact scorestats can still carry progression tier, but retail exposes no raw 0xc0 ownerdraw, placement guard, value, width, or key callback route.
3. Confirmed CG_2ND_PLYR and CG_2ND_PLYR_SCORE use the direct slot-1 name and score helpers, including cached second-player text, "-" fallback, retail %d score formatting, and alignment.
4. Confirmed CG_2ND_PLYR_READY stays on the secondary status/ready helper, preserving LEADS/TIED/TRAILS and READY/NOT READY behavior.
5. Restored the retail %d literal for placement frag ownerdraw text so CG_2ND_PLYR_FRAGS matches the 0x100361F0 slot-1 frag path.
6. Added focused structural coverage for constants, retail target-group membership, HLIL cache/status/format signals, local dispatcher routes, competitive score refresh wiring, callback absence, and the inert tier boundary.
7. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A179: Close selected 187-191 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, src/code/cgame/cg_servercmds.c, src/code/game/g_cmds.c, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_1ST_PLYR_AVG_PICKUP_TIME_MH, CG_1ST_PLYR_EXCELLENT, CG_1ST_PLYR_IMPRESSIVE, CG_1ST_PLYR_HUMILIATION, and CG_1ST_PLYR_PR ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, 0x10037730 pickup-average leaf, and 0x10037990 award-count leaf for raw ownerdraw cases 0xbb through 0xbf.
2. Confirmed CG_1ST_PLYR_AVG_PICKUP_TIME_MH closes the first-player pickup average band, reads the MH average cache, and shares the retail "-" zero-count fallback.
3. Confirmed CG_1ST_PLYR_EXCELLENT, CG_1ST_PLYR_IMPRESSIVE, and CG_1ST_PLYR_HUMILIATION stay on the retail award-count path and are fed by the qagame duel score row plus cgame score parsing.
4. Restored the retail %d literal for placement award-count ownerdraw text.
5. Confirmed CG_1ST_PLYR_PR remains an inert ownerdraw boundary: compact scorestats still transport progression PR/tier, but retail does not route raw 0xbf through ownerdraw, guard, value, width, or key callbacks.
6. Added focused structural coverage for constants, retail target-group membership, HLIL average/award cache signals, qagame send path, cgame parse path, formatter, dispatcher guard, callback absence, and direct-switch absence.
7. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A178: Close selected 182-186 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, src/code/cgame/cg_servercmds.c, src/code/game/g_cmds.c, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 98% -> after 100% for the selected CG_1ST_PLYR_PICKUPS_GA, CG_1ST_PLYR_PICKUPS_MH, CG_1ST_PLYR_AVG_PICKUP_TIME_RA, CG_1ST_PLYR_AVG_PICKUP_TIME_YA, and CG_1ST_PLYR_AVG_PICKUP_TIME_GA ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, 0x10036EB0 pickup-count leaf, and 0x10037730 pickup-average leaf for raw ownerdraw cases 0xb6 through 0xba.
2. Confirmed CG_1ST_PLYR_PICKUPS_GA and CG_1ST_PLYR_PICKUPS_MH stay on the direct retail pickup-count path and consume the same compact GA/MH pickup order used by qagame send and cgame parse.
3. Restored the retail pickup-average zero-count fallback: RA/YA/GA average ownerdraws now paint "-" when the corresponding count is zero and %3.2f when a pickup interval is available.
4. Pinned both compact scorestats and legacy duel-score parsing for the selected pickup counts and average intervals.
5. Added focused structural coverage for constants, retail target-group membership, HLIL count/average cache signals, qagame send path, cgame parse path, visible average fallback, dispatcher guard, callback absence, and direct-switch absence.
6. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A177: Close selected 177-181 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, src/code/cgame/cg_servercmds.c, src/code/game/g_cmds.c, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 97% -> after 100% for the selected CG_1ST_PLYR_ACC_PL, CG_1ST_PLYR_ACC_HMG, CG_1ST_PLYR_PICKUPS, CG_1ST_PLYR_PICKUPS_RA, and CG_1ST_PLYR_PICKUPS_YA ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, 0x10036D60 weapon-accuracy leaf, 0x10036EB0 pickup leaf, and 0x1002E2B0 placement weapon resolver for raw ownerdraw cases 0xb1 through 0xb5.
2. Confirmed the selected accuracy tail resolves to PL/HMG weapon indexes 0xc and 0xe, reads the data_10A6034C accuracy cache, formats through the retail %d%% literal, and inherits the best-row color override.
3. Restored the retail CG_1ST_PLYR_PICKUPS summary strip: nonzero RA/YA/GA/MH pickups draw the retail pickup shader, count text, and %3.2f average interval from the compact scorestats payload.
4. Confirmed CG_1ST_PLYR_PICKUPS_RA and CG_1ST_PLYR_PICKUPS_YA stay on the direct retail pickup-count path and now use the retail %d formatter.
5. Pinned qagame-to-cgame compact pickup wiring: qagame sends RA/YA/GA/MH counts and average intervals, cgame parses them into pickupCounts and pickupAvgSeconds, and the ownerdraws consume the same order.
6. Added focused structural coverage for constants, retail target-group membership, resolver returns, pickup shader handles, server send path, cgame parse path, summary drawing, direct count drawing, dispatcher guard, callback absence, and direct-switch absence.
7. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A176: Close selected 6-10 cgame front-panel status ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, src/code/cgame/cg_main.c, tests/test_cgame_ownerdraw_text_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_GAME_TYPE_MAP, CG_GAME_STATUS, CG_MATCH_DETAILS, CG_MATCH_END_CONDITION, and CG_MATCH_STATUS ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher for raw ownerdraw cases 0x06 through 0x0A.
2. Confirmed CG_GAME_TYPE_MAP, CG_GAME_STATUS, CG_MATCH_DETAILS, CG_MATCH_END_CONDITION, and CG_MATCH_STATUS route to the recovered intro-panel detail, standalone game-status, match-details, match-end-condition, and match-status helpers respectively.
3. Confirmed the selected batch preserves the intended callback boundary: CG_GAME_STATUS and CG_MATCH_STATUS participate in the width callback, while all five stay out of value and key callbacks.
4. Added focused structural coverage for constants, retail target membership, source switch routes, helper behavior, shipped menu reachability, display-context wiring, and width callback splits for the selected five.
5. Refreshed the cgame ownerdraw and mapping ledgers for this selected batch.

Task A175: Close selected 172-176 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, src/code/cgame/cg_servercmds.c, src/code/game/g_cmds.c, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_1ST_PLYR_ACC_RG, CG_1ST_PLYR_ACC_PG, CG_1ST_PLYR_ACC_BFG, CG_1ST_PLYR_ACC_CG, and CG_1ST_PLYR_ACC_NG ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, 0x10036D60 weapon-accuracy leaf, and 0x1002E2B0 placement weapon resolver for raw ownerdraw cases 0xac through 0xb0.
2. Confirmed the selected accuracy band resolves to RG/PG/BFG/CG/NG weapon indexes 7, 8, 9, 0xd, and 0xb, reads the data_10A6034C accuracy cache, and formats through the retail %d%% literal.
3. Confirmed the selected ownerdraws inherit the restored retail comparison color path: the row whose weapon accuracy beats the opposing placement slot paints white with 0.8 alpha.
4. Pinned the qagame-to-cgame compact accuracy wiring: qagame sends accuracy_hits and accuracy_shots, cgame parses both in the same weapon order, then derives weaponAccuracy before the ownerdraw reads it.
5. Added focused structural coverage for constants, retail target-group membership, resolver returns, cache reads, weapon mapping, server send path, cgame parse path, formatter, comparison color, dispatcher guard, callback absence, and direct-switch absence.
6. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A174: Close selected 1-5 cgame front-panel ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, src/code/cgame/cg_servercmds.c, src/code/cgame/cg_main.c, tests/test_cgame_ownerdraw_text_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_SERVER_SETTINGS, CG_STARTING_WEAPONS, CG_GAME_LIMIT, CG_GAME_TYPE, and CG_GAME_TYPE_ICON ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher for raw ownerdraw cases 0x01 through 0x05.
2. Confirmed CG_SERVER_SETTINGS, CG_STARTING_WEAPONS, CG_GAME_LIMIT, CG_GAME_TYPE, and CG_GAME_TYPE_ICON route to the recovered settings, starting-weapon icon strip, game-limit label, gametype text, and gametype icon helpers respectively.
3. Confirmed the selected batch preserves the intended callback boundary: CG_GAME_TYPE alone participates in the width callback, while all five stay out of value and key callbacks.
4. Added focused structural coverage for constants, retail target membership, source switch routes, helper behavior, shipped menu reachability, display-context wiring, and configstring inputs for the selected five.
5. Refreshed the cgame ownerdraw and mapping ledgers for this selected batch.

Task A173: Close selected 167-171 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, src/code/cgame/cg_servercmds.c, src/code/game/g_cmds.c, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 98% -> after 100% for the selected CG_1ST_PLYR_ACC_MG, CG_1ST_PLYR_ACC_SG, CG_1ST_PLYR_ACC_GL, CG_1ST_PLYR_ACC_RL, and CG_1ST_PLYR_ACC_LG ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, 0x10036D60 weapon-accuracy leaf, and 0x1002E2B0 placement weapon resolver for raw ownerdraw cases 0xa7 through 0xab.
2. Confirmed the selected accuracy head resolves to MG/SG/GL/RL/LG weapon indexes 2 through 6, reads the data_10A6034C accuracy cache, and formats through the retail %d%% literal.
3. Restored the retail per-weapon accuracy comparison color: the row whose accuracy beats the opposing placement slot for the same weapon paints white with 0.8 alpha.
4. Pinned the qagame-to-cgame compact accuracy wiring: qagame sends accuracy_hits and accuracy_shots, cgame parses both in the same weapon order, then derives weaponAccuracy before the ownerdraw reads it.
5. Added focused structural coverage for constants, retail target-group membership, resolver returns, cache reads, weapon mapping, server send path, cgame parse path, formatter, comparison color, dispatcher guard, callback absence, and direct-switch absence.
6. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A172: Close selected 90-94 cgame endgame/selected-player ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, tests/test_cgame_ownerdraw_text_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_PLYR_END_GAME_SCORE, CG_PLYR_BEST_WEAPON_NAME, CG_SELECTED_PLYR_TEAM_COLOR, CG_SELECTED_PLYR_ACCURACY, and CG_FOLLOW_PLAYER_NAME ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher for raw ownerdraw cases 0x5A through 0x5E.
2. Confirmed CG_PLYR_END_GAME_SCORE, CG_PLYR_BEST_WEAPON_NAME, CG_SELECTED_PLYR_TEAM_COLOR, CG_SELECTED_PLYR_ACCURACY, and CG_FOLLOW_PLAYER_NAME route to the endgame score, selected best weapon, selected team color, selected accuracy, and follow-name helpers respectively.
3. Confirmed the selected batch preserves the intended competitive-score refresh boundary: only the endgame score row refreshes, while the selected-player and follow-name ownerdraws do not.
4. Added focused structural coverage for constants, retail target membership, source switch routes, helper behavior, shipped menu reachability, competitive refresh, and callback absence for the selected five.
5. Refreshed the cgame ownerdraw and mapping ledgers for this selected batch.

Task A171: Close selected 162-166 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, src/code/cgame/cg_servercmds.c, src/code/game/g_cmds.c, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_1ST_PLYR_DMG_BFG, CG_1ST_PLYR_DMG_CG, CG_1ST_PLYR_DMG_NG, CG_1ST_PLYR_DMG_PL, and CG_1ST_PLYR_DMG_HMG ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, 0x10036C30 weapon-damage leaf, and 0x1002E2B0 placement weapon resolver for raw ownerdraw cases 0xa2 through 0xa6.
2. Confirmed the selected damage tail resolves to BFG/CG/NG/PL/HMG weapon indexes 9, 0xd, 0xb, 0xc, and 0xe, reads the data_10A6040C damage cache, and formats through the retail %d literal.
3. Pinned the qagame-to-cgame compact weapon-damage wiring for the selected tail weapon band.
4. Added focused structural coverage for constants, retail target-group membership, resolver returns, cache reads, weapon mapping, server send path, cgame parse path, dispatcher guard, callback absence, and direct-switch absence.
5. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A170: Close selected 85-89 cgame model/second-place/chat ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, tests/test_cgame_ownerdraw_text_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_1STPLACE_PLYR_MODEL, CG_2NDPLACE, CG_2ND_PLACE_SCORE, CG_PLAYER_OBIT, and CG_AREA_NEW_CHAT ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher for raw ownerdraw cases 0x55 through 0x59.
2. Confirmed CG_1STPLACE_PLYR_MODEL, CG_2NDPLACE, CG_2ND_PLACE_SCORE, CG_PLAYER_OBIT, and CG_AREA_NEW_CHAT route to the first-place model, compact score, wide second-place score, obituary feed, and new-chat helpers respectively.
3. Confirmed the selected batch preserves the intended competitive-score refresh boundary: first-place model and second-place score rows refresh, while obituary and chat ownerdraws do not.
4. Added focused structural coverage for constants, retail target membership, source switch routes, helper behavior, shipped menu reachability, competitive refresh, and callback absence for the selected five.
5. Refreshed the cgame ownerdraw and mapping ledgers for this selected batch.

Task A169: Close selected 80-84 cgame award/scoreboard ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, src/code/game/g_main.c, tests/test_cgame_ownerdraw_text_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_MOST_DAMAGEDEALT_PLYR, CG_SPECTATORS, CG_MATCH_WINNER, CG_1STPLACE, and CG_1ST_PLACE_SCORE ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, 0x1003A0D0 award-player HLIL leaf, and adjacent scoreboard call targets for raw ownerdraw cases 0x50 through 0x54.
2. Confirmed CG_MOST_DAMAGEDEALT_PLYR remains the final pre-switch award-player image route fed by CS_AWARD_MOST_DAMAGEDEALT.
3. Confirmed CG_SPECTATORS, CG_MATCH_WINNER, CG_1STPLACE, and CG_1ST_PLACE_SCORE route to the spectator list, match winner, compact score, and wide first-place score helpers respectively.
4. Added focused structural coverage for constants, retail target membership, pre-switch/switch separation, configstring producer/consumer wiring, shipped menu reachability, competitive refresh, and callback absence for the selected five.
5. Refreshed the cgame ownerdraw and mapping ledgers for this selected batch.

Task A168: Close selected 157-161 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, src/code/cgame/cg_servercmds.c, src/code/game/g_cmds.c, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_1ST_PLYR_DMG_GL, CG_1ST_PLYR_DMG_RL, CG_1ST_PLYR_DMG_LG, CG_1ST_PLYR_DMG_RG, and CG_1ST_PLYR_DMG_PG ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, 0x10036C30 weapon-damage leaf, and 0x1002E2B0 placement weapon resolver for raw ownerdraw cases 0x9d through 0xa1.
2. Confirmed the selected damage band resolves to GL/RL/LG/RG/PG weapon indexes 4 through 8, reads the data_10A6040C damage cache, and formats through the retail %d literal.
3. Pinned the qagame-to-cgame compact weapon-damage wiring for the selected middle weapon band.
4. Added focused structural coverage for constants, retail target-group membership, resolver returns, cache reads, weapon mapping, server send path, cgame parse path, dispatcher guard, callback absence, and direct-switch absence.
5. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A167: Close selected 75-79 cgame award-player ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, src/code/game/g_main.c, tests/test_cgame_ownerdraw_text_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_MOST_VALUABLE_OFFENSIVE_PLYR, CG_MOST_VALUABLE_DEFENSIVE_PLYR, CG_MOST_VALUABLE_PLYR, CG_BEST_ITEMCONTROL_PLYR, and CG_MOST_ACCURATE_PLYR ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and 0x1003A0D0 award-player HLIL leaf for raw ownerdraw cases 0x4B through 0x4F.
2. Confirmed the selected ownerdraws route through the pre-switch award-player helper rather than the ordinary CG_OwnerDraw switch body.
3. Confirmed the selected award-player helpers are fed by qagame-published CS_AWARD_* configstrings and cgame resolves them before drawing the profile image for the winning client.
4. Added focused structural coverage for constants, retail target membership, pre-switch ordering, configstring producer/consumer wiring, shipped menu reachability, competitive refresh, and callback absence for the selected five.
5. Refreshed the cgame ownerdraw and mapping ledgers for this selected batch.

Task A166: Close selected 152-156 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, src/code/cgame/cg_servercmds.c, src/code/game/g_cmds.c, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_1ST_PLYR_SHOTS_PL, CG_1ST_PLYR_SHOTS_HMG, CG_1ST_PLYR_DMG_G, CG_1ST_PLYR_DMG_MG, and CG_1ST_PLYR_DMG_SG ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, 0x10036B00 weapon-shot leaf, 0x10036C30 weapon-damage leaf, and 0x1002E2B0 placement weapon resolver for raw ownerdraw cases 0x98 through 0x9c.
2. Tightened placement weapon-damage formatting to the retail %d literal, matching the shared data_10068E44 format used by the damage helper.
3. Confirmed the selected shot tail resolves to PL/HMG weapon indexes 0xc and 0xe, while the selected damage head resolves to G/MG/SG weapon indexes 1, 2, and 3.
4. Pinned the qagame-to-cgame compact accuracy_shots and weapon-damage wiring used by the selected ownerdraws.
5. Added focused structural coverage for constants, retail target-group membership, resolver returns, cache reads, weapon mapping, server send path, cgame parse path, dispatcher guard, callback absence, and direct-switch absence.
6. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A165: Close gauntlet/impressive/rampage/midair/perfect cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, tests/test_cgame_ownerdraw_text_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_GAUNTLET, CG_IMPRESSIVE, CG_RAMPAGES, CG_MIDAIRS, and CG_PERFECT ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and 0x10035340 medal HLIL leaf for raw ownerdraw cases 0x46 through 0x4A.
2. Confirmed CG_GAUNTLET, CG_IMPRESSIVE, and CG_PERFECT route through the real medal target group and read the selected-score gauntlet, impressive, and perfect fields.
3. Confirmed CG_RAMPAGES and CG_MIDAIRS remain retail no-ops between medal IDs with no authored menu usage and no helper route.
4. Added focused structural coverage for constants, retail switch membership, source medal/no-op separation, selected-score field reads, menu reachability, and callback absence for the selected five.
5. Refreshed the cgame ownerdraw and mapping ledgers for this selected batch.

Task A164: Close assist/capture/combo/defend/excellent cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, tests/test_cgame_ownerdraw_text_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_ASSISTS, CG_CAPTURES, CG_COMBOKILLS, CG_DEFEND, and CG_EXCELLENT ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and 0x10035340 medal HLIL leaf for raw ownerdraw cases 0x41 through 0x45.
2. Confirmed CG_ASSISTS, CG_CAPTURES, CG_DEFEND, and CG_EXCELLENT route through the real medal target group and read the selected-score assist, capture, defend, and excellent fields.
3. Confirmed CG_COMBOKILLS remains a retail no-op between medal IDs with no authored menu usage and no helper route.
4. Added focused structural coverage for constants, retail switch membership, source medal/no-op separation, selected-score field reads, menu reachability, and callback absence for the selected five.
5. Refreshed the cgame ownerdraw and mapping ledgers for this selected batch.

Task A163: Close selected 147-151 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, src/code/cgame/cg_servercmds.c, src/code/game/g_cmds.c, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_1ST_PLYR_SHOTS_RG, CG_1ST_PLYR_SHOTS_PG, CG_1ST_PLYR_SHOTS_BFG, CG_1ST_PLYR_SHOTS_CG, and CG_1ST_PLYR_SHOTS_NG ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, 0x10036B00 weapon-shot leaf, and 0x1002E2B0 placement weapon resolver for raw ownerdraw cases 0x93 through 0x97.
2. Confirmed the selected shot band resolves to RG/PG/BFG/CG/NG weapon indexes 7, 8, 9, 0xd, and 0xb, reads the data_10A6038C shot cache, and formats through the retail %d literal.
3. Pinned the qagame-to-cgame compact accuracy_shots wiring for the selected middle/tail weapon band.
4. Added focused structural coverage for constants, retail target-group membership, resolver returns, cache reads, weapon mapping, server send path, cgame parse path, dispatcher guard, callback absence, and direct-switch absence.
5. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A162: Close selected 142-146 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, src/code/cgame/cg_servercmds.c, src/code/game/g_cmds.c, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_1ST_PLYR_SHOTS_MG, CG_1ST_PLYR_SHOTS_SG, CG_1ST_PLYR_SHOTS_GL, CG_1ST_PLYR_SHOTS_RL, and CG_1ST_PLYR_SHOTS_LG ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, 0x10036B00 weapon-shot leaf, and 0x1002E2B0 placement weapon resolver for raw ownerdraw cases 0x8e through 0x92.
2. Tightened placement weapon-shot formatting to the retail %d literal, matching the shared data_10068E44 format used by the shot helper.
3. Confirmed the selected shot head resolves to MG/SG/GL/RL/LG weapon indexes 2 through 6, reads the data_10A6038C shot cache, and is backed by compact accuracy_shots qagame-to-cgame wiring.
4. Added focused structural coverage for constants, retail target-group membership, resolver returns, cache reads, weapon mapping, server send path, cgame parse path, dispatcher guard, callback absence, and direct-switch absence.
5. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A161: Close powerup/teamcolor/killer/accuracy cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, tests/test_cgame_ownerdraw_text_parity.py, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_CTF_POWERUP, CG_AREA_POWERUP, CG_TEAM_COLOR, CG_KILLER, and CG_ACCURACY ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and HLIL leaves for raw ownerdraw cases 0x3C through 0x40.
2. Confirmed CG_CTF_POWERUP uses the persisted powerup item icon path, CG_AREA_POWERUP uses the retail powerup sprite stack, and CG_TEAM_COLOR lands on the lower team-background helper with the local team value.
3. Corrected the powerup frag badge format to the retail x %i literal for the quad and battlesuit badge path.
4. Confirmed CG_KILLER is draw-gated by the cached killer name and shares the same Fragged by %s text with the width callback.
5. Confirmed CG_ACCURACY remains in the real medal switch group with %i%% text and alpha promotion for nonzero accuracy.
6. Added focused structural coverage for constants, retail targets, source dispatcher wiring, callback participation, menu reachability, and helper literals for the selected five.

Task A160: Close selected 137-141 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, src/code/cgame/cg_servercmds.c, src/code/game/g_cmds.c, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_1ST_PLYR_HITS_BFG, CG_1ST_PLYR_HITS_CG, CG_1ST_PLYR_HITS_NG, CG_1ST_PLYR_HITS_PL, and CG_1ST_PLYR_HITS_HMG ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, 0x100369D0 weapon-hit leaf, and 0x1002E2B0 placement weapon resolver for raw ownerdraw cases 0x89 through 0x8d.
2. Confirmed the selected hit tail resolves to BFG/CG/NG/PL/HMG weapon indexes 9, 0xd, 0xb, 0xc, and 0xe, reads the data_10A603CC hit cache, and formats through the retail %d literal.
3. Pinned the qagame-to-cgame compact accuracy_hits wiring for the selected tail weapon band.
4. Added focused structural coverage for constants, retail target-group membership, resolver returns, cache reads, weapon mapping, server send path, cgame parse path, dispatcher guard, callback absence, and direct-switch absence.
5. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A159: Close selected 132-136 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, src/code/cgame/cg_servercmds.c, src/code/game/g_cmds.c, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_1ST_PLYR_HITS_GL, CG_1ST_PLYR_HITS_RL, CG_1ST_PLYR_HITS_LG, CG_1ST_PLYR_HITS_RG, and CG_1ST_PLYR_HITS_PG ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, 0x100369D0 weapon-hit leaf, and 0x1002E2B0 placement weapon resolver for raw ownerdraw cases 0x84 through 0x88.
2. Confirmed the selected hit band resolves to GL/RL/LG/RG/PG weapon indexes 4 through 8, reads the data_10A603CC hit cache, and formats through the retail %d literal.
3. Pinned the qagame-to-cgame compact accuracy_hits wiring for the selected middle weapon band.
4. Added focused structural coverage for constants, retail target-group membership, resolver returns, cache reads, weapon mapping, server send path, cgame parse path, dispatcher guard, callback absence, and direct-switch absence.
5. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A158: Close flag/harvester/key cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, tests/test_cgame_displaycontext_parity.py, tests/test_cgame_ownerdraw_text_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_PLAYER_HASFLAG, CG_PLAYER_HASFLAG2D, CG_HARVESTER_SKULLS, CG_HARVESTER_SKULLS2D, and CG_PLAYER_HASKEY ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and HLIL leaves for raw ownerdraw cases 0x37, 0x38, 0x39, 0x3A, and 0x3B.
2. Corrected the Harvester 3D cube ownerdraw geometry to match the retail rect->y + 5 offset while preserving the forced-2D x + 3, y + 16 cube-icon path.
3. Confirmed the carried-flag ownerdraw keeps the retail 4-pixel 3D inset, forced-2D zero inset, and 1FCTF neutral-flag dropflag prompt.
4. Confirmed the carried-key ownerdraw uses the replicated key-mask stat, IT_KEY tag lookup, item visual registration, and half-width icon overlap.
5. Added focused structural coverage for constants, dispatcher routes, CG_GetValue participation, width/key callback absence, display-context wiring, retail HLIL anchors, and menu reachability.
6. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A157: Close selected 127-131 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, src/code/cgame/cg_servercmds.c, src/code/game/g_cmds.c, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_1ST_PLYR_FRAGS_NG, CG_1ST_PLYR_FRAGS_PL, CG_1ST_PLYR_FRAGS_HMG, CG_1ST_PLYR_HITS_MG, and CG_1ST_PLYR_HITS_SG ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, 0x100368A0 weapon-frag leaf, 0x100369D0 weapon-hit leaf, and 0x1002E2B0 placement weapon resolver for raw ownerdraw cases 0x7f through 0x83.
2. Tightened placement weapon-hit formatting to the retail %d literal, matching the shared data_10068E44 format used by the hit helper.
3. Confirmed the selected frag tail is backed by the compact frag order and the selected hit head is backed by the compact accuracy_hits order across qagame send and cgame parse paths.
4. Added focused structural coverage for constants, retail target-group membership, resolver returns, cache reads, weapon mapping, server send path, cgame parse path, dispatcher guard, callback absence, and direct-switch absence.
5. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A156: Harden selected 86-90 cgame endgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, src/code/cgame/cg_main.c, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_2NDPLACE, CG_2ND_PLACE_SCORE, CG_PLAYER_OBIT, CG_AREA_NEW_CHAT, and CG_PLYR_END_GAME_SCORE ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and HLIL call sites for raw ownerdraw cases 0x56, 0x57, 0x58, 0x59, and 0x5A.
2. Confirmed current source preserves the compact second-place score helper, wide second-place score branch, obituary feed renderer, timed chat stack, and local endgame summary text family.
3. Added focused structural coverage for source switch wiring, display-context callback absence, shipped menu reachability, competitive cache refresh, and helper behavior for the selected endgame slice.
4. Refreshed the ownerdraw parity index and cgame mapping notes for this selected batch.

Task A155: Harden selected 91-95 cgame selected/follow ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, src/code/cgame/cg_main.c, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_PLYR_BEST_WEAPON_NAME, CG_SELECTED_PLYR_TEAM_COLOR, CG_SELECTED_PLYR_ACCURACY, CG_FOLLOW_PLAYER_NAME, and CG_FOLLOW_PLAYER_NAME_EX ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and HLIL call sites for raw ownerdraw cases 0x5B, 0x5C, 0x5D, 0x5E, and 0x5F.
2. Confirmed current source preserves the selected-score bounds checks, selected-client team-color fill, selected accuracy text path, best-weapon label table, and shared follow-name prefix/tint/alignment helper.
3. Added focused structural coverage for source switch wiring, display-context callback absence, shipped menu reachability, and the helper behavior that backs each selected ownerdraw.
4. Refreshed the ownerdraw parity index and cgame mapping notes for this selected/follow-player slice.

Task A154: Harden selected 81-85 cgame scoreboard ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, src/code/cgame/cg_main.c, src/code/cgame/cg_servercmds.c, src/code/game/g_main.c, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_SPECTATORS, CG_MATCH_WINNER, CG_1STPLACE, CG_1ST_PLACE_SCORE, and CG_1STPLACE_PLYR_MODEL ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, online-service, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and HLIL call sites for raw ownerdraw cases 0x51, 0x52, 0x53, 0x54, and 0x55.
2. Confirmed current source preserves the retail spectator-cache gate, match-winner text branch, compact score helper, wide first-place configstring line, and first-place model/profile preview path.
3. Added focused structural coverage for source switch wiring, display-context callback absence, qagame/cgame configstring producer-consumer paths, and shipped menu reachability.
4. Refreshed the ownerdraw parity index and cgame mapping notes for this selected scoreboard/endgame slice.

Task A153: Close renderer post-process command payload wiring [COMPLETED]

Priority: High Primary areas: src/code/renderer/tr_backend.c, tests/test_renderer_post_process_parity.py, tests/test_renderer_internal_helper_mapping_parity.py Parity estimate: before 99.99% -> after 100% for the scoped renderer post-process command payload wiring lane. The strict renderer estimate remains 100% and repo-wide parity remains 98% because unrelated portability, online-service, packaging, and source-legacy surfaces are unchanged.

Completed work:

1. Rechecked the retail quakelive_steam.exe HLIL for RB_ColorCorrectPostProcessCommand @ 0x00436DC0, RB_BloomPostProcessCommand @ 0x00436EC0, R_AddBloomPostProcessCommand @ 0x004384D0, and R_AddColorCorrectPostProcessCommand @ 0x0043CD10.
2. Rewired the color-correct pass to consume the queued command texture and program handles instead of re-reading the live s_postProcess handles.
3. Rewired the bloom downsample, bright-pass, blur, quarter-pass, combine, and fallback blit path to consume the queued bloom command payload while keeping framebuffer binding under the recovered retail render-target owner.
4. Added focused structural coverage that pins command payload consumption for color-correct and bloom execution.

Task A152: Close referenced-pak publication parity [COMPLETED]

Priority: High Primary areas: src/code/qcommon/files.c, tests/fs_searchpath_harness.c, tests/test_fs_search_paths.py, tests/test_engine_cvar_retail_parity.py Parity estimate: before 99% -> after 100% for the scoped sv_referencedPaks / sv_referencedPakNames filesystem publication and client autodownload decision lane. Repo-wide parity remains 98% because unrelated compatibility, portability, online-service, packaging, and source-legacy surfaces are unchanged.

Completed work:

1. Rechecked the retail quakelive_steam.exe HLIL for FS_ReferencedPakChecksums at 0x004D1DC0 and FS_ReferencedPakNames at 0x004D1E20.
2. Removed the legacy Quake III fs_game-wide publication rule from the referenced-pak checksum and name builders so Quake Live now publishes only pk3s whose retail reference flags are nonzero.
3. Kept the existing server sv_referencedPaks / sv_referencedPakNames publication and client FS_PureServerSetReferencedPaks parse wiring intact.
4. Added filesystem-harness coverage proving an unreferenced mod pk3 is not published until its reference flag is set, plus source-visible HLIL guards for the recovered retail flag-only condition.

Task A151: Tighten native game API handoff parity [COMPLETED]

Priority: High Primary areas: src/code/server/sv_game.c, src/code/client/cl_cgame.c, tests/test_game_native_export_helper_parity.py Parity estimate: before 99.9% -> after 100% for the scoped native qagame/cgame API-version handoff and retained QVM fallback ownership slice. Repo-wide parity remains 98% because unrelated portability, online-service, packaging, and source-legacy surfaces are unchanged.

Completed work:

1. Rechecked the retail qagamex86.dll and cgamex86.dll dllEntry evidence: qagame reports native API 10, and cgame reports native API 8.
2. Split the native-load API constants from the legacy VM fallback constants at the qagame and cgame loader call sites.
3. Added structural coverage that pins native API handoff, legacy fallback API handoff, and the retained retail QVM fallback ownership note.
4. Revalidated the focused native-export and qcommon fallback-VM parity probes.

Task A150: Close item/score/race cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, tests/test_cgame_ownerdraw_text_parity.py, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_PLAYER_ITEM, CG_PLAYER_SCORE, CG_RACE_STATUS, CG_RACE_TIMES, and CG_ONEFLAG_STATUS ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and HLIL leaves for raw ownerdraw cases 0x32, 0x33, 0x34, 0x35, and 0x36.
2. Confirmed current source preserves the retail holdable item branch, compact score wrapper, Race status/timing shared leaf, callback participation, and shipped menu reachability for the selected five.
3. Restored the CG_ONEFLAG_STATUS visible leaf to the retail gfx/2d/flag_status/* shader bank, white draw color, and stolen-flag score-dependent y offset.
4. Added focused structural coverage for constants, dispatcher routes, CG_GetValue participation, width/key callback absence, display-context wiring, retail HLIL anchors, and menu reachability.
5. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A149: Close selected 122-126 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, src/code/cgame/cg_servercmds.c, src/code/game/g_cmds.c, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_1ST_PLYR_FRAGS_LG, CG_1ST_PLYR_FRAGS_RG, CG_1ST_PLYR_FRAGS_PG, CG_1ST_PLYR_FRAGS_BFG, and CG_1ST_PLYR_FRAGS_CG ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher, 0x100368A0 weapon-frag leaf, and 0x1002E2B0 placement weapon resolver for raw ownerdraw cases 0x7a through 0x7e.
2. Confirmed the retail resolver maps LG/RG/PG/BFG to weapon indexes 6 through 9 and maps CG to 0xd, preserving the Quake Live enum gap before nailgun/prox/chaingun.
3. Pinned the qagame-to-cgame compact scorestats weapon-frag wiring around the retail LG, RG, PG, BFG, CG lane inside the full frag order.
4. Added focused structural coverage for constants, retail target-group membership, resolver returns, weapon mapping, server send path, cgame parse path, dispatcher guard, callback absence, and direct-switch absence.
5. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A148: Harden selected 75-79 cgame award ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_MOST_VALUABLE_OFFENSIVE_PLYR, CG_MOST_VALUABLE_DEFENSIVE_PLYR, CG_MOST_VALUABLE_PLYR, CG_BEST_ITEMCONTROL_PLYR, and CG_MOST_ACCURATE_PLYR ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll ownerdraw dispatcher route for raw cases 0x4b through 0x4f, all landing on the shared award-player image leaf at 0x1003A0D0.
2. Verified the selected ownerdraws are handled before the source CG_OwnerDraw switch by CG_DrawAwardPlayer, fed by qagame-produced CS_AWARD_* configstrings and rendered through the profile-image helper.
3. Added focused structural coverage for constants, retail target routing, qagame producer wiring, cgame configstring consumption, shipped end-scoreboard menu reachability, display-context callbacks, and absence from value/width/key callback routes.
4. Refreshed the cgame ownerdraw and mapping ledgers so the award lane now documents the recovered configstring-backed profile-image path.

Task A147: Close selected 117-121 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, src/code/cgame/cg_servercmds.c, src/code/game/g_cmds.c, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_1ST_PLYR_FRAGS_G, CG_1ST_PLYR_FRAGS_MG, CG_1ST_PLYR_FRAGS_SG, CG_1ST_PLYR_FRAGS_GL, and CG_1ST_PLYR_FRAGS_RL ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and HLIL leaf at 0x100368A0, including the placement weapon-index helper, weapon-frag cache read, %d format literal, and baseline text paint.
2. Tightened the placement weapon-frag formatter to the retail %d literal.
3. Pinned the qagame-to-cgame compact scorestats weapon-frag wiring around the retail G, MG, SG, GL, RL prefix order.
4. Added focused structural coverage for constants, retail target-group membership, weapon mapping, server send path, cgame parse path, dispatcher guard, callback absence, and direct-switch absence.
5. Refreshed the cgame ownerdraw and mapping ledgers for the selected five.

Task A146: Close health/ammo cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_ownerdraw_text_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_PLAYER_HEALTH_BAR_100, CG_PLAYER_HEALTH_BAR_200, CG_PLAYER_AMMO_ICON, CG_PLAYER_AMMO_ICON2D, and CG_PLAYER_AMMO_VALUE ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and HLIL leaves for raw ownerdraw cases 0x2d, 0x2e, 0x2f, 0x30, and 0x31.
2. Confirmed current source preserves the retail primary/excess health-bar split, ammo icon 2D/3D branches, finite ammo text/shader rendering, and infinite-ammo icon fallback.
3. Added focused structural coverage for constants, dispatcher routes, CG_GetValue participation, width/key callback absence, display-context wiring, and shipped menu reachability.

Task A145: Close local status cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_ownerdraw_text_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_PLAYER_ARMOR_VALUE, CG_PLAYER_ARMOR_BAR_100, CG_PLAYER_ARMOR_BAR_200, CG_ARMORTIERED_COLORIZED, and CG_PLAYER_HEALTH ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and HLIL leaves for raw ownerdraw cases 0x28, 0x29, 0x2a, 0x2b, and 0x2c.
2. Confirmed current source preserves the retail armor/health shader-or-aligned-number paths, armor primary/excess bar clipping, and replicated armor-tier color fill.
3. Added focused structural coverage for constants, dispatcher routes, CG_GetValue participation, width/key callback absence, display-context wiring, and shipped menu reachability.

Task A144: Close selected 112-116 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 96% -> after 100% for the selected CG_1ST_PLYR_ACC, CG_1ST_PLYR_FLAG, CG_1ST_PLYR_AVATAR, CG_1ST_PLYR_TIMEOUT_COUNT, and CG_1ST_PLYR_HEALTH_ARMOR ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and HLIL leaves for 0x10036570, 0x10036720, 0x100366A0, 0x10033040, plus the raw 0x73 no-op gap.
2. Tightened total accuracy formatting to the retail %d%% literal and kept the placement flag route on the cached country-flag shader path.
3. Restored the placement avatar selection order: native avatar handle when cg_drawProfileImages and a replicated identity are available, then the cached model icon fallback, with no source-only dimming or armor fallback.
4. Rebuilt the spectator comparison ownerdraw around retail-style health and armor bands, including local playerState stat reads, tracked-client fallback, fixed retail colors, the short-lived damage-loss segment, primary-slot right alignment, cvar color gating, and removal of the non-retail text/marker overlay.
5. Added focused structural coverage for the five selected ownerdraw IDs and refreshed the cgame ownerdraw and mapping ledgers.

Task A143: Close spectator/player-preview cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_ownerdraw_text_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_SPEC_MESSAGES, CG_PLAYER_HEAD, CG_PLAYERMODEL, CG_PLAYER_ARMOR_ICON, and CG_PLAYER_ARMOR_ICON2D ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and HLIL leaves for raw ownerdraw cases 0x23, 0x24, 0x25, 0x26, and 0x27.
2. Confirmed current source preserves the spectator message copy family, local head damage/idle interpolation, tracked/local model preview, and shared armor 2D/3D icon behavior.
3. Added focused structural coverage for constants, dispatcher routes, shipped menu reachability, and absence from the value/width/key callback paths.

Task A142: Close vote shot/timer cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_ownerdraw_text_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 99% -> after 100% for the selected CG_VOTESHOT1, CG_VOTESHOT2, CG_VOTESHOT3, CG_VOTECOUNT3, and CG_VOTETIMER ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and HLIL leaves for raw ownerdraw cases 0x19, 0x1a, 0x1b, 0x21, and 0x22.
2. Confirmed current source preserves the retail vote preview path, default preview fallback, third vote-count label, timer text family, cvar payload wiring, and absence of value/width/key callback participation.
3. Added focused structural coverage for dispatcher slots, shipped menu usage, ui_vote%s%i shot/count payload transport, and vote active/string bridge state used by the timer menu.

Task A141: Close vote name/count cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, tests/test_cgame_ownerdraw_text_parity.py, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 95% -> after 100% for the selected CG_VOTENAME1, CG_VOTENAME2, CG_VOTENAME3, CG_VOTECOUNT1, and CG_VOTECOUNT2 ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and HLIL leaves for raw ownerdraw cases 0x1c through 0x20.
2. Restored the retail text-width alignment path for vote-name labels while preserving the Name then Map fallback payload order and raw rect-y paint origin.
3. Added focused structural coverage for dispatcher slots, display-context callbacks, ui_vote%s%i name/count payload wiring, shipped menu usage, and absence of value/width/key callback participation.

Task A140: Close selected 107-111 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 98% -> after 100% for the selected CG_1ST_PLYR_DEATHS, CG_1ST_PLYR_DMG, CG_1ST_PLYR_TIME, CG_1ST_PLYR_PING, and CG_1ST_PLYR_WINS ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and HLIL leaves for 0x10036320, 0x10036770, 0x10036070, and 0x10036450, plus the raw 0x6d no-op gap.
2. Tightened the placement ping ownerdraw to use the retail >40 and >80 warning-color thresholds, forced blue to 0.0, and forced alpha to 0.8 before painting.
3. Restored the wins slot to the retail preformatted %d/%d wins/losses text path rather than the source-only bare wins count.
4. Added focused structural coverage for the five selected ownerdraw IDs, including dispatcher routes, metric helpers, the no-op time case, retail color evidence, and the absence of value/width/key callback wiring.
5. Updated the cgame ownerdraw and mapping ledgers with the exact placement metric and ping-color evidence.

Task A139: Close vote gametype/map cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, tests/test_cgame_ownerdraw_text_parity.py, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 94% -> after 100% for the selected CG_VOTEGAMETYPE1, CG_VOTEGAMETYPE2, CG_VOTEGAMETYPE3, CG_VOTEMAP1, and CG_VOTEMAP2 ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and HLIL leaves for raw ownerdraw cases 0x13 through 0x17.
2. Restored the retail text-width alignment path for vote gametype and raw-map labels while preserving the gametype row's rect->y - 8.0f paint bias.
3. Added focused structural coverage for dispatcher slots, display-context callbacks, ui_vote%s%i cvar payload wiring, menu reachability, and the absence of value/width/key callback participation.

Task A138: Tighten 83-87 cgame placement ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 98% -> after 100% for the selected CG_1STPLACE, CG_1ST_PLACE_SCORE, CG_1STPLACE_PLYR_MODEL, CG_2NDPLACE, and CG_2ND_PLACE_SCORE ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll HLIL and committed Ghidra dispatcher for raw ownerdraw cases 0x53 through 0x57.
2. Restored compact placement score formatting through the shared retail score helper, including Race [-]Ns text and the SCORE_NOT_PRESENT skip.
3. Wired the compact placement ownerdraws into the competitive score-cache touch path used by HUD-only placement layouts.
4. Tightened the wide second-place row to use the retail spectator-team gate before choosing between the local player row and cached runner-up row.
5. Refreshed focused structural tests and ownerdraw/mapping ledgers for the completed placement lane.

Task A137: Close selected 100-105 cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md Parity estimate: before 98% -> after 100% for the selected CG_TEAM_PLYR_COUNT, CG_ENEMY_PLYR_COUNT, CG_1ST_PLYR, CG_1ST_PLYR_READY, and CG_1ST_PLYR_SCORE ownerdraw set. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher and HLIL leaves for 0x10033650, 0x10035BD0, 0x10037BA0, and 0x10035F30.
2. Corrected the friendly/enemy round player-count ownerdraws to paint the cached team count at the raw ownerdraw origin, matching the retail sub_10033650 text path instead of the older centered source behavior.
3. Added focused structural coverage for the five selected ownerdraw IDs, including dispatcher routes, pre-switch CG_1ST_PLYR_READY handling, raw text origins, and the absence of value/width/key callback wiring.
4. Updated the cgame ownerdraw and mapping ledgers with the raw-origin player-count evidence.

Task A136: Close gametype cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_ownerdraw_text_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 98% -> after 100% for the selected CG_GAME_TYPE, CG_GAME_TYPE_ICON, and CG_GAME_TYPE_MAP ownerdraw trio. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail cgamex86.dll dispatcher evidence for raw ownerdraw cases 4, 5, and 6, mapped to CG_DrawGameType, CG_DrawGameTypeIcon, and CG_DrawGameTypeMap.
2. Confirmed the current source preserves the direct draw routes, keeps only CG_GAME_TYPE on the ownerdraw-width callback, and leaves the trio out of the value and key-handler callbacks.
3. Added focused structural coverage for the dispatcher, display-context callbacks, shipped menu reachability, gametype icon registration path, and intro-panel gametype-map wiring.

Task A135: Restore retail post-process gamma capture path [COMPLETED]

Priority: High Primary areas: src/code/renderer/tr_init.c, docs/renderer_cvar_matrix.md, tests/test_engine_cvar_retail_parity.py Parity estimate: before 99% -> after 100% for the selected renderer post-process gamma ownership lane. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail quakelive_steam.exe HLIL/Ghidra evidence for r_enablePostProcess, r_enableColorCorrect, the p_gammaRecip color-correct uniform, and the legacy Win32 gamma ramp fallback.
2. Removed the reconstruction-only r_enablePostProcess force-off so the shader-backed color-correct pass owns final r_gamma output by default, matching retail and improving Windows desktop/PrintScreen capture parity.
3. Updated cvar parity tests and renderer cvar documentation to reflect the restored retail default and fallback boundary.

Task A134: Close player/opponent UI ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/ui/ui_main.c, tests/test_ui_menu_files.py, docs/reverse-engineering/ui-ownerdrawtype-parity-index.md Parity estimate: before 76% -> after 100% for the selected UI_PLAYERMODEL, UI_OPPONENTMODEL, and UI_OPPONENT_NAME ownerdraw trio. Repo-wide parity remains 98% because unrelated compatibility, source-legacy, and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail uix86.dll HLIL/Ghidra evidence for UI_DrawPlayerModel at 0x10005690, UI_DrawOpponent at 0x10006730, UI_DrawOpponentName at 0x100068F0, and UI_OwnerDrawHandleKey at 0x1000A820.
2. Rebuilt UI_PLAYERMODEL around the retail model/headmodel cvar reads, empty team name, cg_loadout/cg_weaponPrimary primary weapon selection, weapon item validation, heavy-machinegun fallback for invalid primary weapon values, retail 5/210/0 view angles, and the pre-draw cached playerInfo_t reset observed in HLIL.
3. Tightened UI_OPPONENTMODEL cvar refresh to use buffered ui_opponentModel reads before rebuilding and drawing the cached playerInfo_t, including the retail 0/170/0 view angles and pre-draw cached preview reset.
4. Tightened UI_OPPONENT_NAME to the retail buffered ui_opponentName paint path, kept the empty width-helper case, and removed the source-only ownerdraw key route that retail does not dispatch.
5. Updated the UI ownerdraw parity index so all retail-routed Needs check rows are now checked.

Verification:

• python -m pytest tests/test_ui_menu_files.py -k "player_opponent_ownerdraws or handicap_netsource_netfilter"
  • Result: 2 passed, 76 deselected.
• python -m pytest tests/test_ui_menu_files.py
  • Result: 78 passed.

Task A133: Close front-panel cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: tests/test_cgame_ownerdraw_text_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 98% -> after 100% for the selected CG_SERVER_SETTINGS, CG_STARTING_WEAPONS, and CG_GAME_LIMIT ownerdraw trio. Repo-wide parity remains 98% because no unrelated compatibility or runtime-evidence surfaces changed.

Completed work:

1. Rechecked the retail CG_OwnerDraw dispatcher cases 1, 2, and 3 against the committed Ghidra/HLIL evidence.
2. Confirmed CG_SERVER_SETTINGS stays on the custom-settings text and modified-weapon icon-strip leaf.
3. Confirmed CG_STARTING_WEAPONS stays on the CS_LOADOUT_MASK icon-strip and queued-primary loadout preview leaf.
4. Confirmed CG_GAME_LIMIT stays on the narrowed retail Cap Limit / Frag Limit / Round Limit / Score Limit label family.
5. Added focused structural coverage that guards switch wiring, display-context callback non-participation, menudef constants, and shipped menu reachability for the two shipped front-panel menu users.

Task A132: Close objective/powerup cgame ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, tests/test_cgame_displaycontext_parity.py, tests/test_cgame_hud_parity.py, tests/test_game_key_item_parity.py, docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 98% -> after 100% for the selected CG_PLAYER_HASKEY, CG_CTF_POWERUP, and CG_AREA_POWERUP ownerdraw trio. Repo-wide parity remains 98% because unrelated dirty worktree and runtime evidence surfaces are unchanged.

Completed work:

1. Rechecked the retail cgame CG_OwnerDraw switch and leaves at 0x10031F90, 0x100310F0, and 0x10031160.
2. Reconstructed CG_PLAYER_HASKEY to use the retail BG_FindItemByTypeAndTag( IT_KEY, tag ) path for silver, gold, and master keys instead of a classname detour.
3. Confirmed CG_CTF_POWERUP remains the direct STAT_PERSISTANT_POWERUP item-icon ownerdraw.
4. Confirmed CG_AREA_POWERUP is directly wired to the retail powerup stack helper and has no cg_drawSprites or cg_drawSpriteSelf ownerdraw gate.
5. Updated the cgame ownerdraw index so all menudef cgame rows are now checked or checked retail no-ops.

Task A131: Add full cgame ownerdraw parity index [COMPLETED]

Priority: High Primary areas: docs/reverse-engineering/cg-ownerdrawtype-parity-index.md, tests/test_cgame_displaycontext_parity.py Parity estimate: before 0% -> after 100% for durable index coverage of the src/ui/menudef.h cgame ownerdraw block. The current ledger has 327 checked rows, 12 checked retail no-ops, and no open partial rows after the objective/powerup ownerdraw follow-up.

Completed work:

1. Added a full tabulated listing for menudef cgame ownerdraw IDs 1..339, excluding CG_SHOW_* flags, UI_* ownerdraws, and ui_shared.h-only source aliases outside the menudef range.
2. Recorded the current cgame implementation route and parity state for every row so future parity-picking has a stable index.
3. Added a structural test that compares the ledger rows back to the menudef cgame block and rejects accidental UI/visibility-flag leakage.

Task A130: Close medal-adjacent cgame ownerdraw no-op parity [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, tests/test_cgame_displaycontext_parity.py, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 95% -> after 100% for the focused CG_COMBOKILLS, CG_RAMPAGES, and CG_MIDAIRS cgame ownerdraw lane. Repo-wide parity remains 98% because unrelated dirty worktree changes and non-Windows/runtime-evidence lanes are unchanged.

Completed work:

1. Rechecked the committed cgame ownerdraw switch evidence for 0x1003B0F0 -> CG_OwnerDraw and the medal helper route at 0x10035340 -> CG_DrawMedal.
2. Confirmed retail routes only CG_ACCURACY, CG_ASSISTS, CG_CAPTURES, CG_DEFEND, CG_EXCELLENT, CG_GAUNTLET, CG_IMPRESSIVE, and CG_PERFECT to the medal helper; CG_COMBOKILLS (0x43), CG_RAMPAGES (0x48), and CG_MIDAIRS (0x49) are absent from the retail ownerdraw switch.
3. Made the three absent ownerdraws explicit no-op cases in CG_OwnerDraw, preserving no menu usage, no CG_GetValue route, no CG_OwnerDrawWidth route, and the retail null key-handler behavior.
4. Added focused structural coverage for the retail switch absence, medal helper exclusion, display-context callback non-participation, and shipped menu reachability.

Verification:

• python -m pytest tests/test_cgame_displaycontext_parity.py -k "combo_rampage_midair or team_pickup_timeheld_and_medal_dispatch_groups" -q
  • Result: 2 passed, 145 deselected.
• python -m pytest tests/test_cgame_displaycontext_parity.py -k "combo_rampage_midair or ownerdraw_dispatch_covers_committed_retail_switch_cases or ownerdraw_width_callback" -q
  • Result: 3 passed, 144 deselected.

Task A129: Close team overlay transport and drawing parity [COMPLETED]

Priority: High Primary areas: src/code/game/g_team.c, src/code/cgame/cg_servercmds.c, src/code/game/bg_public.h, src/code/cgame/cg_draw.c, tests/test_cgame_hud_parity.py Parity estimate: before 98% -> after 100% for the focused team-overlay transport/cache and classic HUD drawing lane. Repo-wide parity remains 98% because unrelated dirty worktree changes and non-Windows/runtime-evidence lanes are unchanged.

Completed work:

1. Rechecked the retail evidence for cgame CG_DrawTeamOverlay at 0x10009DA0, cgame CG_ParseTeamInfo at 0x100487B0, and qagame TeamplayInfoMessage at 0x10068490.
2. Restored the retail eight-player overlay cap and made qagame format the sorted top-eight teammate set it already selected from level.sortedClients.
3. Kept the emitted tinfo row count tied to successfully serialized rows and guarded the cgame receiver against oversized, truncated, or invalid client rows before updating sortedTeamPlayers.
4. Matched the retail drawing gates and row body: invulnerability-expand early return, integer row placement, ITEM_TEXTSTYLE_SHADOWED host text, sans name/location columns, mono health/armor columns, EF_DEAD dead-row gating, and a powerup icon ladder that suppresses the neutral flag and frozen sentinel while preserving the invulnerability item path.
5. Added a focused parity test covering the eight-row cap, sorted top-eight payload order, tinfo row count, and receiver-side bounds.

Verification:

• python -m pytest tests/test_cgame_hud_parity.py -k "team_overlay" -q
  • Result: 3 passed, 29 deselected.
• python -m pytest tests/test_cgame_hud_parity.py tests/test_game_runframe_parity.py -q
  • Result: 35 passed, 2 failed; the two failures are unrelated existing HUD assertions for CG_DrawTeamInfoRow width and CG_DrawAreaPowerUp sprite gating on the dirty worktree.

Task A128: Close map selection UI ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/ui/ui_main.c, tests/test_ui_menu_files.py, docs/reverse-engineering/ui-ownerdrawtype-parity-index.md Parity estimate: before 98% -> after 100% for the focused UI_ALLMAPS_SELECTION, UI_MAPS_SELECTION, and UI_STARTMAPCINEMATIC ownerdraw lane. Repo-wide parity remains 98% because unrelated compatibility and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail uix86.dll HLIL/Ghidra evidence for UI_DrawAllMapsSelection at 0x10006890 and UI_DrawMapCinematic at 0x10005560.
2. Confirmed UI_ALLMAPS_SELECTION uses the retail shared map-name draw path with the net/current selector flag, guards the map index against the retail active map count, and has an empty width-helper case with no key handler.
3. Confirmed UI_MAPS_SELECTION matches the retail current-map name paint path and is reconstructed through the shared source helper with the current map selector flag.
4. Confirmed UI_STARTMAPCINEMATIC uses the retail map cinematic helper with the start/net selector flag, including map index reset, %s.roq lazy start, run/extents/draw behavior, and preview fallback.
5. Updated the ownerdraw parity index so the three selected IDs are now marked checked and the unchecked retail-routed set is down to UI_PLAYERMODEL, UI_OPPONENTMODEL, and UI_OPPONENT_NAME.

Verification:

• python -m pytest tests/test_ui_menu_files.py -k "map_selection_ownerdraws"
  • Result: 1 passed, 76 deselected.
• python -m pytest tests/test_ui_menu_files.py
  • Result: 77 passed.

Task A127: Close server info UI ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/ui/ui_main.c, tests/test_ui_menu_files.py, docs/reverse-engineering/ui-ownerdrawtype-parity-index.md Parity estimate: before 98% -> after 100% for the focused UI_SERVERREFRESHDATE, UI_SERVERMOTD, and UI_GLINFO ownerdraw lane. Repo-wide parity remains 98% because unrelated compatibility and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail uix86.dll HLIL/Ghidra evidence for UI_DrawServerRefreshDate at 0x10008F20, UI_DrawServerMOTD at 0x10009080, and UI_DrawGLInfo at 0x100093B0.
2. Confirmed UI_SERVERREFRESHDATE matches the retail active-refresh pulse path, server-count message, cached ui_lastServerRefresh_%i timestamp path, draw dispatcher route, and UI_OwnerDrawWidth route.
3. Confirmed UI_SERVERMOTD matches the retail MOTD scroll state, wraparound dual-paint behavior, and the cl_motdString fallback seeding to Welcome to Team Arena! before drawing.
4. Confirmed UI_GLINFO matches the retail vendor/version/pixel-format paint lines and the extension-string split/wrap behavior.
5. Confirmed none of the three selected ownerdraws has retail key-handler wiring, and only UI_SERVERREFRESHDATE participates in the width helper.
6. Updated the ownerdraw parity index so the three selected IDs are now marked checked and the unchecked retail-routed set reflects the remaining work.

Verification:

• python -m pytest tests/test_ui_menu_files.py -k "server_info_ownerdraws"
  • Result: 1 passed, 75 deselected.
• python -m pytest tests/test_ui_menu_files.py
  • Result: 76 passed.

Task A126: Close map media UI ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/ui/ui_main.c, tests/test_ui_menu_files.py, docs/reverse-engineering/ui-ownerdrawtype-parity-index.md Parity estimate: before 98% -> after 100% for the focused UI_NETMAPPREVIEW, UI_MAPCINEMATIC, and UI_NETMAPCINEMATIC ownerdraw lane. Repo-wide parity remains 98% because unrelated compatibility and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail uix86.dll HLIL/Ghidra evidence for UI_DrawMapCinematic at 0x10005560, UI_DrawNetMapPreview at 0x100065B0, and UI_DrawNetMapCinematic at 0x10006600.
2. Confirmed UI_NETMAPPREVIEW matches the retail current-server preview shader draw path and the menu/art/unknownmap fallback when no preview shader is available.
3. Confirmed UI_MAPCINEMATIC matches the retail map selector validation, lazy %s.roq cinematic start, run/extents/draw sequence, failure marker, preview fallback, and ownerdraw dispatcher route.
4. Confirmed UI_NETMAPCINEMATIC matches the retail current-server cinematic run/extents/draw sequence, current-net-map reset guard, preview fallback, and ownerdraw dispatcher route.
5. Locked the shared cinematic cleanup wiring: _UI_Init installs UI_StopCinematic, positive handles stop directly, and ownerdraw cleanup handles UI_MAPCINEMATIC and UI_NETMAPCINEMATIC by resetting their backing cinematic handles to -1.
6. Updated the ownerdraw parity index so the three selected IDs are now marked checked and the unchecked retail-routed set reflects the remaining work.

Verification:

• python -m pytest tests/test_ui_menu_files.py -k "map_media_ownerdraws"
  • Result: 1 passed, 74 deselected.
• python -m pytest tests/test_ui_menu_files.py
  • Result: 75 passed.

Task A125: Close skill and map preview UI ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/ui/ui_main.c, tests/test_ui_menu_files.py Parity estimate: before 98% -> after 100% for the focused UI_SKILL, UI_MAPPREVIEW, and UI_MAP_TIMETOBEAT ownerdraw lane. Repo-wide parity remains 98% because unrelated compatibility and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail uix86.dll HLIL/Ghidra evidence for UI_DrawSkill at 0x10005350, UI_DrawMapPreview at 0x100053C0, UI_DrawMapTimeToBeat at 0x100054A0, and UI_Skill_HandleKey at 0x1000A390.
2. Confirmed UI_SKILL clamps g_spSkill into the retail one-to-five range, paints the matching skill table entry, and wires the ownerdraw key handler to the retail click/enter cycle behavior.
3. Confirmed UI_MAPPREVIEW validates the selected map slot, resets the current map cvars on invalid indexes, lazily caches levelshot shaders, falls back to menu/art/unknownmap, and uses the net-map preview path from the ownerdraw dispatcher.
4. Confirmed UI_MAP_TIMETOBEAT validates ui_currentMap, resolves the retail gametype enum before indexing the per-map time table, formats the value as %02i:%02i, and paints it through the ownerdraw dispatcher.
5. Added regression coverage binding all three ownerdraws to the promoted retail evidence addresses, draw dispatch wiring, key dispatch wiring, and fallback/reset behavior.

Verification:

• python -m pytest tests/test_ui_menu_files.py -k "skill_mappreview_maptime"
  • Result: 1 passed, 73 deselected.
• python -m pytest tests/test_ui_menu_files.py
  • Result: 74 passed.

Task A124: Close bot selector UI ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/ui/ui_main.c, tests/test_ui_menu_files.py Parity estimate: before 97% -> after 100% for the focused UI_BOTNAME, UI_BOTSKILL, and UI_REDBLUE ownerdraw lane. Repo-wide parity remains 98% because unrelated compatibility and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail uix86.dll HLIL/Ghidra evidence for UI_DrawBotName at 0x10006b30, UI_DrawBotSkill at 0x10006bc0, UI_DrawRedBlue at 0x10006c10, and the bot-name/bot-skill key handlers at 0x1000a570 and 0x1000a5d0.
2. Made the UI_BOTNAME draw path explicitly mirror the retail invalid-index behavior: clamp high indexes back to zero, print the retail invalid bot warning for remaining invalid indexes, and paint the Sarge fallback while keeping the add-bot selector bound to the bot catalog rather than the character roster.
3. Confirmed UI_BOTSKILL already matches the retail five-entry skill table paint path and wraparound key cycling.
4. Confirmed UI_REDBLUE already matches the retail Red/Blue paint path, toggle handler, and non-returning ownerdraw key dispatch.
5. Locked all three ownerdraw draw dispatches, key dispatches, key wrap paths, fallback behavior, and add-bot catalog wiring with focused regression coverage.

Verification:

• python -m pytest tests/test_ui_menu_files.py -k "botname_botskill_redblue or addbot_uses_bot_catalog"
  • Result: 2 passed, 71 deselected.
• python -m pytest tests/test_ui_menu_files.py
  • Result: 73 passed.

Task A123: Close handicap and server-browser selector UI ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/ui/ui_main.c, tests/test_ui_menu_files.py Parity estimate: before 96% -> after 99% for the focused UI_HANDICAP, UI_NETSOURCE, and UI_NETFILTER ownerdraw lane. Repo-wide parity remains 98% because unrelated compatibility and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail uix86.dll HLIL/Ghidra evidence for the handicap draw and key handler, plus the browser source/filter draw, width, and key handlers at 0x1000a040, 0x1000a420, and 0x1000a4f0.
2. Confirmed UI_HANDICAP already matched the retail 5..100 clamp, label table indexing, five-point key steps, and wrap behavior, then locked that draw/width/key-dispatch path with focused tests.
3. Corrected UI_NETSOURCE draw and ownerdraw-width bounds to reject index == numNetSources, and replaced the stale width check against uiInfo.numJoinGameTypes with the retail browser-source count.
4. Corrected UI_NETFILTER draw and ownerdraw-width bounds to reject index == numServerFilters, matching the retail seven-filter table limit.
5. Locked the Mplayer-skip source cycling, global-refresh gate, server-list rebuilds, non-returning ownerdraw key dispatch, and filter wrap behavior with focused regression coverage.

Verification:

• python -m pytest tests/test_ui_menu_files.py -q --tb=short
  • Result: 72 passed.

Task A122: Close crosshair, next-map, and selected-player UI ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/ui/ui_main.c, tests/test_ui_menu_files.py Parity estimate: before 93% -> after 99% for the focused UI_CROSSHAIR, UI_NEXTMAP, and UI_SELECTEDPLAYER ownerdraw lane. Repo-wide parity remains 98% because unrelated compatibility and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail uix86.dll HLIL/Ghidra evidence for the crosshair preview, next-map label, selected-player draw, player-list refresh, and selected-player key-handler paths.
2. Reconstructed UI_CROSSHAIR preview placement to use the retail 24..40 size clamp and fixed rect->x - 2, rect->y - rect->h + 2 anchor instead of scaling/centering within the ownerdraw rect.
3. Reconstructed UI_NEXTMAP to paint only the retail configstring 0x29A payload and removed the compatibility fallback through rotation-title configstrings from that ownerdraw.
4. Reconstructed UI_SELECTEDPLAYER drawing to use the retail buffered cvar read for either cg_selectedPlayerName or name, while preserving the player-list refresh and selected-player key cycling semantics.
5. Locked draw dispatch, key dispatch, cvar reads/writes, configstring reads, selected-player cache refresh, wrap behavior, and removed fallback paths with focused regression coverage.

Verification:

• python -m pytest tests/test_ui_menu_files.py -q --tb=short
  • Result: 71 passed.

Task A121: Close retail gametype-selector UI ownerdraw parity [COMPLETED]

Priority: High Primary areas: src/code/ui/ui_main.c, tests/test_ui_menu_files.py Parity estimate: before 95% -> after 99% for the focused UI_GAMETYPE, UI_NETGAMETYPE, and UI_JOINGAMETYPE ownerdraw lane. Repo-wide parity remains 98% because unrelated compatibility and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail uix86.dll HLIL/Ghidra evidence for the gametype ownerdraw dispatch cases and their key handlers at 0x1000a110, 0x1000a210, and 0x1000a300.
2. Removed the old non-retail ui_Q3Model side effect from the UI_GAMETYPE key path while preserving retail cycling, cap/frag refresh, best-score reload, and map-feeder reset wiring.
3. Reconstructed the retail UI_NETGAMETYPE hidden-enum skip so the selector cycles past One Flag CTF, Overload, and Harvester before updating ui_netGameType, ui_actualnetGameType, ui_currentNetMap, and the map feeder.
4. Locked the UI_GAMETYPE, UI_NETGAMETYPE, and UI_JOINGAMETYPE draw-dispatch, key-dispatch, cvar, feeder, and server-list wiring with focused regression coverage.

Verification:

• python -m pytest tests/test_ui_menu_files.py -q --tb=short
  • Result: 70 passed.

Task A120: Close three retail UI ownerdraw parity seams [COMPLETED]

Priority: High Primary areas: src/code/ui/ui_main.c, tests/test_ui_menu_files.py Parity estimate: before 94% -> after 99% for the focused UI_VOTESTRING, UI_STARTING_WEAPONS, and UI_ADVERT ownerdraw lane. Repo-wide parity remains 98% because the remaining non-ownerdraw compatibility and packaging surfaces are unchanged.

Completed work:

1. Rechecked the retail uix86.dll HLIL/Ghidra ownerdraw dispatch evidence for UI_VOTESTRING, UI_STARTING_WEAPONS, and UI_ADVERT.
2. Reconstructed UI_VOTESTRING to use the retail buffered cvar read and integer x-anchor centering behavior instead of centering inside the rect width.
3. Reconstructed UI_STARTING_WEAPONS to parse the loadout mask through the retail decimal atoi path and to draw the loadout + marker at the retail x-offset before the queued-primary icon.
4. Locked the advert ownerdraw parse/setup/refresh/paint/activate bridge with focused regression coverage, including default content, cell id, hidden-menu refresh rectangles, shader assignment, draw/update ordering, and activation script forwarding.

Verification:

• python -m pytest tests/test_ui_menu_files.py -q --tb=short
  • Result: 69 passed.

Task A119: Reconstruct retail ownership for ui_* cvars [COMPLETED]

Priority: High Primary areas: src/code/ui/ui_main.c, src/code/ui/ui_local.h, src/code/client/cl_ui.c, tests/test_ui_menu_files.py, tests/test_cgame_hud_parity.py, tests/test_platform_services.py, docs/reverse-engineering/ui-cvar-mapping-round-2026-05-28.md Parity estimate: before 93% -> after 98% for the focused ui_* cvar callback and ownership lane. Repo-wide parity remains 98% because the source still carries bounded GPL-era and compatibility cvar rows outside the retail uix86.dll table.

Completed work:

1. Rechecked uix86.dll Ghidra and HLIL evidence for the retail UI cvar table, including the 0x82 row count, row field order, UI_RegisterCvars, UI_UpdateCvars, and the callback owners at sub_10011240, sub_10011630, sub_10011660, and sub_10011690.
2. Reconstructed the source cvar table callback lane with a retail-shaped row order of vmCvar, cvar name, default string, update callback, and flags.
3. Replaced ad hoc color modification-count checks with the retail UI_UpdateCvars callback rule and added the missing bulk ui_teamColor and ui_enemyColor rows.
4. Routed retail slider-color, force-model, and cg_announcer rows through source-visible callbacks that mirror the committed HLIL behavior.
5. Corrected the native UI S_RegisterSound host shim so the retail announcer callback's non-zero 7 argument remains true after qboolean normalization.
6. Documented the full retail 0x82 row table, host-owned ui_* bridge cvars, and compatibility-only source cvars, then refreshed focused parity tests.

Task A118: Reconstruct cached retail ownership for web_* cvars [COMPLETED]

Priority: High Primary areas: src/code/qcommon/common.c, src/code/client/cl_cgame.c, tests/test_engine_cvar_retail_parity.py, tests/test_engine_client_command_parity.py, tests/test_awesomium_browser_parity.py, tests/test_platform_services.py, tools/ci/verify-awesomium-browser-host-parity.ps1, docs/reverse-engineering/awesomium-browser-wiring.md, docs/reverse-engineering/quakelive_steam_mapping_round_333.md Parity estimate: before 96% -> after 99% for the focused web_* cvar ownership lane. Overall Awesomium WebUI wiring remains 99.1% -> 99.1%; repo-wide parity remains 98%.

Completed work:

1. Rechecked the committed Ghidra and HLIL corpus for all retail web_* cvar names and confirmed the supported set is web_zoom, web_console, and web_browserActive; the other web_* names in this subsystem are command registrations.
2. Reconstructed the core cached web_browserActive cvar pointer in common.c and routed the frame idle-sleep browser-active test through that pointer instead of a repeated name lookup.
3. Reconstructed cached browser-host cvar pointers for web_zoom, web_console, and web_browserActive, added a source-visible mapping table with recovered retail globals/defaults/flags, and routed console-message and live zoom consumers through the cached cvars.
4. Matched the retail web_zoom modified-latch consumption path by clearing the cvar latch after the live Awesomium view consumes a zoom update.
5. Refreshed focused tests, verifier anchors, and mapping notes for the new cvar-owner evidence.

Task A117: Stop no-op Awesomium bridge cvar churn and guard Awesomium launches [COMPLETED]

Priority: High Primary areas: src/code/client/cl_cgame.c, .vscode/build.ps1, .vscode/launch.ps1, tests/test_awesomium_browser_parity.py, tests/test_platform_services.py, tools/ci/verify-awesomium-browser-host-parity.ps1, docs/reverse-engineering/awesomium-browser-wiring.md Parity estimate: before 99.0% -> after 99.1% for the focused Awesomium WebUI wiring lane. Repo-wide parity remains 98% because default online-service exclusion is policy-bounded.

Completed work:

1. Rechecked the Ghidra/HLIL evidence for Cvar_Set2 and confirmed that the retail debug string is emitted before the unchanged-value return path, so the core cvar setter must not be softened to hide compatibility publisher spam.
2. Added a client-side CL_SetCvarIfChanged() guard and routed repeated browser/advert bridge telemetry plus frame-loop web_browserActive mirrors through it, while keeping retail-style open/hide cvar writes direct.
3. Added a build-settings stamp to the VS Code build wrapper and made launch.ps1 -EnableAwesomium refuse a launch when the last build used QLBuildOnlineServices=0, making accidental default-off WebUI tests fail early with a clear message.
4. Refreshed focused tests, verifier anchors, and the Awesomium wiring notes so the no-op cvar publish behavior and explicit Awesomium build path stay pinned.

Task A116: Remove Awesomium SDK replication and require external SDK dependencies [COMPLETED]

Priority: High Primary areas: src/code/client/cl_awesomium_win32.cpp, src/code/awesomium_process.vcxproj, src/code/quakelive_steam.vcxproj, src/code/win32/awesomium_process.cpp, src/code/win32/awesomium_process.rc, tests/test_awesomium_browser_parity.py, tests/test_platform_services.py, tools/ci/verify-awesomium-browser-host-parity.ps1, tools/ci/verify-awesomium-process-parity.ps1, docs/reverse-engineering/awesomium-browser-wiring.md, docs/reverse-engineering/awesomium_process-mapping.md, docs/reverse-engineering/quakelive_steam_mapping_round_332.md Parity estimate: before 64% -> after 94% for the focused Awesomium SDK-clean integration lane. Overall Awesomium WebUI wiring remains estimated at 99.0% -> 99.0%; repo-wide parity remains 98%.

Completed work:

1. Audited the Awesomium client/backend/helper build surface and found local SDK-like replication in the C++ ABI fallback layer, generated .def import library, local ChildProcessMain declaration, and vendor-owned helper metadata.
2. Removed local Awesomium object storage, __thiscall thunks, decorated C++ import fallbacks, vtable-dispatch fallbacks, and raw BitmapSurface field offset reads from the live backend.
3. Corrected live SDK export usage for transparent-state setup and keyboard event injection, using the SDK-owned WebKeyboardEvent object path.
4. Deleted src/code/win32/awesomium.def, required external AwesomiumSdkDir/AWESOMIUM_SDK_DIR for online Awesomium builds, linked awesomium.lib from that SDK for the helper, and declared the external runtime dependency for the launcher.
5. Updated documentation, mapping notes, pytest coverage, and verifier guards so SDK replication does not re-enter silently.

Verification:

• python -m pytest tests/test_awesomium_browser_parity.py tests/test_platform_services.py -q --tb=short
  • Result: 103 passed.
• powershell -ExecutionPolicy Bypass -File tools\ci\verify-awesomium-browser-host-parity.ps1 -RepoRoot E:\Repositories\QuakeLive-reverse
  • Result: all Awesomium browser host parity and SDK-clean dependency anchors present.
• git diff --check -- ...
  • Result: clean; only Git line-ending normalization warnings were reported.
• powershell -ExecutionPolicy Bypass -File tools\ci\build-windows-dlls.ps1 -RepoRoot E:\Repositories\QuakeLive-reverse -Solution src/code/quakelive.sln -Configuration Debug -Platform Win32 -PlatformToolset v143 -DisableOptionalCodecs
  • Result: build succeeded with 29 existing warnings and 0 errors.

Task A115: Reconstruct Awesomium qz method table return flags [COMPLETED]

Priority: High Primary areas: src/code/client/cl_cgame.c, src/code/client/cl_awesomium_win32.cpp, tests/test_awesomium_browser_parity.py, tests/test_platform_services.py, tools/ci/verify-awesomium-browser-host-parity.ps1, docs/reverse-engineering/awesomium-browser-wiring.md, docs/reverse-engineering/quakelive_steam_mapping_round_331.md Parity estimate: before 97% -> after 99.4% for the focused Awesomium qz method-table/return-flag lane. Overall Awesomium WebUI wiring is estimated at 98.9% -> 99.0%; repo-wide parity remains 98%.

Completed work:

1. Rechecked QLJSHandler_LookupMethodId, QLJSHandler_BindQzInstance, QLJSHandler_OnMethodCall, and QLJSHandler_OnMethodCallWithReturnValue against HLIL with Ghidra function/vtable ownership support.
2. Added source-visible retail bounds and row addresses for the data_55c008 qz method table.
3. Corrected SetCvar and ResetCvar to follow the retail return-valued method lane and return boolean success/failure values.
4. Updated the injected startup qz_instance bridge so local cvar set/reset helpers normalize cvar names and return truthy values.
5. Added parity coverage and CI verifier anchors for the qz method table, return flags, startup bridge behavior, and mapping-round evidence.

Verification:

• python -m pytest tests/test_awesomium_browser_parity.py tests/test_platform_services.py -q --tb=short
  • Result: 102 passed.
• powershell -ExecutionPolicy Bypass -File tools\ci\verify-awesomium-browser-host-parity.ps1 -RepoRoot E:\Repositories\QuakeLive-reverse
  • Result: all Awesomium browser host parity anchors present.
• git diff --check -- src/code/client/cl_awesomium_win32.cpp src/code/client/cl_cgame.c src/code/client/client.h tests/test_awesomium_browser_parity.py tests/test_platform_services.py tools/ci/verify-awesomium-browser-host-parity.ps1 docs/reverse-engineering/awesomium-browser-wiring.md docs/reverse-engineering/quakelive_steam_mapping_round_330.md docs/reverse-engineering/quakelive_steam_mapping_round_331.md IMPLEMENTATION_PLAN.md
  • Result: clean; only Git line-ending normalization warnings were reported.
• powershell -ExecutionPolicy Bypass -File tools\ci\build-windows-dlls.ps1 -RepoRoot E:\Repositories\QuakeLive-reverse -Solution src/code/quakelive.sln -Configuration Debug -Platform Win32 -PlatformToolset v143 -DisableOptionalCodecs
  • Result: build succeeded with 0 warnings and 0 errors.

Task A114: Reconstruct Awesomium WebSession lifecycle/cache wiring [COMPLETED]

Priority: High Primary areas: src/code/client/cl_awesomium_win32.cpp, src/code/client/cl_cgame.c, src/code/client/client.h, tests/test_awesomium_browser_parity.py, tools/ci/verify-awesomium-browser-host-parity.ps1, docs/reverse-engineering/awesomium-browser-wiring.md, docs/reverse-engineering/quakelive_steam_mapping_round_330.md Parity estimate: before 96% -> after 99% for the focused Awesomium WebSession lifecycle/cache lane. Overall Awesomium WebUI wiring is estimated at 98.4% -> 98.9%; repo-wide parity remains 98%.

Completed work:

1. Rechecked sub_4F2A10, sub_4F2A30, sub_4F2A60, sub_4F2D30, sub_4F3CB0, and sub_4F3CD0 against HLIL with Ghidra metadata and symbol-alias support.
2. Reconstructed WebSession slot 0x18 initialization immediately after WebCore::CreateWebSession.
3. Reclassified WebSession slot 0x1c as cache clear for web_clearCache and reload, removed it from shutdown, and routed live clear-cache through CL_Awesomium_ClearCache.
4. Matched retail command-tail behavior for web_showError and made live navigation failures enter the mapped load-failure handler.
5. Updated Awesomium wiring notes, mapping-round evidence, pytest coverage, and the browser-host parity verifier.

Verification:

• python -m pytest tests/test_awesomium_browser_parity.py tests/test_platform_services.py -q --tb=short
  • Result: 101 passed.
• powershell -ExecutionPolicy Bypass -File tools\ci\verify-awesomium-browser-host-parity.ps1 -RepoRoot E:\Repositories\QuakeLive-reverse
  • Result: all Awesomium browser host parity anchors present.
• git diff --check -- src/code/client/cl_awesomium_win32.cpp src/code/client/cl_cgame.c src/code/client/client.h docs/reverse-engineering/awesomium-browser-wiring.md docs/reverse-engineering/quakelive_steam_mapping_round_330.md tests/test_awesomium_browser_parity.py tools/ci/verify-awesomium-browser-host-parity.ps1
  • Result: clean; only Git line-ending normalization warnings were reported.
• powershell -ExecutionPolicy Bypass -File tools\ci\build-windows-dlls.ps1 -RepoRoot E:\Repositories\QuakeLive-reverse -Solution src/code/quakelive.sln -Configuration Debug -Platform Win32 -PlatformToolset v143 -DisableOptionalCodecs
  • Result: build succeeded; existing UI/game config warnings were reported.

Task A113: Reconstruct application initialization host wiring [COMPLETED]

Priority: High Primary areas: src/code/win32/win_main.c, references/analysis/quakelive_symbol_aliases.json, tests/test_application_initialization_mapping.py, docs/reverse-engineering/application-initialization-wiring-reconstruction-2026-05-27.md Parity estimate: before 96.5% -> after 98.0% for the scoped application-initialization wiring lane. The repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked WinMain @ 0x004ED830, Com_Init @ 0x004CBFD0, Sys_Init @ 0x004ED400, NET_Init @ 0x004EF320, CL_Init @ 0x004BC690, SV_Init @ 0x004E3AD0, and the adjacent ZMQ/browser startup handoffs against Binary Ninja HLIL with Ghidra metadata support.
2. Promoted sub_4EC580 to Sys_WinkeyHookProc and reconstructed the WinMain-owned low-level keyboard hook that gates VK_LWIN / VK_RWIN behind winkey_disable.
3. Added shutdown unhook wiring to both Sys_Error and Sys_Quit, matching the retail unhook sites in the error and normal-exit paths.
4. Documented policy-aware startup divergences for Steam, ZMQ, and Awesomium so Quake Live-only online services remain behind the repository's disabled defaults.
5. Added static parity coverage for the retail owner map, host call order, hook predicate, shutdown unhook, and policy-adjusted common/client/server startup wiring.

Verification:

• python -m pytest tests/test_application_initialization_mapping.py -q --tb=short
  • Result: 3 passed.
• python -m pytest tests/test_renderer_win32_host_glue_parity.py tests/test_win32_raw_input_parity.py tests/test_engine_client_command_parity.py -q --tb=short
  • Result: 46 passed.
• python -m pytest tests/test_platform_services.py::test_client_steam_callback_owner_reconstructs_retail_frame_pump_and_lifecycle tests/test_platform_services.py::test_client_browser_host_core_reconstructs_retained_runtime_owner tests/test_platform_services.py::test_server_zmq_runtime_reconstructs_retail_publication_and_rcon_owners -q --tb=short
  • Result: 3 passed.
• python -m pytest tests/test_awesomium_browser_parity.py::test_awesomium_runtime_bootstrap_and_surface_pump_reconstruct_retail_host_contract tests/test_awesomium_browser_parity.py::test_awesomium_view_callbacks_reconstruct_tooltip_and_console_contracts -q --tb=short
  • Result: 2 passed.
• MSBuild.exe src\code\quakelive_steam.vcxproj /p:Configuration=Debug /p:Platform=Win32 /m /v:minimal
  • Result: succeeded; existing cl_cgame.c unused-local warnings were reported.
• git diff --check
  • Result: clean; only Git line-ending normalization warnings were reported.

Task A112: Retire generated UI bridge menu assets [COMPLETED]

Priority: High Primary areas: src/code/ui/ui_main.c, src/code/ui/ui_atoms.c, src/code/ui/ui_cdkey.c, src/code/ui/ui_local.h, src/code/ui/ui.vcxproj, src/code/ui/Conscript, tests/test_ui_menu_files.py, tests/test_platform_services.py, docs/ui/scripting-guide.md Parity estimate: before 99.90% -> after 99.96% for the scoped retail UI menu-root purity lane. The strict-retail Windows module target remains 100% and the repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked retail uix86.dll evidence: HLIL/Ghidra both show the default menu root opening ui/menus.txt, and the checked-in src/ui/menus.txt loads retail ui/main.menu.
2. Removed the generated ql_bridge_* menu-script owner and its build entries so offline browser-service handling no longer writes replacement menu assets into the homepath.
3. Kept disabled browser verbs policy-compliant by swallowing unavailable web_showBrowser / web_changeHash paths while the existing retail menu tree remains active.
4. Reframed the native CD-key command as an explicit unavailable-command stub in native UI builds instead of preferring generated credentials menus.
5. Updated parity tests, platform-service tests, docs, and CI font/build references to assert that retail menu files are used directly.

Verification:

• python -m pytest tests/test_ui_menu_files.py tests/test_platform_services.py tests/test_non_windows_portability.py tests/test_ui_full_parity_gate.py -q --tb=short
  • Result: 165 passed, 1 skipped.

Task A111: Reconstruct spectator client-state resync wiring [COMPLETED]

Priority: High Primary areas: src/code/game/bg_public.h, src/code/game/g_cmds.c, src/code/game/g_client.c, src/code/game/g_active.c, src/code/cgame/cg_playerstate.c, src/code/cgame/cg_local.h, tests/test_spectator_client_state_wiring_parity.py, docs/reverse-engineering/spectator-client-state-wiring-reconstruction-2026-05-27.md, docs/reverse-engineering/qagame-mapping.md, docs/reverse-engineering/cgame-mapping.md, references/symbol-maps/qagame.json, references/symbol-maps/cgame.json Parity estimate: before 99.90% -> after 99.97% for the scoped spectator client-state and item-state resync lane. The repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked qagame G_ApplyTeamChange @ 0x10040440 and ClientSpawn @ 0x1003BC30 against Binary Ninja HLIL/Ghidra for the retail 0x4000 spectator client-state eFlag writes.
2. Added EF_SPECTATOR_RESPAWN as the Quake Live alias for that network bit, intentionally sharing the GPL EF_VOTED value because retail uses the same bit as the cgame specresp cue.
3. Reconstructed qagame set/clear wiring in G_ApplyTeamChange and ClientSpawn, and preserved the observer bit when SpectatorClientEndFrame copies followed-player snapshots.
4. Replaced the raw cgame 0x00004000 CG_Respawn check with EF_SPECTATOR_RESPAWN, keeping the retail specresp command handoff.
5. Mapped the surrounding spectator item-state responder: specresp and spectator team entry call G_SyncSpectatorItemStatesForClient, while item pickup broadcast still packs EV_ITEM_PICKUP_SPEC for the cgame cache.
6. Added cross-binary static coverage plus qagame/cgame mapping and symbol-map notes for the eFlag, command, event payload, cache decoder, and overlay update/draw path.

Verification:

• python -m pytest tests/test_spectator_client_state_wiring_parity.py -q --tb=short
  • Result: 8 passed.
• python -m pytest tests/test_spectator_client_state_wiring_parity.py tests/test_game_spectator_connection_parity.py tests/test_game_active_pmove_wiring_parity.py tests/test_game_helper_seam_parity.py tests/test_cgame_spectator_parity.py tests/test_cgame_snapshot_parity.py tests/test_cgame_event_transport_parity.py tests/test_cgame_hud_parity.py::test_spectator_item_timer_text_uses_retail_default_font_lane tests/test_game_vote_clear_parity.py -q --tb=short
  • Result: 111 passed, 1 skipped.
• git diff --check
  • Result: clean; only Git line-ending normalization warnings were reported.
• Broader context: the full tests/test_cgame_hud_parity.py file still has an unrelated test_team_info_overlay_restores_fixed_retail_row_renderer assertion failure outside this spectator client-state lane.

Task A110: Reconstruct StepSlideMove down-clip and wiring map [COMPLETED]

Priority: High Primary areas: src/code/game/bg_slidemove.c, src/code/game/bg_pmove.c, src/code/game/g_pmove.c, src/code/cgame/cg_servercmds.c, src/code/cgame/cg_predict.c, tests/pmove_validation_harness.c, tests/test_pmove_validation_fixtures.py, tests/test_step_jump_gate_parity.py, docs/reverse-engineering/pmove-stepslidemove-wiring-reconstruction-2026-05-27.md Parity estimate: before 98.8% -> after 99.9% for the scoped StepSlideMove lane. The repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked qagame PM_StepSlideMove @ 0x1002EFE0 and cgame PM_StepSlideMove @ 0x100034B0 against Binary Ninja HLIL, with Ghidra companion confirmation for the matching helper call graph.
2. Tightened the down-trace clip guard to the retail dot-product epsilon: DotProduct(velocity, plane.normal) < 0.0f || fabs(dot) < 0.001f.
3. Preserved the retail direct reachability trace before recording step side effects, plus the startsolid rejection on the projected step-jump support probe.
4. Documented the full movement call matrix, step-jump latch split, and cgame prediction smoothing handoff.
5. Added executable and structural parity coverage for the down-clip guard, call-site wiring, normal/crouch takeoff latches, and pmove_t::stepUp prediction export.
6. Pinned the pmove_AirStepFriction, pmove_AirSteps, and pmove_StepHeight server-to-cgame tuning transport that feeds the private step-move globals.

Verification:

• python -m pytest tests/test_step_jump_gate_parity.py tests/test_pmove_validation_fixtures.py -q --tb=short
  • Result: 22 passed.
• python -m pytest tests/test_step_jump_gate_parity.py tests/test_pmove_helper_parity.py tests/test_pmove_validation_fixtures.py tests/test_pmove_movement_fixtures.py tests/test_pmove_selected_cvar_parity.py tools/tests/test_pmove_settings_configstring.py -q --tb=short
  • Result: 134 passed, 107 subtests passed.
• git diff --check
  • Result: clean; only Git line-ending normalization warnings were reported.

Task A109: Reconstruct CA/A-D spectator respawn handoff [COMPLETED]

Priority: High Primary areas: src/code/game/g_team.c, src/code/game/g_client.c, tests/test_game_attack_defend_parity.py, tests/test_game_spectator_connection_parity.py, tests/test_game_active_pmove_wiring_parity.py, docs/reverse-engineering/qagame-spectator-state-movement-reconstruction-2026-05-26.md, docs/reverse-engineering/qagame-client-spawn-reconstruction-2026-05-26.md, docs/reverse-engineering/qagame-mapping.md, references/symbol-maps/qagame.json Parity estimate: before 99.8% -> after 99.9% for the scoped qagame spectator state/movement lane. The repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked G_CAADRespawnAsSpectator @ 0x10035960 and the dependent ClientSpawn @ 0x1003BC30 spectator branch against Binary Ninja HLIL.
2. Collapsed the CA/A-D eliminated-player spectator helper back to the retail sequence: CopyToBodyQue, pre-seed PM_SPECTATOR, ClientSpawn, active team counts, and FollowCycle only when both teams still have players.
3. Removed the source-only direct follow-target picker, post-spawn relink, and free/scoreboard fallback assignment from that helper.
4. Taught ClientSpawn to preserve a pre-existing PM_SPECTATOR state as the spectator-spawn selector, so active-team CA/A-D eliminations can reuse the retail spectator placement path without pretending the session team changed.
5. Updated attack/defend, spectator, and active-pmove static parity coverage, plus the qagame mapping, client-spawn note, spectator note, and symbol map.

Verification:

• python -m pytest tests/test_game_attack_defend_parity.py tests/test_game_spectator_connection_parity.py -q --tb=short
  • Result: 10 passed.
• python -m pytest tests/test_game_active_pmove_wiring_parity.py::test_first_18_g_cvars_match_retail_defaults_flags_and_wiring tests/test_game_active_pmove_wiring_parity.py::test_spectator_impacts_and_predictable_event_sidecars_match_retail_wiring -q --tb=short
  • Result: 2 passed.
• python -m pytest tests/test_game_attack_defend_parity.py tests/test_game_spectator_connection_parity.py tests/test_game_active_pmove_wiring_parity.py tests/test_game_helper_seam_parity.py tests/test_pmove_helper_parity.py tests/test_game_factory_regen_parity.py tests/test_bg_misc_validation_fixtures.py tests/test_spawn_spec_cvars.py tests/test_cgame_spectator_parity.py -q --tb=short
  • Result: 164 passed.
• git diff --check
  • Result: clean; only Git line-ending normalization warnings were reported.

Task A108: Reconstruct initial and team-begin spawn fallback modes [COMPLETED]

Priority: High Primary areas: src/code/game/g_client.c, src/code/game/g_team.c, tests/test_game_active_pmove_wiring_parity.py, tests/test_game_helper_seam_parity.py, docs/reverse-engineering/qagame-client-spawn-reconstruction-2026-05-26.md, docs/reverse-engineering/qagame-mapping.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil.txt, references/reverse-engineering/ghidra/qagamex86/decompile_top_functions.c Parity estimate: before 98.8% -> after 99.0% for the scoped qagame client-spawn band. The repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked G_SelectRankedSpawnPoint @ 0x10039080 and G_SelectClientSpawnPoint @ 0x10039730 against Binary Ninja HLIL and the committed Ghidra companion body.
2. Reconstructed the neutral initial-spawn admission mode: ranked neutral initial selection now requires spawnflags bit 1 while still excluding bit 2.
3. Added a ranked initial-spawn wrapper so SelectInitialSpawnPoint ranks all eligible initial info_player_deathmatch starts instead of taking the first matching entity from G_Find.
4. Restored the team TEAM_BEGIN fallback retry from team_CTF_*player to the same team's team_CTF_*spawn class before neutral deathmatch fallback.
5. Updated helper/active parity coverage and the qagame client-spawn mapping notes, leaving the training-map-specific ranked-spawn path as the next explicit follow-up.

Verification:

• python -m pytest tests/test_game_helper_seam_parity.py tests/test_game_active_pmove_wiring_parity.py -q --tb=short
  • Result: 63 passed.
• python -m pytest tests/test_game_spectator_connection_parity.py tests/test_game_active_pmove_wiring_parity.py tests/test_game_helper_seam_parity.py tests/test_pmove_helper_parity.py tests/test_game_factory_regen_parity.py tests/test_bg_misc_validation_fixtures.py tests/test_spawn_spec_cvars.py tests/test_cgame_spectator_parity.py -q --tb=short
  • Result: 161 passed.
• git diff --check -- src/code/game/g_client.c src/code/game/g_team.c tests/test_game_helper_seam_parity.py tests/test_game_active_pmove_wiring_parity.py docs/reverse-engineering/qagame-client-spawn-reconstruction-2026-05-26.md docs/reverse-engineering/qagame-mapping.md
  • Result: pass; only repository LF-to-CRLF conversion warnings.

Task A107: Correct spectator follow command surface [COMPLETED]

Priority: High Primary areas: src/code/game/g_cmds.c, tests/test_game_spectator_connection_parity.py, docs/reverse-engineering/qagame-spectator-state-movement-reconstruction-2026-05-26.md, docs/reverse-engineering/qagame-spectator-follow-reconstruction-2026-05-22.md, docs/reverse-engineering/qagame-mapping.md, references/symbol-maps/qagame.json Parity estimate: before 99.5% -> after 99.8% for the scoped qagame spectator command surface. The repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked Cmd_Follow_f @ 0x10040F30 and Cmd_FollowCycle_f @ 0x100412E0 against the committed Binary Ninja HLIL after the broader spectator-state pass.
2. Corrected Cmd_Follow_f so follow1 and follow2 remain owned by SetTeam, while the follow command resolves player strings through ClientNumberFromString.
3. Restored the retail cgame "pw" suffix guard that preserves the current followed flag carrier via BG_PlayerCarryingFlag, and changed explicit target rejection to the retail ps.pm_type == PM_SPECTATOR check.
4. Removed source-only training-map print gates from Cmd_Follow_f and Cmd_FollowCycle_f; no-arg follow cleanup remains handled by the retail follow/end-frame paths.
5. Rebuilt the public follow-cycle entry around the retail sess.sessionTeam spectator handoff before tailcalling the inner FollowCycle worker.
6. Updated static parity coverage, the spectator reconstruction notes, the qagame mapping ledger, and the qagame symbol-map comments.

Verification:

• python -m pytest tests/test_game_spectator_connection_parity.py tests/test_spawn_spec_cvars.py tests/test_cgame_spectator_parity.py -q --tb=short
  • Result: 18 passed.
• python -m pytest tests/test_game_active_pmove_wiring_parity.py::test_spectator_impacts_and_predictable_event_sidecars_match_retail_wiring tests/test_pmove_helper_parity.py -q --tb=short
  • Result: 41 passed.
• python -m pytest tests/test_game_helper_seam_parity.py::test_last_alive_alert_helpers_match_recovered_retail_boundaries -q --tb=short
  • Result: 1 passed.
• Wider adjacent suite note: python -m pytest tests/test_game_spectator_connection_parity.py tests/test_game_active_pmove_wiring_parity.py tests/test_game_helper_seam_parity.py tests/test_pmove_helper_parity.py tests/test_game_factory_regen_parity.py tests/test_bg_misc_validation_fixtures.py tests/test_spawn_spec_cvars.py tests/test_cgame_spectator_parity.py -q --tb=short currently reports 159 passed, 2 failed; both failures are in unrelated ranked-spawn assertions in tests/test_game_active_pmove_wiring_parity.py and tests/test_game_helper_seam_parity.py.
• git diff --check
  • Result: pass; only repository LF-to-CRLF conversion warnings.

Task A106: Reconstruct ranked-spawn spawnflag exclusion [COMPLETED]

Priority: High Primary areas: src/code/game/g_client.c, tests/test_game_active_pmove_wiring_parity.py, tests/test_game_helper_seam_parity.py, docs/reverse-engineering/qagame-client-spawn-reconstruction-2026-05-26.md, docs/reverse-engineering/qagame-mapping.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil.txt, references/reverse-engineering/ghidra/qagamex86/decompile_top_functions.c Parity estimate: before 98.7% -> after 98.8% for the scoped qagame client-spawn band. The repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked G_SelectRankedSpawnPoint @ 0x10039080 against Binary Ninja HLIL and the committed Ghidra companion body.
2. Identified the retail candidate admission guard that reads spawnflags at entity + 0x248 and rejects candidates with bit 2 set before distance scoring.
3. Added RANKED_SPAWN_EXCLUDE_FLAG and G_RankedSpawnPointAllowed, then routed both the ranked-scoring path and the no-ranked-candidate fallback through the recovered filter.
4. Pinned the helper, constant, and ranked-picker call sites in the qagame helper and active pmove/spawn parity tests.
5. Updated the client-spawn reconstruction note and qagame mapping ledger with the HLIL/Ghidra evidence and the remaining bit-1 fallback question.

Verification:

• python -m pytest tests/test_game_helper_seam_parity.py tests/test_game_active_pmove_wiring_parity.py -q --tb=short
  • Result: 63 passed.
• git diff --check -- src/code/game/g_client.c src/code/game/g_local.h tests/test_game_helper_seam_parity.py tests/test_game_active_pmove_wiring_parity.py docs/reverse-engineering/qagame-client-spawn-reconstruction-2026-05-26.md docs/reverse-engineering/qagame-mapping.md IMPLEMENTATION_PLAN.md
  • Result: pass; only repository LF-to-CRLF conversion warnings.

Task A105: Split renderer framebuffer and shader procedure gates [COMPLETED]

Priority: High Primary areas: src/code/renderer/tr_backend.c, tests/test_renderer_internal_helper_mapping_parity.py, docs/reverse-engineering/quakelive_steam_mapping_round_328.md, docs/reverse-engineering/renderer-full-parity-audit-and-implementation-plan-2026-04-09.md, docs/reverse-engineering/source-file-parity-ledger-2026-04-22.md, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/ Parity estimate: before 99.985% -> after 99.99% for the scoped post-process proc gate lane. The strict renderer estimate remains 100% and the repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked retail sub_4500B0 and sub_450640 against HLIL to separate framebuffer target prerequisites from post-effect shader/program prerequisites.
2. Added RBPP_LoadFramebufferProcs for the retail framebuffer-only target lane: FBO/renderbuffer/rectangle entry points plus max rectangle texture size.
3. Added independent framebuffer procedure loaded/supported state to ppState_t.
4. Rewired RBPP_CreateRenderTarget to call the framebuffer-only loader, while RBPP_LoadProgram remains behind the full shader/uniform RBPP_LoadProcs gate.
5. Updated the renderer mapping, audit, and source-file ledger; tr_backend.c now tracks 79 functions.

Verification:

• python -m pytest tests/test_renderer_post_process_parity.py tests/test_renderer_internal_helper_mapping_parity.py tests/test_renderer_export_tail_parity.py tests/test_renderer_full_parity_gate.py -q --tb=short
  • Result: 46 passed, 1 skipped.
• python -m pytest tests/test_engine_cvar_retail_parity.py::test_engine_cvar_fortysecond_renderer_postprocess_state_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_twentyninth_renderer_postprocess_extension_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_thirtysecond_renderer_bloom_picmip_tranche_matches_retail_contracts -q --tb=short
  • Result: 3 passed.
• python -m pytest tests/test_engine_client_command_parity.py::test_postprocess_restart_routes_through_renderer_export_not_renderer_cmd_registration -q --tb=short
  • Result: 1 passed.
• git diff --check
  • Result: pass; only repository LF-to-CRLF conversion warnings.

Task A104: Reconstruct renderer post-process GL error and link wiring [COMPLETED]

Priority: High Primary areas: src/code/renderer/tr_backend.c, src/code/renderer/tr_init.c, src/code/renderer/tr_local.h, tests/test_renderer_internal_helper_mapping_parity.py, tests/test_renderer_post_process_parity.py, docs/reverse-engineering/quakelive_steam_mapping_round_327.md, docs/reverse-engineering/renderer-full-parity-audit-and-implementation-plan-2026-04-09.md, docs/reverse-engineering/source-file-parity-ledger-2026-04-22.md, references/analysis/quakelive_symbol_aliases.json, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/, references/reverse-engineering/ghidra/quakelive_steam/functions.csv Parity estimate: before 99.97% -> after 99.985% for the scoped post-process GL error and link lane. The strict renderer estimate remains 100% and the repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked retail sub_447E40, sub_4500B0, sub_4505F0, and sub_450640 against HLIL and the committed Ghidra function inventory.
2. Restored GL_CheckErrors as a return-valued helper so post-process callers can branch on unignored GL errors while existing callers may ignore the result.
3. Rewired RBPP_CreateRenderTarget to use GL_CheckErrors() after rectangle texture allocation and after framebuffer completeness, and removed the source-only renderbuffer bind from the attachment path.
4. Promoted sub_4505F0 as RBPP_LinkProgram, moving post-effect program create/attach/link and GL-error-gated success into its own retail-shaped helper.
5. Dropped the source-only GL_OBJECT_LINK_STATUS_ARB link-status query from RBPP_LoadProgram and routed color-correct texture allocation through GL_CheckErrors().

Verification:

• python -m pytest tests/test_renderer_post_process_parity.py tests/test_renderer_internal_helper_mapping_parity.py tests/test_renderer_export_tail_parity.py tests/test_renderer_full_parity_gate.py -q --tb=short
  • Result: 45 passed, 1 skipped.
• python -m pytest tests/test_engine_cvar_retail_parity.py::test_engine_cvar_fortysecond_renderer_postprocess_state_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_twentyninth_renderer_postprocess_extension_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_thirtysecond_renderer_bloom_picmip_tranche_matches_retail_contracts -q --tb=short
  • Result: 3 passed.
• python -m pytest tests/test_engine_client_command_parity.py::test_postprocess_restart_routes_through_renderer_export_not_renderer_cmd_registration -q --tb=short
  • Result: 1 passed.
• git diff --check
  • Result: pass; only repository LF-to-CRLF conversion warnings.

Task A103: Tighten spectator follow state and POI divergence [COMPLETED]

Priority: High Primary areas: src/code/game/g_active.c, src/code/game/g_cmds.c, src/code/game/g_local.h, references/symbol-maps/qagame.json, docs/reverse-engineering/qagame-mapping.md, docs/reverse-engineering/qagame-spectator-state-movement-reconstruction-2026-05-26.md, tests/test_game_spectator_connection_parity.py, tests/test_game_active_pmove_wiring_parity.py, tests/test_game_helper_seam_parity.py Parity estimate: before 96% -> after 99.5% for scoped qagame spectator follow-state and end-frame parity; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked SpectatorClientEndFrame @ 0x10035470, StopFollowing @ 0x10040D10, Cmd_Follow_f @ 0x10040F30, and FollowCycle @ 0x10041130 against the committed Binary Ninja HLIL.
2. Removed the source-only POI spectator camera storage and direct admin commands; retail follow cycling is client-only.
3. Rebuilt StopFollowing around the retail SPECTATOR_FREE reset with PM_SPECTATOR, PMF_FOLLOW clearing, self spectatorClient/clientNum, and linked-entity unlinking.
4. Routed SpectatorThink and stale end-frame targets through the recovered inner FollowCycle worker, including active-team fallback cycling when both teams still have live players.
5. Tightened FollowCycle filters for disconnected, PM_SPECTATOR, PMF_FOLLOW, and cross-team targets.

Verification:

• python -m pytest tests/test_game_spectator_connection_parity.py tests/test_game_active_pmove_wiring_parity.py tests/test_game_helper_seam_parity.py tests/test_pmove_helper_parity.py tests/test_game_factory_regen_parity.py tests/test_bg_misc_validation_fixtures.py -q --tb=short
  • Result: 149 passed.
• python -m pytest tests/test_spawn_spec_cvars.py tests/test_cgame_spectator_parity.py -q --tb=short
  • Result: 11 passed.
• git diff --check
  • Result: pass; only repository LF-to-CRLF conversion warnings.

Task A102: Reconstruct renderer post-process renderbuffer cache [COMPLETED]

Priority: High Primary areas: src/code/renderer/tr_backend.c, tests/test_renderer_internal_helper_mapping_parity.py, docs/reverse-engineering/quakelive_steam_mapping_round_326.md, docs/reverse-engineering/renderer-full-parity-audit-and-implementation-plan-2026-04-09.md, docs/reverse-engineering/source-file-parity-ledger-2026-04-22.md, references/analysis/quakelive_symbol_aliases.json, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/, references/reverse-engineering/ghidra/quakelive_steam/functions.csv Parity estimate: before 99.95% -> after 99.97% for the scoped post-process depth-stencil renderbuffer cache lane. The strict renderer estimate remains 100% and the repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked retail sub_44FFD0, sub_4500B0, sub_450710, and sub_450780 against HLIL and the committed Ghidra function inventory.
2. Promoted sub_44FFD0 as RBPP_CreateDepthStencilRenderbuffer and added the retail eight-entry width/height renderbuffer cache to ppState_t.
3. Routed RBPP_CreateRenderTarget through the cached GL_DEPTH24_STENCIL8_EXT renderbuffer helper before rectangle-texture and framebuffer creation.
4. Reset the reconstructed renderbuffer cache metadata during bloom resource shutdown, matching the retail rebuild/shutdown cache reset lane.
5. Updated the renderer mapping, audit, source-file ledger, and symbol alias evidence; tr_backend.c now tracks 77 functions.

Verification:

• python -m pytest tests/test_renderer_post_process_parity.py tests/test_renderer_internal_helper_mapping_parity.py tests/test_renderer_export_tail_parity.py tests/test_renderer_full_parity_gate.py -q --tb=short
  • Result: 44 passed, 1 skipped.
• python -m pytest tests/test_engine_cvar_retail_parity.py::test_engine_cvar_fortysecond_renderer_postprocess_state_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_twentyninth_renderer_postprocess_extension_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_thirtysecond_renderer_bloom_picmip_tranche_matches_retail_contracts -q --tb=short
  • Result: 3 passed.
• python -m pytest tests/test_engine_client_command_parity.py::test_postprocess_restart_routes_through_renderer_export_not_renderer_cmd_registration -q --tb=short
  • Result: 1 passed.
• git diff --check
  • Result: pass; only repository LF-to-CRLF conversion warnings.

Task A101: Correct ranked spawn candidate window [COMPLETED]

Priority: High Primary areas: src/code/game/g_client.c, tests/test_game_active_pmove_wiring_parity.py, tests/test_game_helper_seam_parity.py, docs/reverse-engineering/qagame-client-spawn-reconstruction-2026-05-26.md, docs/reverse-engineering/qagame-mapping.md Parity estimate: before 98.5% -> after 98.7% for the scoped qagame client-spawn band. The repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked G_SelectRankedSpawnPoint @ 0x10039080 against HLIL and the Ghidra companion body.
2. Restored the retail 0x1a retained ranked-spawn candidate cap by changing MAX_RANKED_SPAWN_POINTS from 32 to 26.
3. Pinned the cap in the active pmove/spawn wiring and helper seam parity tests.
4. Updated the qagame client-spawn reconstruction note and mapping ledger with the ranked-picker evidence.

Verification:

• python -m pytest tests/test_game_helper_seam_parity.py tests/test_game_active_pmove_wiring_parity.py -q --tb=short
  • Result: 1 failed, 62 passed; failure is an unrelated pre-existing g_cmds.c spectator free-cam expectation in test_team_loadout_bot_and_drop_cvars_keep_retail_behavioral_wiring.
• python -m pytest tests/test_game_helper_seam_parity.py::test_client_spawn_uses_recovered_loadout_and_rr_helpers tests/test_game_active_pmove_wiring_parity.py::test_first_18_g_cvars_match_retail_defaults_flags_and_wiring -q --tb=short
  • Result: 2 passed.

Task A100: Reconstruct renderer bloom teardown and program cleanup ordering [COMPLETED]

Priority: High Primary areas: src/code/renderer/tr_backend.c, tests/test_renderer_post_process_parity.py, tests/test_renderer_internal_helper_mapping_parity.py, docs/reverse-engineering/quakelive_steam_mapping_round_325.md, docs/reverse-engineering/renderer-full-parity-audit-and-implementation-plan-2026-04-09.md, docs/reverse-engineering/source-file-parity-ledger-2026-04-22.md, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/, src2/ghidra/quakelive_steam/quakelive_steam_decomp.cpp Parity estimate: before 99.93% -> after 99.95% for the scoped post-process bloom teardown lane. The strict renderer estimate remains 100% and the repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked retail sub_437DA0 and sub_4506A0 against HLIL and the Ghidra companion corpus.
2. Updated RBPP_DestroyBloomPrograms to destroy bloom programs in retail memory order: brightpass, downsample, blurvertical, blurhoriz, combine.
3. Rebuilt RBPP_ShutdownBloomResources around the retail grouped delete order: all textures, then all framebuffers, then all depth-stencil renderbuffers, followed by target slot clearing.
4. Added the retail glDetachObjectARB path to RBPP_DestroyProgram so shader objects detach before program deletion.
5. Kept tr_backend.c at 76 tracked functions while tightening lifecycle parity in place.

Verification:

• python -m pytest tests/test_renderer_post_process_parity.py tests/test_renderer_internal_helper_mapping_parity.py tests/test_renderer_export_tail_parity.py tests/test_renderer_full_parity_gate.py -q --tb=short
  • Result: 43 passed, 1 skipped.
• python -m pytest tests/test_engine_cvar_retail_parity.py::test_engine_cvar_fortysecond_renderer_postprocess_state_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_twentyninth_renderer_postprocess_extension_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_thirtysecond_renderer_bloom_picmip_tranche_matches_retail_contracts -q --tb=short
  • Result: 3 passed.
• python -m pytest tests/test_engine_client_command_parity.py::test_postprocess_restart_routes_through_renderer_export_not_renderer_cmd_registration -q --tb=short
  • Result: 1 passed.
• git diff --check
  • Result: pass; only repository LF-to-CRLF conversion warnings.

Task A99: Restore client-spawn team gametype selection band [COMPLETED]

Priority: High Primary areas: src/code/game/g_client.c, tests/test_game_helper_seam_parity.py, docs/reverse-engineering/qagame-client-spawn-reconstruction-2026-05-26.md, docs/reverse-engineering/qagame-mapping.md Parity estimate: before 98% -> after 98.5% for the scoped qagame client-spawn band. The repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked G_SelectClientSpawnPoint @ 0x10039730 against the HLIL gametype - 4 <= 7 branch and the current source gametype enum.
2. Replaced the broad source g_gametype >= GT_CTF spawn-class shortcut with G_GametypeUsesTeamSpawnSelection.
3. Restored the retail edge cases: Clan Arena now uses team spawn classes, and Red Rover no longer routes through the CTF-style team spawn family before its Red Rover role/loadout finalizer.
4. Extended the client-spawn reconstruction note and qagame mapping ledger.

Verification:

• python -m pytest tests/test_game_helper_seam_parity.py tests/test_game_spectator_connection_parity.py tests/test_game_active_pmove_wiring_parity.py -q --tb=short
  • Result: 69 passed.

Task A98: Reconstruct spectator state and fly-move parity [COMPLETED]

Priority: High Primary areas: src/code/game/g_active.c, src/code/game/bg_pmove.c, src/code/game/g_items.c, src/code/game/g_main.c, src/code/game/bg_public.h, references/symbol-maps/qagame.json, docs/reverse-engineering/qagame-mapping.md, docs/reverse-engineering/qagame-spectator-state-movement-reconstruction-2026-05-26.md, tests/test_game_active_pmove_wiring_parity.py, tests/test_game_factory_regen_parity.py, tests/test_pmove_helper_parity.py Parity estimate: before 93% -> after 99% for scoped spectator state/movement and shared fly-move parity; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked qagame SpectatorThink at 0x10033E30, adjacent spectator helpers, qagame/cgame PM_FlyMove, and Pickup_Powerup against the committed HLIL and symbol maps.
2. Restored the retail spectator pmove ordering: unlink linked spectators before pmove, set PM_SPECTATOR with speed 480, run Pmove, copy the moved origin, and touch triggers afterward.
3. Restored the retail spectator input edge split: attack-edge follow cycling returns before the later BUTTON_ANY edge can drop follow state.
4. Collapsed shared PM_FlyMove back to the retail four-call path and removed source-only Flight fuel/thrust stat seeding from powerup pickup while keeping the retail cvar registrations.

Verification:

• python -m pytest tests/test_game_spectator_connection_parity.py tests/test_game_active_pmove_wiring_parity.py tests/test_pmove_helper_parity.py tests/test_game_factory_regen_parity.py tests/test_bg_misc_validation_fixtures.py -q --tb=short
  • Result: 118 passed.
• git diff --check
  • Result: pass; only repository LF-to-CRLF conversion warnings.

Task A97: Reconstruct client-spawn no-spawn retry path [COMPLETED]

Priority: High Primary areas: src/code/game/g_client.c, src/code/game/g_local.h, tests/test_game_helper_seam_parity.py, docs/reverse-engineering/qagame-client-spawn-reconstruction-2026-05-26.md, docs/reverse-engineering/qagame-mapping.md Parity estimate: before 96% -> after 98% for the scoped qagame client-spawn band. The repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked ClientSpawn @ 0x1003BC30, G_SelectClientSpawnPoint @ 0x10039730, G_SelectRankedSpawnPoint @ 0x10039080, Team_SelectDominationSpawnPoint @ 0x10038B60, G_InitClientSpawnState @ 0x1003B6C0, G_FinalizeSpawnLoadout @ 0x1003B5A0, and G_GiveItemByName @ 0x1003BB90 against HLIL, Ghidra, and the curated symbol map.
2. Restored the retail active-player no-spawn branch: ClientSpawn now defers for 600 ms, writes PM_SPECTATOR, increments the retry counter, and returns before the full client reset when the spawn wrapper cannot produce an eligible point.
3. Replaced the inherited unbounded spawnpoint eligibility loop with a bounded FL_NO_BOTS / FL_NO_HUMANS guard and a scheduled retry callback.
4. Added a dedicated reconstruction note and expanded qagame mapping coverage for the client-spawn helper family.

Verification:

• python -m pytest tests/test_game_helper_seam_parity.py tests/test_game_spectator_connection_parity.py tests/test_game_active_pmove_wiring_parity.py -q --tb=short
  • Result: 69 passed.

Task A96: Reconstruct renderer framebuffer owner and retire legacy scratch path [COMPLETED]

Priority: High Primary areas: src/code/renderer/tr_backend.c, tests/test_renderer_post_process_parity.py, tests/test_renderer_internal_helper_mapping_parity.py, docs/reverse-engineering/quakelive_steam_mapping_round_324.md, docs/reverse-engineering/renderer-full-parity-audit-and-implementation-plan-2026-04-09.md, docs/reverse-engineering/source-file-parity-ledger-2026-04-22.md, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/, src2/ghidra/quakelive_steam/quakelive_steam_decomp.cpp Parity estimate: before 99.9% -> after 99.93% for the scoped post-process framebuffer owner lane. The strict renderer estimate remains 100% and the repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked retail sub_4500B0, sub_437E40, sub_438790, and sub_4387D0 against HLIL, the Ghidra companion corpus, and the alias ledger.
2. Confirmed the live retail-shaped framebuffer owner is the GL_TEXTURE_RECTANGLE_ARB RBPP_CreateRenderTarget path with shared depth-stencil renderbuffer attachment.
3. Removed the disconnected legacy GL_TEXTURE_2D scene-target structs, FBO-loader helpers, and fixed-function scratch-bloom helper family from tr_backend.c.
4. Rewired end-of-frame release and queued post-process reset handling to call RBPP_ReleaseSceneRenderTarget and RBPP_ResetIfNeeded directly.
5. Updated the renderer audit and source-file ledger; tr_backend.c now tracks 76 functions after retiring the duplicate helper family.

Verification:

• python -m pytest tests/test_renderer_post_process_parity.py tests/test_renderer_internal_helper_mapping_parity.py tests/test_renderer_export_tail_parity.py tests/test_renderer_full_parity_gate.py -q --tb=short
  • Result: 42 passed, 1 skipped.
• python -m pytest tests/test_engine_cvar_retail_parity.py::test_engine_cvar_fortysecond_renderer_postprocess_state_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_twentyninth_renderer_postprocess_extension_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_thirtysecond_renderer_bloom_picmip_tranche_matches_retail_contracts -q --tb=short
  • Result: 3 passed.
• python -m pytest tests/test_engine_client_command_parity.py::test_postprocess_restart_routes_through_renderer_export_not_renderer_cmd_registration -q --tb=short
  • Result: 1 passed.
• git diff --check
  • Result: pass; only repository LF-to-CRLF conversion warnings.

Task A95: Reconstruct teleporting state and related wiring [COMPLETED]

Priority: High Primary areas: src/code/game/g_misc.c, tests/test_teleport_state_reconstruction.py, docs/reverse-engineering/qagame-teleport-state-reconstruction-2026-05-26.md, docs/reverse-engineering/qagame-mapping.md, docs/reverse-engineering/cgame-mapping.md, references/symbol-maps/qagame.json, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/, references/hlil/quakelive/cgamex86.dll/cgamex86.dll_hlil_split/ Parity estimate: before 96% -> after 99% for the scoped teleporting state and related wiring lane. The repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked retail TeleportPlayer at 0x1005A420, Weapon_HookFree at 0x1006E330, target/trigger teleporter callbacks, holdable teleporter event handling, spectator-door teleporting, respawn teleport-in effects, dropped-powerup teleporter handling, bot movement flags, and cgame teleport consumers against HLIL, Ghidra, and symbol-map evidence.
2. Restored the missing TeleportPlayer active-hook cleanup edge: Weapon_HookFree( player->client->hook ) now runs after EF_TELEPORT_BIT toggles and before destination view-angle / killbox handling.
3. Added focused parity tests pinning the server helper, server producers, cgame prediction/snapshot/view/event consumers, bot PMF_TIME_KNOCKBACK -> MFL_TELEPORTED bridge, and dropped-powerup teleporter movement side path.
4. Added a dedicated reconstruction note and updated qagame/cgame mapping ledgers with the observed facts, source delta, confidence, and open evidence boundary.

Verification:

• python -m pytest tests/test_teleport_state_reconstruction.py -q --tb=short
  • Result: 4 passed.
• python -m pytest tests/test_teleport_state_reconstruction.py tests/test_pmove_helper_parity.py tests/test_game_item_runframe_parity.py tests/test_game_target_parity.py tests/test_game_runthink_parity.py tests/test_cgame_snapshot_parity.py tests/test_cgame_event_transport_parity.py -q --tb=short
  • Result: 88 passed.
• git diff --check
  • Result: pass; only repository LF-to-CRLF conversion warnings.

Task A94: Split renderer color-correct init and live browser override [COMPLETED]

Priority: High Primary areas: src/code/renderer/tr_backend.c, tests/test_renderer_post_process_parity.py, tests/test_renderer_internal_helper_mapping_parity.py, tests/test_engine_cvar_retail_parity.py, docs/reverse-engineering/quakelive_steam_mapping_round_323.md, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/, src2/ghidra/quakelive_steam/quakelive_steam_decomp.cpp Parity estimate: before 99.85% -> after 99.9% for the scoped post-process color-correct browser override lane. The strict renderer estimate remains 100% and the repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked retail sub_43CD60 and sub_43CFE0 against HLIL and the Ghidra companion corpus.
2. Confirmed that the browser-active override belongs to the live uniform refresh helper, while the init helper seeds gamma reciprocal and contrast directly from cvars.
3. Added a browserOverride parameter to the shared source uniform writer so RBPP_InitColorCorrectResources passes qfalse and RBPP_SetColorCorrectUniformsFromCvars passes qtrue.

Verification:

• python -m pytest tests/test_renderer_post_process_parity.py tests/test_renderer_internal_helper_mapping_parity.py tests/test_renderer_export_tail_parity.py tests/test_renderer_full_parity_gate.py -q --tb=short
  • Result: 41 passed, 1 skipped.
• python -m pytest tests/test_engine_cvar_retail_parity.py::test_engine_cvar_fortysecond_renderer_postprocess_state_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_twentyninth_renderer_postprocess_extension_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_thirtysecond_renderer_bloom_picmip_tranche_matches_retail_contracts -q --tb=short
  • Result: 3 passed.
• python -m pytest tests/test_engine_cvar_retail_parity.py::test_engine_cvar_twentythird_renderer_runtime_tuning_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_thirtyeighth_renderer_image_quality_tranche_matches_retail_contracts -q --tb=short
  • Result: 2 passed.
• python -m pytest tests/test_engine_client_command_parity.py::test_postprocess_restart_routes_through_renderer_export_not_renderer_cmd_registration -q --tb=short
  • Result: 1 passed.
• git diff --check
  • Result: pass; only repository LF-to-CRLF conversion warnings.

Task A93: Reconstruct renderer live post-process cvar refresh [COMPLETED]

Priority: High Primary areas: src/code/renderer/tr_cmds.c, src/code/renderer/tr_backend.c, src/code/renderer/tr_init.c, src/code/renderer/tr_local.h, tests/test_renderer_post_process_parity.py, tests/test_renderer_internal_helper_mapping_parity.py, tests/test_engine_cvar_retail_parity.py, docs/reverse-engineering/quakelive_steam_mapping_round_322.md, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/, src2/ghidra/quakelive_steam/quakelive_steam_decomp.cpp Parity estimate: before 99.7% -> after 99.85% for the scoped post-process live cvar refresh lane. The strict renderer estimate remains 100% and the repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked the retail RE_BeginFrame post-texture-mode branch against HLIL and the Ghidra companion corpus.
2. Moved live post-process modified-latch consumption out of R_UpdatePostProcessCvars, preserving that function as the enable/restart state owner.
3. Added R_RefreshLivePostProcessCvars in tr_cmds.c so frame-begin now refreshes color-correct uniforms from r_gamma/r_contrast, refreshes bloom uniforms from the five live bloom cvars, and leaves r_gamma->modified for the later R_SetColorMappings path.
4. Exposed the two active-state guarded backend uniform helpers through tr_local.h while keeping the low-level uniform writers private.

Verification:

• python -m pytest tests/test_renderer_post_process_parity.py tests/test_renderer_internal_helper_mapping_parity.py tests/test_renderer_export_tail_parity.py tests/test_renderer_full_parity_gate.py -q --tb=short
  • Result: 40 passed, 1 skipped.
• python -m pytest tests/test_engine_cvar_retail_parity.py::test_engine_cvar_fortysecond_renderer_postprocess_state_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_twentyninth_renderer_postprocess_extension_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_thirtysecond_renderer_bloom_picmip_tranche_matches_retail_contracts -q --tb=short
  • Result: 3 passed.
• python -m pytest tests/test_engine_client_command_parity.py::test_postprocess_restart_routes_through_renderer_export_not_renderer_cmd_registration -q --tb=short
  • Result: 1 passed.
• Direct source sentinel for the test_engine_cvar_fortyfirst_renderer_platform_scene_tranche_matches_retail_contracts post-process assertions:
  • Result: passed.
• python -m pytest tests/test_engine_cvar_retail_parity.py::test_engine_cvar_fortyfirst_renderer_platform_scene_tranche_matches_retail_contracts -q --tb=short
  • Result: failed before the new post-process assertions on an unrelated read-only src/ui/ui_main.c r_inGameVideo expectation.
• git diff --check
  • Result: pass; only repository LF-to-CRLF conversion warnings.

Task A92: Reconstruct renderer color-correct uniform helper [COMPLETED]

Priority: High Primary areas: src/code/renderer/tr_backend.c, tests/test_renderer_post_process_parity.py, tests/test_renderer_internal_helper_mapping_parity.py, docs/reverse-engineering/quakelive_steam_mapping_round_321.md, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/, src2/ghidra/quakelive_steam/quakelive_steam_decomp.cpp Parity estimate: before 99.5% -> after 99.7% for the scoped post-process color-correct uniform helper lane. The strict renderer estimate remains 100% and the repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked sub_43CD60, sub_43CFE0, and sub_436DC0 against HLIL and the Ghidra companion corpus.
2. Split color-correct cvar-to-uniform writes out of the draw pass into RBPP_SetColorCorrectUniforms and the active-state guarded RBPP_SetColorCorrectUniformsFromCvars helper.
3. Seeded the color-correct program uniforms during RBPP_InitColorCorrectResources, matching the initialization write sequence recovered from sub_43CFE0.
4. Left RBPP_ApplyColorCorrectPass focused on framebuffer copy, live uniform refresh through the recovered helper, and full-screen draw execution.

Verification:

• python -m pytest tests/test_renderer_post_process_parity.py tests/test_renderer_internal_helper_mapping_parity.py tests/test_renderer_export_tail_parity.py tests/test_renderer_full_parity_gate.py -q --tb=short
  • Result: 39 passed, 1 skipped.
• python -m pytest tests/test_engine_client_command_parity.py::test_postprocess_restart_routes_through_renderer_export_not_renderer_cmd_registration -q --tb=short
  • Result: 1 passed.
• python -m pytest tests/test_engine_cvar_retail_parity.py::test_engine_cvar_fortysecond_renderer_postprocess_state_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_twentyninth_renderer_postprocess_extension_tranche_matches_retail_contracts -q --tb=short
  • Result: 2 passed.
• git diff --check
  • Result: pass; only repository LF-to-CRLF conversion warnings.

Task A91: Correct renderer post-process bloom uniform slot order [COMPLETED]

Priority: High Primary areas: src/code/renderer/tr_backend.c, src/code/renderer/tr_local.h, tests/test_renderer_post_process_parity.py, tests/test_renderer_internal_helper_mapping_parity.py, docs/reverse-engineering/quakelive_steam_mapping_round_320.md, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/, src2/ghidra/quakelive_steam/quakelive_steam_decomp.cpp Parity estimate: before 99% -> after 99.5% for the scoped post-process bloom uniform naming lane. The strict renderer estimate remains 100% and the repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked sub_4380F0, sub_438590, and sub_4386D0 against HLIL and the Ghidra companion corpus for the combine uniform slot order.
2. Corrected the private five-float source signature so the last two arguments are sceneSaturation and then sceneIntensity.
3. Reordered the source-side combine uniform fields, uniform lookup, and uniform write path to mirror the retail slots: p_bloomsaturation, p_scenesaturation, p_bloomintensity, p_sceneintensity.
4. Added a dedicated mapping note and parity sentinels so the argument/slot order cannot drift back to shader-declaration order.

Verification:

• python -m pytest tests/test_renderer_post_process_parity.py tests/test_renderer_internal_helper_mapping_parity.py tests/test_renderer_export_tail_parity.py tests/test_renderer_full_parity_gate.py -q --tb=short
  • Result: 38 passed, 1 skipped.
• python -m pytest tests/test_engine_client_command_parity.py::test_postprocess_restart_routes_through_renderer_export_not_renderer_cmd_registration -q --tb=short
  • Result: 1 passed.
• python -m pytest tests/test_engine_cvar_retail_parity.py::test_engine_cvar_thirtysecond_renderer_bloom_picmip_tranche_matches_retail_contracts -q --tb=short
  • Result: 1 passed.
• git diff --check
  • Result: pass; only repository LF-to-CRLF conversion warnings.

Task A90: Reconstruct qagame knockback application and blocking [COMPLETED]

Priority: High Primary areas: src/code/game/g_combat.c, src/game/g_config.c, tests/test_game_weapon_parity.py, tests/test_pmove_helper_parity.py, docs/reverse-engineering/qagame-knockback-reconstruction-2026-05-26.md, docs/reverse-engineering/qagame-mapping.md, docs/gameplay/cvars.md, docs/gameplay-cvars-and-steam.md, references/symbol-maps/qagame.json, references/symbol-maps/cgame.json, src/code/game/ai_dmq3.c, src/code/game/g_target.c, src/code/game/g_missile.c, src/code/botlib/be_aas_move.c, src/code/botlib/be_ai_move.c, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/ Parity estimate: before 82% -> after 98% for the scoped qagame knockback application, no-knockback producer, and movement-blocking lane. The repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked retail G_KnockbackScaleForMOD at 0x10048A10 and G_Damage at 0x10048C30 against HLIL, Ghidra, and the existing symbol map.
2. Reconstructed signed knockback application so negative scalars invert the damage direction and still latch PMF_TIME_KNOCKBACK.
3. Kept FL_NO_KNOCKBACK and DAMAGE_NO_KNOCKBACK as zeroing gates before velocity or timer side effects.
4. Removed non-retail G_Damage vertical boost and crouch/low-health cripple velocity reduction; g_knockback_cripple now remains only as the retail knockback timer floor.
5. Added focused source/evidence sentinels and a dedicated reconstruction note for the signed knockback and blocking seam, including the shared pmove friction, walk acceleration, drop-timer, spawn-side timer wiring, and teleport-side timer wiring.
6. Removed the inherited Quake III damage_knockback feedback slot/write/reset after rechecking retail G_Damage and P_DamageFeedback: Quake Live stores only armor, blood, damage_from, and damage_fromWorld in that record.
7. Pinned the bot-side move initializer wiring where active PMF_TIME_KNOCKBACK plus positive pm_time raises MFL_TELEPORTED before the waterjump movement flag.
8. Pinned the no-knockback producers (target_laser_think and juiced prox discharge), the null-direction DAMAGE_NO_KNOCKBACK conversion, the fatal FL_NO_KNOCKBACK ordering, and the companion botlib weapon-jump predictor.
9. Aligned the g_max_knockback help text in both registration tables with the reconstructed positive-only clamp behavior.
10. Rechecked reconstructed source style: added the required function headers for G_ReadKnockbackCvar and G_UpdateKnockbackConfig, clarified the G_Damage knockback comment, and tab-indented knockbackConfig_t.
11. Added a graph-level wiring map to the reconstruction note and pinned the live knockback, no-knockback, pmove, feedback, bot bridge, and botlib predictor nodes in the focused parity suite.
12. Removed the source-only g_debugDamage knockback summary instrumentation after the committed HLIL/string corpus showed only the retail health/damage/armor debug print in G_Damage.

Verification:

• python -m pytest tests/test_game_weapon_parity.py -q --tb=short
  • Result: 35 passed.
• python -m pytest tests/test_game_active_pmove_wiring_parity.py -q --tb=short
  • Result: 32 passed.
• python -m pytest tests/test_pmove_helper_parity.py -q --tb=short
  • Result: 40 passed.
• python -m pytest tests/test_game_target_parity.py -q --tb=short
  • Result: 11 passed.
• python -m pytest tests/test_game_weapon_parity.py tests/test_game_active_pmove_wiring_parity.py tests/test_pmove_helper_parity.py tests/test_game_target_parity.py -q --tb=short
  • Result: 118 passed.
• git diff --check
  • Result: pass; only repository LF-to-CRLF warnings were reported.

Task A89: Reconstruct renderer post-process private refexport tail [COMPLETED]

Priority: High Primary areas: src/code/renderer/tr_backend.c, src/code/renderer/tr_init.c, src/code/renderer/tr_local.h, src/code/renderer/tr_public.h, tests/test_renderer_export_tail_parity.py, tests/test_renderer_post_process_parity.py, tests/test_renderer_internal_helper_mapping_parity.py, docs/reverse-engineering/quakelive_steam_mapping_round_319.md, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/ Parity estimate: before 96% -> after 99% for the scoped post-process private-tail lane. The strict renderer estimate remains 100% and the repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked the retail GetRefAPI tail assignments around data_5878c0 through data_5878cc, including j_sub_4384D0 and sub_451420.
2. Added the missing RetailBloomPostProcessCommand refexport slot between scene-target capture and post-process restart.
3. Assigned the bloom command and five-float bloom uniform setter through GetRefAPI.
4. Factored the bloom uniform writes into RBPP_SetBloomUniforms and added the cvar-refresh helper matching sub_438590.
5. Mirrored the retail dirty-flag behavior so temporary tail-set bloom uniforms are restored from cvars after the next bloom command.

Verification:

• python -m pytest tests/test_renderer_post_process_parity.py tests/test_renderer_internal_helper_mapping_parity.py tests/test_renderer_export_tail_parity.py tests/test_renderer_full_parity_gate.py -q --tb=short
  • Result: 37 passed, 1 skipped.
• python -m pytest tests/test_engine_client_command_parity.py::test_postprocess_restart_routes_through_renderer_export_not_renderer_cmd_registration -q --tb=short
  • Result: 1 passed.
• git diff --check
  • Result: pass; only repository LF-to-CRLF conversion warnings.

Task A88: Reconstruct renderer post-process backend command ABI [COMPLETED]

Priority: High Primary areas: src/code/renderer/tr_backend.c, src/code/renderer/tr_cmds.c, src/code/renderer/tr_init.c, src/code/renderer/tr_local.h, src/code/renderer/tr_shader.c, tests/test_renderer_post_process_parity.py, tests/test_renderer_internal_helper_mapping_parity.py, docs/reverse-engineering/quakelive_steam_mapping_round_318.md, references/analysis/quakelive_symbol_aliases.json, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/ Parity estimate: before 88% -> after 96% for the scoped post-process command ABI lane. The strict renderer estimate remains 100% and the repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked the retail command emitter and executor evidence around sub_43CD10, sub_4384D0, sub_43CBA0, sub_436DC0, sub_436EC0, and the sub_437A50 backend dispatch table.
2. Reconstructed command payloads for color correction (0x10 bytes), bloom (0x38 bytes), and scene-target binding (0x04 bytes), preserving retail command IDs 9, 10, and 11.
3. Added frontend emitters and backend handlers for the post-process command lane, then routed draw-surface scene capture and end-frame post-processing through queued commands instead of the swap-buffer shortcut.
4. Updated command-list repair walking and the private refexport capture slot so the newly reconstructed commands are visible in related wiring.
5. Promoted command-handler aliases for sub_436DC0 and sub_436EC0 to match their observed command-size returns.

Verification:

• python -m pytest tests/test_renderer_post_process_parity.py tests/test_renderer_internal_helper_mapping_parity.py tests/test_renderer_export_tail_parity.py tests/test_renderer_full_parity_gate.py -q --tb=short passed: 35 passed, 1 skipped.
• python -m pytest tests/test_engine_client_command_parity.py::test_postprocess_restart_routes_through_renderer_export_not_renderer_cmd_registration -q --tb=short passed: 1 passed.
• git diff --check completed without whitespace errors; Git reported only line-ending normalization warnings for modified text files.

Task A87: Reconstruct renderer post-process bloom scene-target wiring [COMPLETED]

Priority: High Primary areas: src/code/renderer/tr_backend.c, src/code/renderer/tr_image.c, src/code/renderer/tr_local.h, tests/test_renderer_post_process_parity.py, tests/test_renderer_internal_helper_mapping_parity.py, docs/reverse-engineering/quakelive_steam_mapping_round_317.md, docs/reverse-engineering/renderer-full-parity-audit-and-implementation-plan-2026-04-09.md, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/ Parity estimate: before 98% -> after 99.5% for the scoped renderer post-process scene-target and bloom/color-correct owner boundary. The strict renderer estimate remains 100% and the repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked the retail quakelive_steam.exe post-process evidence around sub_4384A0, sub_4380F0, sub_438790, screenshot readback, and the backend command ID 0x0b scene-target rebind path.
2. Promoted the source-side RBPP_BloomEnabled() predicate so the scene render target is gated by the same broad post-process, shader-support, and r_bloomActive checks recovered from retail.
3. Moved the full-resolution scene target into the bloom resource lifecycle, matching the retail split where color correction owns a default-framebuffer copy/pass and bloom owns offscreen scene capture.
4. Rewired scene-target binding, screenshot release/rebind, and post-process submission so color-correct-only frames no longer route world rendering through the bloom scene framebuffer.
5. Added focused parity sentinels and a dedicated mapping note for the bloom-owned scene-target boundary.

Verification:

• python -m pytest tests/test_renderer_post_process_parity.py tests/test_renderer_internal_helper_mapping_parity.py tests/test_renderer_full_parity_gate.py -q --tb=short passed: 27 passed, 1 skipped.
• python -m pytest tests/test_renderer_export_tail_parity.py tests/test_engine_client_command_parity.py::test_postprocess_restart_routes_through_renderer_export_not_renderer_cmd_registration -q --tb=short passed: 7 passed.
• git diff --check completed without whitespace errors; Git reported only line-ending normalization warnings for modified text files.

Task A86: Reconstruct retail qagame score, vote, and cheat command ladder [COMPLETED]

Priority: High Primary areas: src/code/game/g_cmds.c, tests/test_game_helper_seam_parity.py, docs/reverse-engineering/qagame-score-vote-cheat-command-reconstruction-2026-05-25.md, docs/reverse-engineering/qagame-mapping.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/ Parity estimate: before 72% -> after 93% for the scoped ten-command score/stat/ready/vote/cheat ClientCommand surface (score, acc, pstats, readyup, vote, give, god, notarget, noclip, and kill) plus the adjacent ragequit boundary and next-map vote wiring. The repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked the qagame HLIL ClientCommand ladder at sub_10045DEC, the command strings around data_10084EF8 through data_10084F40, and helper strings for accuracy payloads, vote handling, cheat toggles, and kill rejection.
2. Revalidated symbol-map ownership for Cmd_Score_f, Cmd_Acc_f, Cmd_Pstats_f, Cmd_ReadyUp_f, Cmd_Vote_f, Cmd_Give_f, Cmd_God_f, Cmd_Notarget_f, Cmd_Noclip_f, Cmd_Kill_f, and ClientCommand.
3. Reordered ClientCommand so acc, pstats, readyup, and vote sit in the recovered retail pre-intermission cluster after score.
4. Moved source-local compatibility branches (ready, notready, unready, players, teams, and cvar) after the retail vote branch.
5. Removed the later post-intermission vote branch, making G_HandleNextMapVote reachable during intermission through Cmd_Vote_f.
6. Added a focused parity sentinel tying the ten command names to source dispatch order, helper bodies, HLIL command strings, and recovered vote/cheat/self-kill messages.

Open boundary:

• Retail ragequit is visible between readyup and vote, but its target is a one-argument native import at data_104B13AC + 0x300. This pass leaves it documented rather than reconstructing behavior before the import owner is cross-validated.

Verification:

• python -m pytest tests/test_game_helper_seam_parity.py -q --tb=short -k "score_vote_cheat_client_command_tranche" passed: 1 passed, 29 deselected.
• python -m pytest tests/test_game_helper_seam_parity.py -q --tb=short passed: 30 passed.
• python -m pytest tests/test_game_exit_rules_parity.py::test_nextmap_vote_pipeline_matches_retail_intermission_vote_flow tests/test_vote_ui_throttle.py::test_vote_ui_throttle_transitions -q --tb=short passed/skipped: 1 passed, 1 skipped.

Task A85: Reconstruct retail qagame chat and voice command ladder [COMPLETED]

Priority: High Primary areas: src/code/game/g_cmds.c, tests/test_game_helper_seam_parity.py, docs/reverse-engineering/qagame-chat-voice-command-reconstruction-2026-05-25.md, docs/reverse-engineering/qagame-mapping.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/ Parity estimate: before 78% -> after 95% for the scoped ten-command chat/voice ClientCommand surface (say, say_team, tell, botSay, vsay, vsay_team, vtell, vosay, vosay_team, and votell) plus the adjacent vtaunt boundary and voice/tell transport helpers. The repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked the qagame HLIL ClientCommand ladder at sub_10045DEC, the command strings at data_10084EAC through data_10084EF0, and the helper bodies around sub_10041B60, sub_10041B90, sub_10041CC0, sub_10041E40, sub_10041E60, and sub_10041F50.
2. Revalidated symbol-map ownership for Cmd_Say_f, Cmd_Tell_f, Cmd_BotSay_f, Cmd_Voice_f, Cmd_VoiceTell_f, Cmd_VoiceTaunt_f, and ClientCommand.
3. Reordered ClientCommand so the retail chat/voice cluster flows directly from botSay into vsay; the source-local complaint hook remains supported, but now sits after the adjacent vtaunt retail boundary.
4. Added a focused parity sentinel tying the ten command names to source dispatch order, handler calls, HLIL strings, voice transport commands, private tell echo behavior, bot chat payloads, and voice flood labels.
5. Added a dedicated reconstruction note and refreshed the qagame mapping ledger with this source-reconstruction pass.

Verification:

• python -m pytest tests/test_game_helper_seam_parity.py -q --tb=short -k "chat_voice_client_command_tranche or game_say_reconstructs" passed: 2 passed, 27 deselected.
• python -m pytest tests/test_game_helper_seam_parity.py -q --tb=short passed: 29 passed.

Task A84: Reconstruct retail qagame server-command wiring tranche [COMPLETED]

Priority: High Primary areas: src/code/game/g_cmds.c, src/code/game/g_svcmds.c, src/code/game/g_bot.c, src/code/game/g_mem.c, tests/test_game_helper_seam_parity.py, docs/reverse-engineering/qagame-server-command-wiring-reconstruction-2026-05-25.md, docs/reverse-engineering/qagame-mapping.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/ Parity estimate: before 82% -> after 94% for the scoped ten-command qagame server-command surface (addscore, addteamscore, setmatchtime, entitylist, forceteam, game_memory, addbot, botlist, game_crash, and reload_access) plus the related direct-table, console-dispatch, bot media-refresh, and helper wiring. The repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked the owning qagamex86.dll corpus metadata/import/export/function evidence, the qagame symbol map, the data_10080750 direct command table, ConsoleCommand at sub_10066B90, and Svcmd_AddBot_f at sub_10037910.
2. Revalidated the retail direct command table tail rows for addscore, addteamscore, and setmatchtime against the recovered command strings, handlers, privilege floors, and observable score/time messages.
3. Revalidated server-console dispatch for entitylist, forceteam, game_memory, addbot, botlist, game_crash, and reload_access against source helpers and HLIL command-token comparisons.
4. Corrected Svcmd_AddBot_f so its local-client media refresh sends loaddeferred\n, matching retail Quake Live and the existing G_AddTrainerBot path.
5. Added a focused parity sentinel tying the ten command names to source dispatch, helper bodies, symbol-map identities, HLIL strings, and the corrected addbot server-command spelling.
6. Added a dedicated reconstruction note and refreshed the qagame mapping ledger with this source-reconstruction pass.

Verification:

• python -m pytest tests/test_game_helper_seam_parity.py -q --tb=short -k "server_command_wiring or console_tail or direct_score_time" passed: 3 passed, 25 deselected.
• python -m pytest tests/test_game_helper_seam_parity.py -q --tb=short passed: 28 passed.

Task A83: Reconstruct retail qagame direct command table tail [COMPLETED]

Priority: High Primary areas: src/code/game/g_cmds.c, tests/test_game_helper_seam_parity.py, docs/reverse-engineering/qagame-direct-command-table-tail-reconstruction-2026-05-25.md, docs/reverse-engineering/qagame-mapping.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/ Parity estimate: before 18% -> after 91% for the scoped remaining direct-command table tail (addscore, addteamscore, setmatchtime, and the help-visible null-handler rcon row). The repo-wide parity estimate remains 98%.

Completed work:

1. Rechecked the data_10080750 tail rows at 0x100808F4, 0x10080908, 0x1008091C, and 0x10080930, plus the sub_10061670, sub_10061730, sub_10062CE0, and sub_10070B40 HLIL bodies.
2. Added direct table rows for addscore, addteamscore, setmatchtime, and help-visible rcon.
3. Updated direct dispatch so matched rows with null handlers fall through instead of being invoked, matching the recovered rcon row.
4. Added score adjustment handlers for player and team score deltas with the recovered Player score adjusted., Team score adjusted., and increase/decrease print strings.
5. Added setmatchtime handling that rewrites CS_LEVEL_START_TIME from level.time, formats whole seconds as %i:%i%i, and broadcasts the recovered match-time confirmation.
6. Added a reconstruction note, qagame mapping rows, and parity sentinels for the remaining table offsets, strings, and source wiring.

Verification:

• python -m pytest tests/test_game_helper_seam_parity.py -q --tb=short -k "direct_score_time or direct_admin_access or direct_command_table" passed: 3 passed, 24 deselected.

Task A82: Reconstruct retail qagame direct admin command tranche [COMPLETED]

Priority: High Primary areas: src/code/game/g_cmds.c, src/code/game/g_main.c, src/code/game/g_local.h, tests/test_game_helper_seam_parity.py, docs/reverse-engineering/qagame-direct-admin-command-reconstruction-2026-05-25.md, docs/reverse-engineering/qagame-mapping.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/ Parity estimate: before 34% -> after 87% for the scoped ten-command direct admin/access command surface (mute, unmute, tempban, ban, listaccess, unban, addadmin, addmod, demote, and abort) plus the related access-list page printer, access-entry mutation helpers, kick/drop helper, and opsay privilege-floor correction. The repo-wide parity estimate remains 98%; later direct table rows such as addscore, addteamscore, and setmatchtime are closed by Task A83.

Completed work:

1. Rechecked the qagame HLIL direct command table at data_10080750, the symbol map rows for 0x10061550, 0x10062470 through 0x10062C60, and the access-list page helper at 0x100327D0.
2. Extended the direct command table with the ten selected retail rows and corrected opsay from admin to moderator privilege based on the adjacent table-entry byte layout.
3. Reworked mute and unmute around the shared PlayerID resolver and recovered broadcast strings, including the same-or-higher target guard for mute.
4. Added source-side access-list mutation helpers and G_PrintAccessListPage with the recovered 20-entry page size, separator, tier labels, and TEMP/PERM mode column.
5. Rebuilt tempban and ban through the retail-style shared helper: direct rows use PlayerID lookup, refuse privileged targets, write temporary or permanent ban entries, and drop clients with was kicked.
6. Added addadmin, addmod, demote, unban, listaccess, and abort handlers, including the recovered promotion/demotion broadcasts, priv %i refreshes, unban confirmation, timeout-active abort rejection, PRE_GAME reset, and map_restart 3.
7. Added a dedicated reconstruction note, expanded qagame mapping rows, and added helper seam parity tests covering the selected table offsets, exact HLIL strings, access-list format, and source wiring.

Verification:

• python -m pytest tests/test_game_helper_seam_parity.py -q --tb=short -k "direct_admin_access or direct_mute or direct_command_table or timeout_race" passed: 4 passed, 22 deselected.
• python -m pytest tests/test_game_helper_seam_parity.py -q --tb=short passed: 26 passed.

Task A81: Restore retail parity for access, A&D bonus, flood, drop-health, knockback, and vampiric g_* CVar tranche [COMPLETED]

Priority: High Primary areas: src/code/game/g_main.c, src/game/g_config.c, src/code/game/g_combat.c, src/code/game/g_items.c, src/code/game/g_team.c, src/code/game/g_active.c, src/code/game/g_cmds.c, src/code/game/g_svcmds.c, tests/test_game_active_pmove_wiring_parity.py, docs/gameplay/cvars.md, docs/attack_defend_bonuses.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/ Parity estimate: before 69% -> after 100% for the scoped ten-CVar registration/default/flag/wiring surface. This batch explicitly skipped the vote/complaint rows and the A78-A80 rows already closed in this local run, while also avoiding rows called out as inspected by session 019e6082-1189-74e3-8bdc-423ce9f7ef4d.

Completed work:

1. Rechecked retail qagame HLIL rows, defaults, and registration flags for g_accessFile, g_adTouchScoreBonus, g_adElimScoreBonus, g_adCaptureScoreBonus, g_floodprot_maxcount, g_floodprot_decay, g_dropDamagedHealth, g_max_knockback, g_returnFlagOnSuicide, and g_vampiricDamage.
2. Corrected the Attack & Defend bonus rows from source-local archive flags to retail CVAR_SERVERINFO | CVAR_GAMERULE while preserving the recovered 1, 2, and 3 defaults and score/announcement wiring.
3. Corrected flood protection to retail runtime-only rows, including the g_floodprot_maxcount=10 default shared by active-client decay, command flood gating, and floodstatus.
4. Corrected g_dropDamagedHealth to retail default 0 with CVAR_TEMP | 0x00040000 | CVAR_GAMERULE, and aligned the custom-settings digest so the DROP DAMAGED HEALTH bit only appears when the cvar is enabled.
5. Corrected g_max_knockback to retail default 120 and CVAR_GAMERULE across both the legacy game cvar table and the config helper table, then moved the invalid-value fallback used by knockback config/damage code to the same value.
6. Corrected g_returnFlagOnSuicide to retail default 0 and CVAR_GAMERULE, while preserving the cached flag config wiring.
7. Corrected g_vampiricDamage to retail 0x00040000 | CVAR_GAMERULE and revalidated the combat reward, custom-settings, and Steam tag hooks.
8. Documented the tranche in gameplay CVar notes and refreshed the A&D bonus tuning note so it no longer describes the retail gamerule rows as archived.

Verification:

• python -m pytest tests/test_game_active_pmove_wiring_parity.py -q --tb=short passed after the tranche update.
• python -m pytest tests/test_game_weapon_parity.py tests/test_bot_resource_loading.py tests/test_spawn_spec_cvars.py -q --tb=short passed for adjacent knockback/access/spawn documentation sentinels.

Task A80: Restore retail parity for Freeze reset, last-man, and Red Rover zombie g_* CVar tranche [COMPLETED]

Priority: High Primary areas: src/code/game/g_main.c, src/code/game/g_active.c, src/code/game/g_freeze.c, src/code/game/g_team.c, src/code/game/g_client.c, tests/test_game_active_pmove_wiring_parity.py, tests/test_freeze_cvars.py, tests/test_game_helper_seam_parity.py, docs/gameplay/cvars.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/ Parity estimate: before 57% -> after 100% for the scoped ten-CVar Freeze reset, last-man-standing, and Red Rover zombie spawn-trait registration/default/wiring surface. This batch explicitly excluded the g_* cvars already closed by session 019e6082-1189-74e3-8bdc-423ce9f7ef4d and the local A55-A79 ledger.

Completed work:

1. Rechecked retail qagame HLIL rows, defaults, and registration flags for g_freezeThawWinningTeam, g_freezeThawThroughSurface, g_freezeResetWeaponsOnRound, g_freezeResetHealthOnRound, g_freezeResetArmorOnRound, g_freezeRemovePowerupsOnRound, g_lastManStandingWarning, g_lastManStandingMessage, g_rrInfectedZombieHealthBonus, and g_rrInfectedZombieSpeed.
2. Corrected the six Freeze reset/thaw rows to retail CVAR_GAMERULE surfaces, including the retail g_freezeThawWinningTeam=1 default.
3. Corrected the last-man rows to zero flags and corrected g_lastManStandingMessage and its empty-cvar fallback to the retail You are the last standing payload.
4. Corrected the Red Rover infected zombie health and speed rows to retail CVAR_GAMERULE surfaces while preserving the recovered 50 and 1.15 defaults.
5. Revalidated functionality through Freeze config caching, thaw-through-wall line-of-sight gates, winner thaw release, reset weapon/health/armor/powerup handling, CA/FZ/A&D/RR last-alive notification, Red Rover infected spawn health finalization, and infected speed scaling.
6. Added HLIL-backed parity sentinels and updated gameplay cvar notes for the audited controls.

Verification:

• python -m pytest tests/test_game_active_pmove_wiring_parity.py -q --tb=short passed: 24 passed.
• python -m pytest tests/test_freeze_cvars.py tests/test_game_helper_seam_parity.py::test_client_spawn_uses_recovered_loadout_and_rr_helpers tests/test_game_helper_seam_parity.py::test_red_rover_helpers_match_recovered_retail_boundaries tests/test_game_round_controller_helper_parity.py::test_red_rover_controller_readback_helpers_match_retail_mapping_surface -q --tb=short passed: 7 passed.
• python -m pytest tests/test_cgame_event_transport_parity.py::test_last_alive_shared_helper_emits_retail_last_standing_sound -q --tb=short passed: 1 passed.

Task A79: Reconstruct retail qagame direct client-command tranche [COMPLETED]

Priority: High Primary areas: src/code/game/g_cmds.c, src/code/game/g_local.h, tests/test_game_helper_seam_parity.py, docs/reverse-engineering/qagame-client-command-reconstruction-2026-05-25.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/ Parity estimate: before 42% -> after 89% for the scoped ten-command direct client-command surface (players, timeout, timein, allready, pause, unpause, lock, unlock, putteam, and opsay) plus the related ? help-table wiring. The repo-wide parity estimate remains 98% while other command-table rows and compatibility-only surfaces remain separate focused tasks.

Completed work:

1. Rechecked the qagamex86.dll metadata/import/export corpus, the symbol map, and the HLIL direct command table at data_10080750.
2. Replaced the single-entry direct opsay dispatcher with a shared retail-like direct command table, privilege floor checks, and help filtering.
3. Rebuilt /players around the retail %2d %llu %c %s row format, cached SteamID64 values, and the recovered " MA*" privilege marker map.
4. Added the shared direct-command state gate, PlayerID parser, and team-token parser with the exact recovered rejection strings.
5. Added the per-team lock latch used by lock, unlock, and G_TeamJoinAllowed while preserving g_teamSpawnAsSpec as a compatibility guard.
6. Reworked allready, putteam, lock, unlock, timeout, pause, timein, and opsay wiring so the selected tranche follows the recovered table and helper boundaries.
7. Added a dedicated reconstruction note and expanded the helper seam parity tests for the selected table rows, helper names, HLIL strings, and source wiring.

Verification:

• python -m pytest tests/test_game_helper_seam_parity.py::test_game_direct_command_table_reconstructs_retail_client_command_tranche tests/test_game_helper_seam_parity.py::test_team_join_guard_and_connect_broadcast_match_retail_flow tests/test_game_helper_seam_parity.py::test_timeout_race_and_direct_command_helpers_match_recovered_boundaries -q --tb=short passed: 3 passed.

Task A78: Restore retail parity for Red Rover infection g_* CVar tranche [COMPLETED]

Priority: High Primary areas: src/code/game/g_main.c, src/code/game/g_client.c, src/code/game/g_active.c, src/code/game/g_cmds.c, src/code/game/g_team.c, tests/test_game_active_pmove_wiring_parity.py, tests/test_game_round_controller_helper_parity.py, tests/test_game_helper_seam_parity.py, tests/test_cgame_event_transport_parity.py, docs/gameplay/cvars.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/ Parity estimate: before 46% -> after 100% for the scoped ten-CVar Red Rover infection registration/default/wiring surface. Defaults were mostly present but one literal default differed and all selected rows had legacy archive/norestart flag surfaces instead of retail gamerule/latch surfaces. The repo-wide parity estimate remains 98% while remaining g_* rows continue to be audited in focused batches.

Completed work:

1. Rechecked retail qagame HLIL rows, defaults, and registration flags for g_rrAllowNegativeScores, g_rrDamageScoreBonus, g_rrInfected, g_rrInfectedSpreadTime, g_rrInfectedSpreadWarningTime, g_rrInfectedSurvivorMinSpeed, g_rrInfectedSurvivorPingRate, g_rrInfectedSurvivorScoreBonus, g_rrInfectedSurvivorScoreMethod, and g_rrInfectedSurvivorScoreRate.
2. Corrected the nine infection scoring/spread/survivor rows to CVAR_GAMERULE and corrected g_rrInfected to the retail CVAR_LATCH | GAME_CVAR_FLAG_RETAIL_40000 | CVAR_GAMERULE surface.
3. Corrected g_rrInfectedSurvivorMinSpeed to the retail literal default 500.0f.
4. Revalidated functional wiring through negative-score clamping, damage-based survivor scoring, survival bonus cadence, infection spread and warning timers, survivor slow-movement pings, ping-rate configstrings, custom-settings export, infected autojoin routing, infection seeding, round completion, death conversion, and last-alive notification paths.
5. Added HLIL-backed parity sentinels and updated gameplay cvar notes for the audited Red Rover infection controls.

Verification:

• python -m pytest tests/test_game_active_pmove_wiring_parity.py -q --tb=short passed: 22 passed.
• python -m pytest tests/test_game_round_controller_helper_parity.py::test_red_rover_controller_readback_helpers_match_retail_mapping_surface tests/test_game_round_controller_helper_parity.py::test_red_rover_autojoin_helper_routes_team_selection tests/test_game_helper_seam_parity.py::test_client_spawn_uses_recovered_loadout_and_rr_helpers tests/test_game_helper_seam_parity.py::test_client_begin_and_respawn_dispatch_freeze_and_rr_like_retail tests/test_game_helper_seam_parity.py::test_red_rover_helpers_match_recovered_retail_boundaries tests/test_cgame_event_transport_parity.py::test_qagame_infected_event_uses_retail_recipient_slot tests/test_cgame_event_transport_parity.py::test_red_rover_survival_bonus_emits_retail_global_team_sound -q --tb=short passed: 7 passed.

Task A77: Restore retail parity for round and Freeze timing g_* CVar tranche [COMPLETED]

Priority: High Primary areas: src/code/game/g_main.c, src/code/game/g_active.c, src/code/game/g_client.c, src/code/game/g_freeze.c, src/code/game/g_combat.c, tests/test_game_active_pmove_wiring_parity.py, tests/test_freeze_cvars.py, tests/test_game_round_controller_helper_parity.py, docs/gameplay/cvars.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/ Parity estimate: before 58% -> after 100% for the scoped ten-CVar round draw, Freeze warmup, thaw, protection, environmental respawn, and auto-thaw registration/default/wiring surface. The repo-wide parity estimate remains 98% while remaining g_* rows continue to be audited in focused batches.

Completed work:

1. Rechecked retail qagame HLIL rows, defaults, and registration flags for g_roundWarmupDelay, g_roundDrawLivingCount, g_roundDrawHealthCount, g_freezeThawTime, g_freezeThawTick, g_freezeThawRadius, g_freezeRoundDelay, g_freezeProtectedSpawnTime, g_freezeEnvironmentalRespawnDelay, and g_freezeAutoThawTime.
2. Corrected registration flags: draw gates and most Freeze timing rows now carry CVAR_GAMERULE, while g_roundWarmupDelay and g_freezeRoundDelay now match the retail CVAR_SERVERINFO | CVAR_GAMERULE surface.
3. Corrected recovered retail defaults: g_freezeThawTick is now 1, g_freezeProtectedSpawnTime is now 0, g_freezeEnvironmentalRespawnDelay is now 5000, and g_freezeAutoThawTime is now 120000.
4. Revalidated functionality through CA/FZ draw resolution, end-round living and health prints, warmup-delay rescheduling, Freeze config caching, thaw radius and tick accumulation, post-round delay, post-thaw protection, environmental respawn release, and auto-thaw deadlines.
5. Added HLIL-backed parity sentinels and updated gameplay cvar notes for the audited round and Freeze timing controls.

Verification:

• python -m pytest tests/test_game_active_pmove_wiring_parity.py -q --tb=short passed: 20 passed.
• python -m pytest tests/test_game_active_pmove_wiring_parity.py tests/test_freeze_cvars.py tests/test_game_round_controller_helper_parity.py tests/test_game_runframe_parity.py -q --tb=short passed: 34 passed.
• A broader adjacent sweep that also included tests/test_game_exit_rules_parity.py still exposes the unrelated existing ExitLevel mapname sentinel mismatch outside this tranche; the touched round/Freeze timing surface is covered by the passing focused and adjacent suites above.

Task A76: Restore retail parity for player-appearance, lag rewind, and instagib g_* CVar tranche [COMPLETED]

Priority: High Primary areas: src/code/game/g_main.c, src/code/game/g_active.c, src/code/game/g_client.c, src/code/game/g_pmove.c, src/code/game/g_weapon.c, tests/test_game_active_pmove_wiring_parity.py, tests/test_cgame_displaycontext_parity.py, docs/gameplay/cvars.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/ Parity estimate: before 72% -> after 100% for the scoped ten-CVar player-appearance, lag rewind, and instagib registration/default/wiring surface. The repo-wide parity estimate remains 98% while remaining g_* rows continue to be audited in focused batches.

Completed work:

1. Rechecked retail qagame HLIL rows, defaults, and registration flags for g_allowCustomHeadmodels, g_playerCylinders, g_playerheadmodelOverride, g_playerheadScale, g_playerheadScaleOffset, g_playermodelOverride, g_playerModelScale, g_lagHaxHistory, g_lagHaxMs, and g_instaGib.
2. Added the recovered 0x00010000 game cvar flag alias and corrected the seven player-appearance rows so they match retail: six 0x00010000 | CVAR_GAMERULE rows and the dedicated g_playerCylinders CVAR_GAMERULE row.
3. Corrected g_lagHaxHistory and g_lagHaxMs to retail CVAR_LATCH rows while preserving the recovered 4 and 80 defaults.
4. Corrected g_instaGib to the retail CVAR_SERVERINFO | 0x00040000 | CVAR_GAMERULE registration surface.
5. Revalidated functional wiring through player-cylinder and player-appearance configstrings, forced model/head userinfo rewrites, admin config caching, lag rewind allocation/store/shift/restore, weapon hitscan rewind bracketing, pmove instagib no-player-clip caching, factory reset, custom-settings, and rank payload export.
6. Added HLIL-backed parity sentinels and updated gameplay cvar notes for the audited player-appearance, lag rewind, and instagib controls.

Verification:

• python -m pytest tests/test_game_active_pmove_wiring_parity.py -q --tb=short passed: 18 passed.
• python -m pytest tests/test_cgame_displaycontext_parity.py::test_game_cvar_transport_redundancy_stays_off_serverinfo_slabs tests/test_cgame_displaycontext_parity.py::test_cgame_player_cylinders_configstring_retains_direct_retail_parser_boundary tests/test_cgame_displaycontext_parity.py::test_cgame_player_appearance_configstring_reconstruction_uses_retail_parser_and_head_gate -q --tb=short passed: 3 passed.
• python -m pytest tests/test_game_active_pmove_wiring_parity.py tests/test_cgame_displaycontext_parity.py::test_game_cvar_transport_redundancy_stays_off_serverinfo_slabs tests/test_cgame_displaycontext_parity.py::test_cgame_player_cylinders_configstring_retains_direct_retail_parser_boundary tests/test_cgame_displaycontext_parity.py::test_cgame_player_appearance_configstring_reconstruction_uses_retail_parser_and_head_gate tests/test_engine_netcode_parity.py tests/test_pmove_selected_cvar_parity.py tests/test_game_factory_regen_parity.py -q --tb=short passed: 74 passed.
• A broader python -m pytest tests/test_engine_cvar_retail_parity.py -q --tb=short sweep still exposes an unrelated existing Steamworks sentinel mismatch in src/code/qcommon/common.c; the touched g_* tranche is covered by the passing focused and adjacent suites above.

Task A75: Restore retail parity for flag-physics and forced-override g_* CVar tranche [COMPLETED]

Priority: High Primary areas: src/code/game/g_main.c, src/code/game/g_local.h, src/code/game/g_combat.c, src/code/game/g_team.c, tests/test_game_active_pmove_wiring_parity.py, docs/gameplay/cvars.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/ Parity estimate: before 54% -> after 100% for the scoped ten-CVar flag-physics, friendly-fire dampening, and forced-override registration/wiring surface. The repo-wide parity estimate remains 98% while remaining g_* rows continue to be audited in focused batches.

Completed work:

1. Rechecked retail qagame HLIL rows, defaults, and registration flags for g_flagBounce, g_flagPhysics, g_throwFlagVelocity, g_throwFlagForwardMult, g_tackleFlag, g_droppedFlagBonus, g_neutralFlagPingRate, g_friendlyFireDampen, g_forceDmgThroughSurface, and g_forceAtmosphericEffects.
2. Corrected the flag cvar rows to retail defaults and flags, including g_flagBounce default 0.25/CVAR_GAMERULE, g_throwFlagForwardMult default 2.5 with no flags, g_droppedFlagBonus default 1/CVAR_TEMP, and the retail g_neutralFlagPingRate registration name/default/flag surface.
3. Added the missing g_friendlyFireDampen row and wired it into same-team damage when g_friendlyFire allows friendly hits.
4. Updated the flag config cache so the throw forward multiplier is stored as a float and both item drops and carried-flag tosses preserve classic behavior unless g_flagPhysics enables the retail physics path.
5. Revalidated forced override wiring for atmosphere and damage-through-surface configstring publication.

Verification:

• python -m pytest tests/test_game_active_pmove_wiring_parity.py -q --tb=short passed: 16 passed.
• python -m pytest tests/test_game_active_pmove_wiring_parity.py tests/test_game_callvote_option_parity.py tests/test_game_item_runframe_parity.py tests/test_game_target_parity.py tests/test_game_runframe_parity.py tests/test_game_native_export_helper_parity.py tests/test_game_visibility_queue_parity.py tests/test_engine_netcode_parity.py -q --tb=short passed: 68 passed.
• git diff --check -- ... reported no whitespace errors; Git only warned that touched text files will be normalized from LF to CRLF the next time Git writes them.

Task A74: Restore retail parity for Domination and auto-shuffle g_* CVar tranche [COMPLETED]

Priority: High Primary areas: src/code/game/g_main.c, src/code/game/g_team.c, tests/test_game_active_pmove_wiring_parity.py, docs/gameplay/cvars.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/ Parity estimate: before 38% -> after 100% for the scoped ten-CVar Domination and auto-shuffle registration/wiring surface. The repo-wide parity estimate remains 98% while remaining g_* rows continue to be audited in focused batches.

Completed work:

1. Rechecked retail qagame HLIL rows, defaults, and registration flags for g_domCapTime, g_domTeammateCapScale, g_domDistressThreshold, g_domEnableContention, g_domNeutralFlag, g_domScoreRate, g_shuffle_timedelay, g_shuffle_minplayers, g_shuffle_automatic, and g_shuffle_automatic_minplayers.
2. Corrected the six Domination rows from legacy archived controls to retail CVAR_GAMERULE rows while preserving the recovered defaults.
3. Corrected the four auto-shuffle rows to retail zero-flag rows, including g_shuffle_timedelay default 5000, g_shuffle_minplayers default 3, g_shuffle_automatic default 0, and g_shuffle_automatic_minplayers default 6.
4. Corrected Team_UpdateAutoShuffleState so g_shuffle_timedelay is treated as milliseconds and passed directly into G_AutoShuffleCountdown_Arm, while log/UI announcements round the millisecond value up to seconds.
5. Revalidated functional wiring through Domination capture/scoring cadence, teammate capture scaling, distress threshold, contention/neutralization rules, capture-time configstring publication, auto-shuffle threshold resolution, automatic shuffle gating, and countdown arming.
6. Added HLIL-backed parity sentinels and updated gameplay cvar notes for the audited Domination and auto-shuffle controls.

Verification:

• python -m pytest tests/test_game_active_pmove_wiring_parity.py -q --tb=short passed: 14 passed.
• python -m pytest tests/test_game_active_pmove_wiring_parity.py tests/test_auto_shuffle_countdown.py tests/test_clanarena_shuffle.py tests/test_match_state_configstring.py tests/test_game_runframe_parity.py tests/test_cgame_domination_count_parity.py tests/test_game_callvote_option_parity.py tests/test_game_module_retail_parity_gate.py -q --tb=short passed: 37 passed, 6 skipped.
• python -m pytest tests/test_game_helper_seam_parity.py::test_team_balance_helper_is_split_out_for_setteam_and_readyup tests/test_game_helper_seam_parity.py::test_input_spawn_and_host_map_paths_keep_retail_factory_gates tests/test_game_helper_seam_parity.py::test_custom_settings_cvar_helper_matches_recovered_boundary -q --tb=short passed: 3 passed.
• A broader adjacent sweep that included all of tests/test_game_helper_seam_parity.py still exposes an unrelated existing G_Say token sentinel mismatch outside this tranche.

Task A73: Restore retail parity for match-flow/warmup/timeout g_* CVar tranche [COMPLETED]

Priority: High Primary areas: src/code/game/g_main.c, src/code/game/g_client.c, src/code/game/g_combat.c, src/code/game/g_gametype_lifecycle.inc, src/code/game/g_mem.c, src/code/game/g_spawn.c, src/code/game/g_team.c, src/game/g_match_config.c, tests/test_game_active_pmove_wiring_parity.py, docs/gameplay/cvars.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/ Parity estimate: before 55% -> after 100% for the scoped ten-CVar match-state, friendly-fire, debug-allocation, spawn-grant, warmup, ready-delay, and timeout registration/wiring surface. The repo-wide parity estimate remains 98% while remaining g_* rows continue to be audited in focused batches.

Completed work:

1. Rechecked retail qagame HLIL rows, defaults, and registration flags for g_gameState, g_friendlyFire, g_debugAlloc, g_grantItemOnSpawn, g_doWarmup, g_warmup, g_warmupReadyDelay, g_warmupReadyDelayAction, g_timeoutCount, and g_timeoutLen.
2. Corrected g_friendlyFire and g_grantItemOnSpawn to retail CVAR_GAMERULE rows.
3. Corrected warmup/ready-delay rows: g_warmup now defaults to retail 10, g_doWarmup now carries CVAR_ARCHIVE, and both ready-delay controls are zero-flag rows.
4. Corrected timeout rows: g_timeoutLen now carries CVAR_GAMERULE, and g_timeoutCount now defaults to 0 with CVAR_SERVERINFO | CVAR_GAMERULE.
5. Revalidated behavior/wiring through game-state publication, friendly-fire damage gates, debug allocation prints, spawn grant scripts, warmup countdowns, duel ready-delay actions, match-config timeout loading, timeout pool refills, and match-state configstring publication.
6. Added HLIL-backed parity sentinels and documented the audited match-flow, warmup, timeout, and spawn-grant controls for server operators.

Verification:

• python -m pytest tests/test_game_active_pmove_wiring_parity.py -q --tb=short passed: 12 passed.
• python -m pytest tests/test_game_active_pmove_wiring_parity.py tests/test_game_duel_ready_delay_parity.py tests/test_gametype_lifecycle.py tests/test_match_state_configstring.py tests/test_match_sim_harness.py tests/test_game_helper_seam_parity.py tests/test_game_callvote_option_parity.py tests/test_spawn_spec_cvars.py tests/test_game_module_retail_parity_gate.py -q --tb=short passed: 81 passed, 8 skipped.

Task A72: Restore retail parity for server access/factory match g_* CVar tranche [COMPLETED]

Priority: High Primary areas: src/code/game/g_main.c, src/code/game/g_utils.c, src/code/game/g_factory.c, src/code/game/g_svcmds.c, src/code/game/g_client.c, src/game/g_match_config.c, tests/test_game_active_pmove_wiring_parity.py, docs/gameplay/cvars.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/ Parity estimate: before 64% -> after 100% for the scoped ten-CVar server access, ban/filter, factory selection, overtime/mercy, automatic action, and malformed-userinfo registration and wiring surface. The repo-wide parity estimate remains 98% while remaining g_* rows continue to be audited in focused batches.

Completed work:

1. Rechecked retail qagame HLIL rows, defaults, and registration flags for g_password, g_needpass, g_banIPs, g_filterBan, g_factory, g_factoryTitle, g_overtime, g_mercytime, g_autoAction, and g_kickBadUserinfo.
2. Corrected registration surfaces: g_factory is now the retail CVAR_SERVERINFO | CVAR_ROM row, g_overtime is CVAR_SERVERINFO | CVAR_GAMERULE, g_mercytime is CVAR_NORESTART | CVAR_GAMERULE, g_autoAction defaults to 0 with zero flags, and g_kickBadUserinfo now carries zero flags.
3. Preserved the already-matching retail rows for password/needpass, ban/filter, and factory title while tightening indentation around the audited rows.
4. Revalidated runtime wiring through password-to-needpass mirroring, IP ban and filter decisions, factory selection/title application, match overtime config loading, mercy-rule timing, malformed-userinfo kick/fallback handling, and g_autoAction parsing.
5. Added HLIL-backed parity sentinels and documented the audited server access/factory/match-flow controls for server operators.

Verification:

• python -m pytest tests/test_game_active_pmove_wiring_parity.py -q --tb=short passed: 10 passed.
• python -m pytest tests/test_game_active_pmove_wiring_parity.py tests/test_game_factory_regen_parity.py tests/test_game_helper_seam_parity.py tests/test_game_attack_defend_parity.py tests/test_game_round_controller_helper_parity.py tests/test_engine_netcode_parity.py tests/test_game_module_retail_parity_gate.py -q --tb=short passed: 81 passed, 1 skipped.
• python -m pytest tests/test_match_state_configstring.py tests/test_match_sim_harness.py -q --tb=short passed: 19 passed, 7 skipped.
• A broader adjacent sweep that also included tests/test_game_exit_rules_parity.py still exposes an unrelated existing ExitLevel sentinel mismatch: the source reads SERVERINFO_KEY_MAPNAME while the test expects the literal "mapname" string.

Task A71: Restore retail parity for Team Arena environment g_* CVar tranche [COMPLETED]

Priority: High Primary areas: src/code/game/g_main.c, src/code/cgame/cg_main.c, src/code/game/g_arenas.c, src/code/game/g_team.c, src/code/game/g_combat.c, src/code/game/g_spawn.c, tests/test_game_active_pmove_wiring_parity.py, docs/gameplay/cvars.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/, references/hlil/quakelive/cgamex86.dll/cgamex86.dll_hlil_split/ Parity estimate: before 46% -> after 100% for the scoped ten-CVar Team Arena environment, podium, Obelisk/Cube, MOTD, and dust registration surface. The repo-wide parity estimate remains 98% while remaining g_* rows continue to be audited in focused batches.

Completed work:

1. Rechecked retail qagame HLIL rows, defaults, and registration flags for g_allTalk, g_motd, g_podiumDist, g_podiumDrop, g_obeliskHealth, g_obeliskRegenAmount, g_obeliskRegenPeriod, g_obeliskRespawnDelay, g_cubeTimeout, and g_enableDust.
2. Corrected qagame registration flags: g_allTalk, g_motd, and g_enableDust are plain zero-flag rows; podium, Obelisk, and Cube timing rows now carry CVAR_GAMERULE instead of legacy zero/serverinfo/archive surfaces.
3. Corrected the cgame mirror for g_obeliskRespawnDelay to the retail zero-flag row while preserving cgame's retail CVAR_SERVERINFO mirror for g_enableDust.
4. Revalidated behavior/wiring through all-talk chat gating, MOTD publication, podium placement, Obelisk health/regen/respawn timing, Harvester cube timeout, worldspawn dust key handling, and cgame dust/effect consumers.
5. Added focused HLIL-backed parity sentinels and documented the audited Team Arena environment controls for server operators.

Verification:

• python -m pytest tests/test_game_active_pmove_wiring_parity.py -q --tb=short passed: 8 passed.
• python -m pytest tests/test_game_active_pmove_wiring_parity.py tests/test_game_factory_regen_parity.py tests/test_game_weapon_parity.py tests/test_game_battlesuit_parity.py tests/test_engine_netcode_parity.py tests/test_game_module_retail_parity_gate.py -q --tb=short passed: 85 passed, 1 skipped.
• python -m pytest tests/test_cgame_hud_parity.py tests/test_engine_netcode_parity.py -q --tb=short passed: 47 passed.
• Broader dirty-tree sweeps that included unrelated cgame display/impact and Steam platform cvar tests still expose pre-existing failures outside this tranche; the touched CVar surfaces above are covered by the passing focused and adjacent suites.
• git diff --check -- ... reported no whitespace errors; Git only warned that touched text files will be normalized from LF to CRLF the next time Git writes them.

Task A70: Restore retail parity for classic gameplay/admin g_* CVar tranche [COMPLETED]

Priority: High Primary areas: src/code/game/g_main.c, src/code/game/g_active.c, src/code/game/g_local.h, tests/test_game_active_pmove_wiring_parity.py, docs/gameplay/cvars.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/ Parity estimate: before 74% -> after 100% for the scoped ten-CVar classic gameplay/admin registration and inactivity-wiring surface. The repo-wide parity estimate remains 98% while remaining g_* rows continue to be audited in focused batches.

Completed work:

1. Rechecked retail qagame HLIL rows, defaults, and registration flags for g_speed, g_gravity, g_weaponRespawn, g_inactivity, g_inactivityWarning, g_dropInactive, g_debugMove, g_debugDamage, g_teamAutoJoin, and g_teamForceBalance.
2. Corrected registration flags for g_speed, g_gravity, g_weaponRespawn, and g_teamForceBalance so the source rows match the retail serverinfo/archive/gamerule bits recovered from HLIL.
3. Restored the retail g_inactivityWarning row with default 10, exported the cvar to game code, and replaced the hardcoded ten-second warning path with the retail cvar-driven threshold and pluralized centerprint text.
4. Wired g_dropInactive into ClientInactivityTimer: timed-out inactive clients are moved to spectator when the cvar is disabled, and are dropped only when it is enabled, matching the retail branch.
5. Added the retail-style inactivity warning reset latch for g_inactivity changes, and revalidated existing wiring for debug movement/damage, team autojoin/force-balance, speed, gravity, and weapon-respawn behavior.
6. Added focused HLIL-backed parity sentinels and documented the audited classic gameplay/admin controls for server operators.

Verification:

• python -m pytest tests/test_game_active_pmove_wiring_parity.py -q --tb=short passed: 6 passed.
• python -m pytest tests/test_game_active_pmove_wiring_parity.py tests/test_game_helper_seam_parity.py tests/test_game_factory_regen_parity.py tests/test_game_weapon_parity.py tests/test_cvar_alias_console.py tests/test_engine_netcode_parity.py tests/test_game_module_retail_parity_gate.py -q --tb=short passed: 103 passed, 1 skipped.
• git diff --check -- ... reported no whitespace errors; Git only warned that touched text files will be normalized from LF to CRLF the next time Git writes them.

Task A69: Restore retail parity for second weapon-special g_* CVar tranche [COMPLETED]

Priority: High Primary areas: src/code/game/g_main.c, src/code/game/g_missile.c, src/code/game/g_pmove.c, src/code/game/bg_pmove.c, src/code/game/bg_public.h, src/code/game/g_local.h, src/code/cgame/cg_servercmds.c, src/game/g_config.c, tests/test_game_weapon_parity.py, tests/test_game_factory_regen_parity.py, docs/gameplay/cvars.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/ Parity estimate: before 59% -> after 100% for the scoped ten-CVar weapon-special, Quad Hog, and prox-mine registration/wiring surface. The repo-wide parity estimate remains 98% while remaining g_* rows continue to be audited in focused batches.

Completed work:

1. Rechecked retail qagame HLIL rows and string/default storage for g_midAirMinHeight, g_nailspeed, g_nailspread, g_damagePlums, g_quadDamageFactor, g_quadHog, g_quadHogIdle, g_quadHogPingRate, g_quadHogTime, and g_proxMineTimeout.
2. Corrected registration defaults and flags: midair height, damage plums, Quad Damage, Quad Hog timers, and prox mine timeout now match the retail table; Nailgun speed/spread and prox mine timeout now carry the recovered 0x00040000 | CVAR_GAMERULE flag combination; Quad Hog now carries CVAR_LATCH | 0x00040000 | CVAR_GAMERULE.
3. Revalidated functional wiring through G_InitWeaponConfig, G_IsMidAirEligibleTarget, Nailgun projectile spread/speed sampling, Quad Damage scaling, server-settings publication, Quad Hog pickup/frame timers, pmove cache/compact payload mirroring, custom-settings masks, and proximity mine activation.
4. Corrected retail units: Quad Hog ping rate is milliseconds (1500) and is no longer multiplied by 1000, while Quad Hog time/idle and prox mine timeout remain seconds-based and are converted at the timer boundary.
5. Added focused HLIL-backed parity sentinels and documented the audited controls for server operators.

Verification:

• python -m pytest tests/test_game_weapon_parity.py -q --tb=short passed: 33 passed.
• python -m pytest tests/test_game_weapon_parity.py tests/test_game_factory_regen_parity.py tests/test_game_ammopack_parity.py tests/test_cvar_console_write_parity.py tests/test_game_module_retail_parity_gate.py tests/test_pmove_selected_cvar_parity.py -q --tb=short passed: 82 passed, 1 skipped.
• python -m pytest tests/test_pmove_helper_parity.py -q --tb=short passed: 38 passed.
• git diff --check -- ... reported no whitespace errors; Git only warned that touched text files will be normalized from LF to CRLF the next time Git writes them.

Task A68: Restore retail parity for first weapon-special g_* CVar tranche [COMPLETED]

Priority: High Primary areas: src/code/game/g_main.c, src/code/game/g_missile.c, src/code/game/g_weapon.c, src/code/game/g_pmove.c, tests/test_game_weapon_parity.py, tests/test_game_factory_regen_parity.py, docs/gameplay/cvars.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/ Parity estimate: before 64% -> after 100% for the scoped ten-CVar weapon-special registration and wiring surface. The repo-wide parity estimate remains 98% while remaining weapon-special and item/powerup cvar rows continue to be audited in focused batches.

Completed work:

1. Rechecked retail qagame HLIL rows and string/default storage for g_velocity_gh, g_guidedRocket, g_lightningDischarge, g_railJump, g_gauntletSpeedFactor, g_headShotDamage_rg, g_ironsights_mg, g_nailbounce, g_nailbouncepercentage, and g_nailcount.
2. Corrected registration flags: grapple velocity, lightning discharge, rail jump, rail headshot damage, and the three audited Nailgun rows now use 0x00040000 | CVAR_GAMERULE; guided rockets and gauntlet speed factor now use CVAR_GAMERULE; machinegun ironsights now use retail CVAR_TEMP | 0x00040000 | CVAR_GAMERULE.
3. Revalidated functional wiring through G_InitWeaponConfig, G_SynchronizeGrappleConfig, guided rocket think arming, G_ApplyRailJump, lightning discharge activation, G_PmoveCacheSettings, Nailgun count/bounce spawn paths, custom-settings mask participation, G_UpdateCvars, and factory config refreshes.
4. Added focused HLIL-backed parity sentinels and documented the audited weapon-special controls for server operators.

Verification:

• python -m pytest tests/test_game_weapon_parity.py -q --tb=short passed: 31 passed.
• python -m pytest tests/test_game_weapon_parity.py tests/test_game_factory_regen_parity.py tests/test_game_ammopack_parity.py tests/test_cvar_console_write_parity.py tests/test_game_module_retail_parity_gate.py tests/test_pmove_selected_cvar_parity.py -q --tb=short passed: 80 passed, 1 skipped.

Task A67: Restore retail parity for velocity/acceleration g_* CVar tranche [COMPLETED]

Priority: High Primary areas: src/code/game/g_main.c, src/code/game/g_missile.c, src/game/g_config.c, tests/test_game_weapon_parity.py, docs/gameplay/cvars.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/ Parity estimate: before 66% -> after 100% for the scoped ten-CVar projectile velocity/acceleration registration and wiring surface. The repo-wide parity estimate remains 98% while remaining weapon-special cvar rows continue to be audited in focused batches.

Completed work:

1. Rechecked retail qagame HLIL rows and string/default storage for g_accelFactor_bfg, g_accelFactor_pg, g_accelFactor_rl, g_accelRate_bfg, g_accelRate_pg, g_accelRate_rl, g_velocity_bfg, g_velocity_gl, g_velocity_pg, and g_velocity_rl.
2. Corrected retail defaults and flags: acceleration factors now use default 1 plus 0x00040000 | CVAR_GAMERULE; acceleration rates now use default 16 plus CVAR_GAMERULE; audited velocity rows now use the retail 1800, 700, 2000, and 1000 defaults plus 0x00040000 | CVAR_GAMERULE.
3. Revalidated functional wiring through G_InitWeaponConfig, G_SynchronizeRocketConfig, the rocket/plasma/BFG acceleration think callbacks, G_GetMissileAccelerationThinkTime, grenade/plasma/rocket/BFG projectile constructors, custom-settings mask participation, G_UpdateCvars, and factory config refreshes.
4. Added focused HLIL-backed parity sentinels and documented the audited projectile velocity/acceleration controls for server operators.

Verification:

• python -m pytest tests/test_game_weapon_parity.py -q --tb=short passed: 29 passed.
• python -m pytest tests/test_game_weapon_parity.py tests/test_game_factory_regen_parity.py tests/test_game_ammopack_parity.py tests/test_cvar_console_write_parity.py tests/test_game_module_retail_parity_gate.py tests/test_pmove_selected_cvar_parity.py -q --tb=short passed: 78 passed, 1 skipped.

Task A66: Restore retail parity for splash radius/falloff g_* CVar tranche [COMPLETED]

Priority: High Primary areas: src/code/game/g_main.c, src/code/game/g_weapon.c, src/code/game/g_missile.c, src/game/g_config.c, tests/test_game_weapon_parity.py, docs/gameplay/cvars.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/ Parity estimate: before 68% -> after 100% for the scoped ten-CVar splash/radius/falloff registration and wiring surface. The repo-wide parity estimate remains 98% while adjacent velocity, acceleration, and remaining weapon-special cvar rows continue to be audited in focused batches.

Completed work:

1. Rechecked retail qagame HLIL rows and string/default storage for g_splashdamage_bfg, g_splashdamage_pl, g_splashradius_bfg, g_splashradius_gl, g_splashradius_pg, g_splashradius_pl, g_splashradius_rl, g_range_lg_falloff, g_range_sg_falloff, and g_rocketsplashOffset.
2. Corrected registration names and flags: BFG/prox splash damage and all audited splash radius rows now use retail lowercase names and 0x00040000 | CVAR_GAMERULE; falloff ranges and rocket splash offset now use CVAR_GAMERULE.
3. Corrected retail defaults: BFG splash radius is 80, both falloff ranges are 768, and g_rocketsplashOffset is -10.0.
4. Revalidated functional wiring through G_InitWeaponConfig, shotgun and lightning falloff helpers, grenade/plasma/rocket/BFG/prox missile spawn paths, rocket splash-origin offset handling, custom-settings mask participation, G_UpdateCvars, and factory config refreshes.

Verification:

• python -m pytest tests/test_game_weapon_parity.py -q --tb=short passed: 27 passed.
• python -m pytest tests/test_game_weapon_parity.py tests/test_game_factory_regen_parity.py tests/test_game_ammopack_parity.py tests/test_cvar_console_write_parity.py tests/test_game_module_retail_parity_gate.py tests/test_pmove_selected_cvar_parity.py -q --tb=short passed: 76 passed, 1 skipped.

Task A65: Restore retail parity for second weapon damage g_* CVar tranche [COMPLETED]

Priority: High Primary areas: src/code/game/g_main.c, src/code/game/g_weapon.c, src/code/game/g_missile.c, src/code/game/g_combat.c, src/game/g_config.c, tests/test_game_weapon_parity.py, docs/gameplay/cvars.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/ Parity estimate: before 70% -> after 100% for the scoped ten-CVar weapon damage/splash registration and wiring surface. The repo-wide parity estimate remains 98% while the remaining splash radius, BFG splash, and weapon range/velocity cvar rows continue to be walked in focused batches.

Completed work:

1. Rechecked retail qagame HLIL rows and string/default storage for g_damage_pg, g_damage_pl, g_damage_rg, g_damage_rl, g_damage_sg, g_damage_sg_falloff, g_damage_sg_outer, g_splashdamage_gl, g_splashdamage_pg, and g_splashdamage_rl.
2. Corrected the seven remaining direct/shotgun damage rows to use the recovered 0x00040000 | CVAR_GAMERULE pair via GAME_CVAR_FLAG_RETAIL_40000.
3. Corrected the audited splash damage registrations to retail lowercase cvar names (g_splashdamage_*) and retail flags, and changed rocket splash damage from the legacy 100 default to Quake Live's 84 default across registration, cache fallback, and custom-settings comparisons.
4. Revalidated functional wiring through G_InitWeaponConfig, G_ClampModDamage, shotgun/rail weapon paths, grenade/plasma/rocket/prox missile spawn paths, custom-settings mask participation, G_UpdateCvars, and factory config refreshes.

Verification:

• python -m pytest tests/test_game_weapon_parity.py -q --tb=short passed: 25 passed.
• python -m pytest tests/test_game_weapon_parity.py tests/test_game_factory_regen_parity.py tests/test_game_ammopack_parity.py tests/test_cvar_console_write_parity.py tests/test_game_module_retail_parity_gate.py tests/test_pmove_selected_cvar_parity.py -q --tb=short passed: 74 passed, 1 skipped.

Task A64: Restore retail parity for first ten weapon damage g_* CVars [COMPLETED]

Priority: High Primary areas: src/code/game/g_main.c, src/code/game/g_weapon.c, src/code/game/g_missile.c, src/code/game/g_combat.c, src/game/g_config.c, tests/test_game_weapon_parity.py, docs/gameplay/cvars.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/ Parity estimate: before 72% -> after 100% for the scoped ten-CVar weapon damage registration and wiring surface. The repo-wide parity estimate remains 98% while the remaining weapon damage/splash/range cvar rows are still being walked in focused batches.

Completed work:

1. Rechecked retail qagame HLIL rows and string/default storage for g_damage_bfg, g_damage_cg, g_damage_g, g_damage_gh, g_damage_hmg, g_damage_gl, g_damage_lg, g_damage_lg_falloff, g_damage_mg, and g_damage_ng.
2. Confirmed the retail default strings (100, 8, 50, 10, 6, 5, 12, and 0) and corrected the ten registration flags to the recovered 0x00040000 | CVAR_GAMERULE pair via GAME_CVAR_FLAG_RETAIL_40000.
3. Revalidated functional wiring through G_InitWeaponConfig, G_ClampModDamage, weapon and missile fire paths, custom-settings mask participation where retail exposes it, G_UpdateCvars, and factory config refreshes.
4. Added focused parity sentinels against the HLIL registration rows/default strings and documented the audited damage tranche for server operators.

Verification:

• python -m pytest tests/test_game_weapon_parity.py -q --tb=short passed: 23 passed.
• python -m pytest tests/test_game_weapon_parity.py tests/test_game_factory_regen_parity.py tests/test_game_ammopack_parity.py tests/test_cvar_console_write_parity.py tests/test_game_module_retail_parity_gate.py tests/test_pmove_selected_cvar_parity.py -q --tb=short passed: 72 passed, 1 skipped.

Task A63: Restore retail parity for second knockback g_* CVar tranche [COMPLETED]

Priority: High Primary areas: src/code/game/g_main.c, src/game/g_config.c, src/code/game/g_combat.c, tests/test_game_weapon_parity.py, docs/gameplay/cvars.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/ Parity estimate: before 70% -> after 100% for the scoped ten-CVar knockback base/scalar surface. The repo-wide parity estimate remains 98% while adjacent weapon cvar families continue to be walked in focused batches.

Completed work:

1. Rechecked retail qagame HLIL rows and string/default storage for g_knockback, g_knockback_bfg, g_knockback_gh, g_knockback_ng, g_knockback_pl, g_knockback_cg, g_knockback_hmg, g_knockback_z, g_knockback_z_self, and g_knockback_cripple.
2. Corrected default spellings for the remaining weapon-specific scalars, vertical boosters, and cripple scalar to match retail byte strings: 1, -5, 24, and 0 instead of source-only float-suffixed forms.
3. Corrected flags: global g_knockback, g_knockback_z, g_knockback_z_self, and g_knockback_cripple now use CVAR_GAMERULE; BFG, grapple, nailgun, proximity launcher, chaingun, and heavy machinegun rows now use CONFIG_CVAR_FLAG_RETAIL_40000 | CVAR_GAMERULE.
4. Revalidated wiring through G_InitKnockbackConfig, G_KnockbackScaleForMOD, and the final g_knockback.value scaling in G_Damage; the later A90 pass removed the source-only vertical boost and cripple-velocity helpers after the retail G_Damage branch was remapped.

Verification:

• python -m pytest tests/test_game_weapon_parity.py -q --tb=short passed: 21 passed.
• python -m pytest tests/test_game_weapon_parity.py tests/test_game_factory_regen_parity.py tests/test_game_ammopack_parity.py tests/test_cvar_console_write_parity.py tests/test_game_module_retail_parity_gate.py tests/test_pmove_selected_cvar_parity.py -q --tb=short passed: 70 passed, 1 skipped.
• git diff --check -- src/code/game/g_main.c src/game/g_config.c tests/test_game_weapon_parity.py docs/gameplay/cvars.md IMPLEMENTATION_PLAN.md passed with only the repository's existing CRLF normalization warnings.

Task A62: Restore retail parity for first ten weapon knockback g_* CVars [COMPLETED]

Priority: High Primary areas: src/game/g_config.c, src/code/game/g_combat.c, tests/test_game_weapon_parity.py, tests/test_game_factory_regen_parity.py, docs/gameplay/cvars.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/ Parity estimate: before 68% -> after 100% for the scoped ten-CVar weapon knockback registration and combat wiring surface. The repo-wide parity estimate remains 98% while adjacent weapon cvar families continue to be walked in focused batches.

Completed work:

1. Rechecked retail qagame HLIL rows and string/default storage for g_knockback_g, g_knockback_mg, g_knockback_sg, g_knockback_gl, g_knockback_rl, g_knockback_rl_self, g_knockback_lg, g_knockback_rg, g_knockback_pg, and g_knockback_pg_self.
2. Corrected the registration default spellings to match retail byte strings: 1, 1.10, 0.90, 1.75, 0.85, and 1.30 instead of source-only float-suffixed forms.
3. Corrected the ten registration flags from plain 0 to the recovered retail 0x00040000 | CVAR_GAMERULE pair, and generalized the local helper to CONFIG_CVAR_FLAG_RETAIL_40000 because the same bit also appears on non-spawn-item retail cvars.
4. Revalidated functional wiring through G_InitKnockbackConfig, G_UpdateKnockbackConfig, factory refreshes, G_UpdateCvars, and G_KnockbackScaleForMOD into G_Damage.

Verification:

• python -m pytest tests/test_game_weapon_parity.py -q --tb=short passed: 19 passed.
• python -m pytest tests/test_game_factory_regen_parity.py tests/test_game_ammopack_parity.py -q --tb=short passed: 29 passed.

Task A61: Restore retail parity for factory item-spawn and respawn g_* CVars [COMPLETED]

Priority: High Primary areas: src/game/g_config.c, src/code/game/g_main.c, src/code/game/g_items.c, src/code/game/g_vote.c, tests/test_game_factory_regen_parity.py, tests/test_game_ammopack_parity.py, docs/gameplay/cvars.md, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/ Parity estimate: before 72% -> after 100% for the scoped ten-CVar factory item-spawn/respawn surface. The repo-wide parity estimate remains 98% because this closes table-flag/default mismatches inside an already mostly wired gameplay seam.

Completed work:

1. Rechecked retail qagame HLIL registration rows and string/default storage for g_ammoPack, g_ammoPackHack, g_ammoRespawn, g_powerupRespawn, g_spawnItemPowerup, g_spawnItemHoldable, g_spawnItemWeapons, g_spawnItemHealth, g_spawnItemArmor, and g_spawnItemAmmo.
2. Corrected the local registration flags: ammo-pack toggles now use CVAR_LATCH | CVAR_GAMERULE, ammo/powerup respawn timers use CVAR_GAMERULE, and the six g_spawnItem* cvars use the recovered 0x00040000 | CVAR_GAMERULE retail bit pair via CONFIG_CVAR_FLAG_RETAIL_40000.
3. Revalidated functional wiring through the factory cvar cache, item spawn gating, global-versus-weapon ammo family switching, ammo/powerup respawn timing, and loadout ammo vote toggles.
4. Added focused parity sentinels against the HLIL table rows/default strings and updated the stale ammo-pack parity expectation that still pinned the pre-fix archive flags.

Verification:

• python -m pytest tests/test_game_factory_regen_parity.py -q --tb=short passed: 25 passed.
• python -m pytest tests/test_game_factory_regen_parity.py tests/test_game_ammopack_parity.py tests/test_bg_misc_validation_fixtures.py tests/test_bg_misc_helper_parity.py tests/test_bg_itemlist_indexes.py tests/test_bg_itemlist_retail_metadata.py tests/test_game_weapon_parity.py tests/test_cvar_console_write_parity.py tests/test_game_module_retail_parity_gate.py -q --tb=short passed: 101 passed, 1 skipped.
• git diff --check -- src/code/game/g_main.c src/game/g_config.c tests/test_game_factory_regen_parity.py tests/test_game_ammopack_parity.py docs/gameplay/cvars.md IMPLEMENTATION_PLAN.md passed with only the repository's existing CRLF normalization warnings.

Task A60: Resolve remaining source-visible cl_* cvar surfaces [COMPLETED]

Priority: Medium Primary areas: src/code/client/cl_main.c, src/code/client/client.h, src/code/client/cl_console.c, src/code/macosx/Q3Controller.m, tests/test_engine_cvar_retail_parity.py, docs/client_cvars.md, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/quakelive_steam.exe_hlil_part04.txt, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/quakelive_steam.exe_hlil_part06.txt, references/reverse-engineering/ghidra/quakelive_steam/decompile_top_functions.c Parity estimate: before 99% -> after 100% for the remaining source-visible cl_* cvar/name inventory. Repo-wide remains 98% pending the active compatibility-only and runtime-evidence gaps.

Completed work:

1. Rechecked the remaining source-visible names against the recovered retail CL_Init cvar slab, string table, dynamic connect path, and companion Ghidra corpus.
2. Kept cl_currentServerAddress as the retail dynamic-only connect cvar: CL_Connect_f sets it from the requested server string, and neither retail nor source registers it in CL_Init.
3. Folded cl_matchmakingProvider, cl_matchmakingPolicy, cl_statsProvider, cl_statsPolicy, cl_socialOverlayProvider, and cl_socialOverlayPolicy into the guarded service-disclosure contract: CVAR_ROM defaults, refresh through CL_RefreshPlatformServiceCvars, and explicit absence from the retail client cvar slab.
4. Removed the non-retail Quake III/PunkBuster/platform carryovers cl_allowDownload, cl_contimestamps, cl_conXOffset, cl_guid, cl_noprint, cl_punkbuster, and cl_showBanner from the checked source surface; console notify drawing now uses the recovered base x-adjust, and the legacy pak-compare fallback no longer exposes a client disable cvar.
5. Added focused regression coverage and refreshed docs/client_cvars.md with the final inventory classification.
6. Verification: python -m pytest tests/test_engine_cvar_retail_parity.py::test_engine_cvar_thirtyninth_client_service_disclosure_tranche_matches_guarded_retail_divergence_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_fortieth_client_remaining_cl_surface_matches_resolved_retail_contracts -q --tb=short -> 2 passed; focused client cl_* parity sweep through the thirty-fourth to fortieth tranches -> 7 passed; python -m pytest tests/test_client_full_parity_gate.py -q --tb=short -> 2 passed, 1 skipped; python -m pytest tests/test_cgame_console_surface_parity.py::test_voice_menu_timer_uses_retail_separate_latch -q --tb=short -> 1 passed; source/tracking inventory reports REMAINING=0; git diff --check on touched files passed with CRLF warnings only.

Task A59: Re-audit retail con_* cvar parity [COMPLETED]

Priority: High Primary areas: src/code/client/cl_console.c, tests/test_engine_cvar_retail_parity.py, tests/test_cl_console_cgame_parity.py, tests/test_engine_command_parity.py, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/quakelive_steam.exe_hlil_part04.txt, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/quakelive_steam.exe_hlil_part06.txt Parity estimate: before 100% -> after 100% for the scoped retail con_* console cvar surface. No production source patch was required because the checked-in registrations and consumers already matched the recovered retail evidence.

Completed work:

1. Rechecked the engine-owned retail evidence for console cvars against metadata.txt, imports.txt, exports.txt, functions.csv, analysis_symbols.txt, and the Binary Ninja HLIL rows around 004b4a30.
2. Confirmed the exhaustive Quake Live retail con_* set contains eight cvars, not ten: con_background, con_height, con_matchlimit, con_noprint, con_opacity, con_scale, con_speed, and con_timestamps.
3. Rejected the tempting Quake III carryovers as non-retail for Quake Live: con_notifytime is absent from the Quake Live HLIL, and scr_conspeed is replaced by the retail con_speed bounded cvar.
4. Added a focused regression guard that now proves the exact retail names, defaults, flags, bounded ranges, HLIL registration rows, and source wiring for every retail con_* cvar, including background selection, console height, find-match limit, print suppression, opacity, scale, scroll speed, and timestamp/cgame physics-time routing.
5. Verification: python -m pytest tests/test_engine_cvar_retail_parity.py::test_console_cvar_surface_matches_retail_hlil tests/test_engine_cvar_retail_parity.py::test_console_cvar_functional_wiring_matches_retail_hlil_surface tests/test_engine_cvar_retail_parity.py::test_engine_cvar_second_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_sixth_client_tranche_matches_retail_contracts -q --tb=short -> 4 passed; python -m pytest tests/test_cl_console_cgame_parity.py tests/test_engine_command_parity.py::test_console_and_alias_command_families_match_retail_wiring -q --tb=short -> 10 passed. A broader full-file sweep of tests/test_engine_cvar_retail_parity.py currently reports 49 passed, 3 failed in unrelated renderer/Steam cvar tranches.

Task A58: Re-audit client/server usercmd movement transport [COMPLETED]

Priority: High Primary areas: src/code/client/cl_input.c, src/code/client/cl_cgame.c, src/code/cgame/cg_predict.c, src/code/cgame/cg_syscalls.c, src/code/qcommon/msg.c, src/code/server/sv_client.c, src/code/server/sv_game.c, tests/test_usercmd_movement_transport_parity.py, docs/reverse-engineering/quakelive_steam_mapping_round_17.md, docs/reverse-engineering/quakelive_steam_mapping_round_57.md, docs/reverse-engineering/quakelive_steam_mapping_round_62.md, docs/reverse-engineering/quakelive_steam_mapping_round_126.md, docs/reverse-engineering/quakelive_steam_mapping_round_277.md, references/analysis/quakelive_symbol_aliases.json Parity estimate: before 100% -> after 100% for the scoped usercmd movement transport seam: client command creation, packet packing, keyed qcommon deltas, server replay into qagame, and cgame prediction access to the host command ring. No production source patch was required.

Completed work:

1. Rechecked the retained retail mappings for CL_CreateCmd, CL_CreateNewCommands, CL_ReadyToSendPacket, CL_WritePacket, CL_SendCmd, MSG_WriteDeltaUsercmdKey, MSG_ReadDeltaUsercmdKey, SV_ClientThink, SV_UserMove, SV_ExecuteClientMessage, QLCGImport_GetCurrentCmdNumber, and QLCGImport_GetUserCmd.
2. Added a focused end-to-end sentinel for the client command path, including final weapon / weaponPrimary / fov storage, serverTime and view-angle capture, primed command generation, packet throttle checks, cl_packetdup resend window, clc_move versus clc_moveNoDelta, checksum-feed keying, ordered MSG_WriteDeltaUsercmdKey, and outPackets bookkeeping.
3. Pinned the qcommon keyed usercmd delta pair against Quake Live extensions: weaponPrimary and fov now stay covered alongside the signed forwardmove, rightmove, and upmove bytes in both changed-field and copied-baseline read paths.
4. Added server replay coverage for SV_UserMove and SV_ClientThink, including delta-message selection, command-count guards, keyed decode, message-ack ping timestamping, pure-client gates, first-command SV_ClientEnterWorld, stale-command suppression, lastUsercmd, qagame G_GET_USERCMD, bot BOTLIB_USER_COMMAND, and GAME_CLIENT_THINK.
5. Added cgame import coverage proving prediction pulls commands through trap_GetCurrentCmdNumber / trap_GetUserCmd, while host-side CL_GetUserCmd still rejects future and overwritten command numbers before copying from the command ring.
6. Verification: python -m pytest tests/test_usercmd_movement_transport_parity.py -q --tb=short -> 5 passed; python -m pytest tests/test_usercmd_movement_transport_parity.py tests/test_engine_client_command_parity.py::test_usercmd_cgame_bridge_matches_retail_weapon_primary_and_fov_bytes tests/test_engine_client_command_parity.py::test_client_input_mapping_round_277_promotes_console_input_and_usercmd_symbols -q --tb=short -> 7 passed; python -m pytest tests/test_usercmd_movement_transport_parity.py tests/test_playerstate_replication.py tests/test_game_active_pmove_wiring_parity.py tests/test_cgame_snapshot_parity.py tests/test_cgame_playerstate_transition_parity.py tests/test_cgame_item_respawn_timer_parity.py tests/test_game_native_export_helper_parity.py tests/test_pmove_validation_fixtures.py tests/test_pmove_air_control_runtime_parity.py tests/test_pmove_jump_timing_parity.py tests/test_pmove_movement_fixtures.py tests/test_pmove_helper_parity.py tests/test_pmove_acceleration_scope_parity.py tests/test_pmove_crouch_time_parity.py tests/test_pmove_crouch_slide_friction_parity.py tests/test_pmove_reload_fallback_parity.py tests/test_pmove_water_scale_parity.py tests/test_bg_playerstate_bridge_parity.py tools/tests/test_pmove_settings_configstring.py tests/test_pmove_selected_cvar_parity.py -q --tb=short -> 199 passed, 107 subtests passed; python -m pytest tests/test_game_module_retail_parity_gate.py tests/test_qcommon_full_parity_gate.py tests/test_client_full_parity_gate.py -q --tb=short -> 6 passed, 3 skipped.

Task A57: Restore retail parity for spawn and sudden-death g_* CVars [COMPLETED]

Priority: High Primary areas: src/code/game/g_main.c, src/game/g_config.c, src/code/game/g_client.c, src/code/game/g_combat.c, src/code/game/g_items.c, src/game/g_match_config.c, docs/gameplay/cvars.md, tests/test_game_factory_regen_parity.py, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/qagamex86.dll.bndb_hlil_part02.txt, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/qagamex86.dll.bndb_hlil_part03.txt Parity estimate: before 74% -> after 100% for the selected ten-cvar surface: g_startingWeapons, g_startingHealth, g_startingHealthBonus, g_startingArmor, g_suddenDeathRespawn, g_suddenDeathRespawnStart, g_suddenDeathRespawnTick, g_suddenDeathRespawnMax, g_suddenDeathRespawnIncrement, and g_suddenDeathRespawnPrint. Repo-wide remains 98% pending the active compatibility-only and runtime-evidence gaps.

Completed work:

1. Rechecked the qagame HLIL registration rows and string/default data for the selected spawn and sudden-death cvars, including the retail g_startingWeapons default mask (3) and the shared CVAR_GAMERULE flag across the sudden-death controller.
2. Corrected g_startingHealth to the retail CVAR_SERVERINFO | CVAR_GAMERULE row, corrected g_startingArmor and g_startingWeapons to CVAR_GAMERULE, and changed g_startingWeapons from source-local 0 to retail 3.
3. Corrected g_startingHealthBonus and g_suddenDeathRespawn in the config cvar registration layer so duplicate registration preserves the recovered CVAR_GAMERULE flags.
4. Corrected g_suddenDeathRespawnStart, g_suddenDeathRespawnTick, g_suddenDeathRespawnMax, g_suddenDeathRespawnIncrement, and g_suddenDeathRespawnPrint to their retail CVAR_GAMERULE rows.
5. Revalidated wiring for spawn loadout masks, starting health/armor fallback, factory config caching, sudden-death match config caching, respawn-delay calculation, centerprint announcements, item respawn suppression, and match state publication.
6. Verification: python -m pytest tests/test_game_factory_regen_parity.py -q --tb=short -> 23 passed; python -m pytest tests/test_game_factory_regen_parity.py tests/test_match_state_configstring.py tests/test_gametype_lifecycle.py tests/test_game_helper_seam_parity.py tests/test_cvar_console_write_parity.py tests/test_game_module_retail_parity_gate.py -q --tb=short -> 61 passed, 6 skipped.

Task A56: Re-audit server/client snapshot playerState transport [COMPLETED]

Priority: High Primary areas: src/code/server/sv_snapshot.c, src/code/client/cl_parse.c, src/code/client/cl_cgame.c, src/code/cgame/cg_syscalls.c, tests/test_playerstate_replication.py, docs/reverse-engineering/quakelive_steam_mapping_round_17.md, docs/reverse-engineering/quakelive_steam_mapping_round_65.md, docs/reverse-engineering/quakelive_steam_mapping_round_127.md, references/analysis/quakelive_symbol_aliases.json Parity estimate: before 100% -> after 100% for the scoped snapshot playerState transport seam: server snapshot playerState build/write, client snapshot parse/ring storage, CL_GetSnapshot exposure, and cgame syscall retrieval. No production source patch was required.

Completed work:

1. Rechecked the retained retail mappings for SV_BuildClientSnapshot, SV_WriteSnapshotToClient, SV_SendClientSnapshot, CL_ParseSnapshot, QLCGImport_GetCurrentSnapshotNumber, and QLCGImport_GetSnapshot against mapping rounds 17, 65, 127, and the committed alias ledger.
2. Expanded tests/test_playerstate_replication.py so the server-side snapshot path now pins SV_GameClientNum copying into frame->ps, own client suppression, view-origin selection, visible-entity gathering, svc_snapshot field order, area-mask write, playerState delta write, and packet-entity emission order.
3. Added client-side snapshot parse coverage for serverCommandSequence, server time, delta metadata, snap flags, areamask, MSG_ReadDeltaPlayerstate, packet-entity parsing, invalid-frame discard, ping calculation from ps.commandTime, backup-ring storage, and cl.newSnapshots.
4. Added cgame import coverage showing trap_GetSnapshot reaches CL_GetSnapshot, which validates the ring entry and copies areamask, playerState, and bounded packet entities into the cgame-visible snapshot.
5. Verification: python -m pytest tests/test_playerstate_replication.py -q --tb=short -> 14 passed; python -m pytest tests/test_playerstate_replication.py tests/test_game_active_pmove_wiring_parity.py tests/test_cgame_snapshot_parity.py tests/test_cgame_playerstate_transition_parity.py tests/test_cgame_item_respawn_timer_parity.py tests/test_game_native_export_helper_parity.py tests/test_pmove_validation_fixtures.py tests/test_pmove_air_control_runtime_parity.py tests/test_pmove_jump_timing_parity.py tests/test_pmove_movement_fixtures.py tests/test_pmove_helper_parity.py tests/test_pmove_acceleration_scope_parity.py tests/test_pmove_crouch_time_parity.py tests/test_pmove_crouch_slide_friction_parity.py tests/test_pmove_reload_fallback_parity.py tests/test_pmove_water_scale_parity.py tests/test_bg_playerstate_bridge_parity.py tools/tests/test_pmove_settings_configstring.py tests/test_pmove_selected_cvar_parity.py -q --tb=short -> 194 passed, 107 subtests passed; python -m pytest tests/test_game_module_retail_parity_gate.py tests/test_qcommon_full_parity_gate.py tests/test_client_full_parity_gate.py -q --tb=short -> 6 passed, 3 skipped. The full server parity gate still reports the unrelated retained SV-G01 Steam GameServer lifecycle/callback/auth artifact lane as failing.

Task A55: Restore retail parity for cosmetic, item-timer, and loadout g_* CVars [COMPLETED]

Priority: High Primary areas: src/code/game/g_main.c, src/code/game/g_local.h, src/game/g_config.c, docs/gameplay/cvars.md, tests/test_game_callvote_option_parity.py, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/qagamex86.dll.bndb_hlil_part02.txt, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/qagamex86.dll.bndb_hlil_part03.txt Parity estimate: before 82% -> after 100% for the selected ten-cvar surface: g_customSettings, g_itemTimers, g_itemHeight, g_forceSmallScoreboardMessage, g_forceSendConfigstring, g_forceAtmosphericEffects, g_forceDmgThroughSurface, g_specItemTimers, g_loadout, and g_infiniteAmmo. Repo-wide remains 98% pending the active compatibility-only and runtime-evidence gaps.

Completed work:

1. Rechecked the qagame HLIL registration rows and string/default data for the selected ten g_* CVars, including the recovered CVAR_GAMERULE bit and the g_itemHeight retail default string at 10086c34 (35).
2. Aligned source registration defaults and flags: g_customSettings now registers as default 0 with CVAR_SERVERINFO; g_itemTimers and g_itemHeight use CVAR_SERVERINFO | CVAR_GAMERULE; the forced cosmetics cvars use their retail no-flag or CVAR_GAMERULE rows; g_loadout and g_infiniteAmmo use their retail game-rule surfaces.
3. Added the missing retail g_specItemTimers registration as a plain qagame cvar, matching the recovered table's default 1 and no-flag row.
4. Revalidated wiring for the selected cvars: custom-settings digest rebuilds, item timer broadcasts and late-join sync, forced-cosmetics configstring updates, loadout vote/cgame mirroring, and infinite-ammo factory/spawn/weapon paths.
5. Updated gameplay cvar notes so the documented defaults/flags distinguish the retail g_itemHeight default (35) from the local invalid-height fallback.
6. Verification: python -m pytest tests/test_game_callvote_option_parity.py -q --tb=short -> 9 passed; python -m pytest tests/test_cvar_console_write_parity.py tests/test_game_callvote_option_parity.py tests/test_game_factory_regen_parity.py tests/test_game_forfeit_parity.py tests/test_game_module_retail_parity_gate.py tests/test_game_helper_seam_parity.py tests/test_game_native_export_helper_parity.py tests/test_game_compact_scoreboard_parity.py -q --tb=short -> 74 passed, 1 skipped.

Task A54: Re-audit qcommon playerState delta-codec wiring [COMPLETED]

Priority: High Primary areas: src/code/qcommon/msg.c, src/code/game/q_shared.h, tests/test_playerstate_replication.py, docs/reverse-engineering/quakelive_steam_mapping_round_57.md, docs/reverse-engineering/qcommon-full-parity-audit-and-implementation-plan-2026-04-10.md, references/analysis/quakelive_symbol_aliases.json Parity estimate: before 100% -> after 100% for the scoped qcommon playerState snapshot replication bridge: the Quake Live-expanded scalar netfield table, signed command-axis mirrors, stats/persistant/ammo/powerup array-mask sections, and server-to-cgame delta round trips. No production source patch was required.

Completed work:

1. Rechecked the retained retail mapping for sub_4D5D50 -> MSG_WriteDeltaPlayerstate and sub_4D66C0 -> MSG_ReadDeltaPlayerstate against the committed alias ledger and mapping round 57.
2. Expanded tests/test_playerstate_replication.py with executable coverage for all four playerState array-mask sections, proving changed stats[], persistant[], ammo[], and powerups[] slots round-trip while unchanged baseline slots survive the delta copy.
3. Added structural coverage for the full Quake Live scalar playerStateFields order and bit widths, including the movement-critical command-time, origin/velocity/viewangle, jump/crouch, weapon-primary, location, fov, and signed forwardmove / rightmove / upmove tail fields.
4. Pinned the signed-byte codec helpers so command-axis mirrors are compared as signed bytes, serialized as unsigned byte payloads, and restored through MSG_SetPlayerStateFieldValue on both changed and copied-from-baseline scalar fields.
5. Verification: python -m pytest tests/test_playerstate_replication.py -q --tb=short -> 10 passed; python -m pytest tests/test_playerstate_replication.py tests/test_game_active_pmove_wiring_parity.py tests/test_cgame_snapshot_parity.py tests/test_cgame_playerstate_transition_parity.py tests/test_cgame_item_respawn_timer_parity.py tests/test_pmove_validation_fixtures.py tests/test_pmove_air_control_runtime_parity.py tests/test_pmove_jump_timing_parity.py tests/test_pmove_movement_fixtures.py tests/test_pmove_helper_parity.py tests/test_pmove_acceleration_scope_parity.py tests/test_pmove_crouch_time_parity.py tests/test_pmove_crouch_slide_friction_parity.py tests/test_pmove_reload_fallback_parity.py tests/test_pmove_water_scale_parity.py tests/test_bg_playerstate_bridge_parity.py tools/tests/test_pmove_settings_configstring.py tests/test_pmove_selected_cvar_parity.py -q --tb=short -> 184 passed, 107 subtests passed; python -m pytest tests/test_game_module_retail_parity_gate.py tests/test_qcommon_full_parity_gate.py -q --tb=short -> 4 passed, 2 skipped.

Task A53: Restore retail ammo_pack item sentinel wiring [COMPLETED]

Priority: High Primary areas: src/code/game/bg_misc.c, src/code/game/g_items.c, src/game/g_config.c, tests/test_game_ammopack_parity.py, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil.txt, references/symbol-maps/qagame.json Parity estimate: before 96% -> after 100% for scoped ammo_pack metadata, grab eligibility, factory-family gating, Pickup_Ammo forwarding, and Add_Ammo sentinel fan-out. Repo-wide remains 98% pending the active compatibility-only and runtime-evidence gaps.

Completed work:

1. Rechecked the qagame item table and symbol-map evidence for ammo_pack, Pickup_Ammo, Add_Ammo, g_ammoPack, g_ammoPackHack, and g_ammoRespawn.
2. Kept the shared item metadata aligned with retail: ammo_pack remains an IT_AMMO item tagged with the WP_NUM_WEAPONS sentinel and the recovered pickup sound, model, icon, pickup name, and quantity.
3. Reshaped game-side ammo pickup wiring so Pickup_Ammo resolves the entity count/item quantity and forwards the item tag once, while Add_Ammo owns the retail WP_NUM_WEAPONS sentinel path that scans owned weapons and adds per-weapon ammo-pack quantities.
4. Confirmed the existing shared grab rule still scans owned weapons below their maximum ammo, and the factory gate still switches between global ammo-pack and weapon-specific ammo families via g_ammoPack / g_ammoPackHack.
5. Added focused regression coverage for the retail evidence, source metadata, grab rule, Add_Ammo sentinel fan-out, Pickup_Ammo forwarding, and factory-family gate.
6. Verification: python -m pytest tests/test_game_ammopack_parity.py -q --tb=short -> 4 passed; python -m pytest tests/test_bg_misc_validation_fixtures.py tests/test_bg_misc_helper_parity.py tests/test_bg_itemlist_indexes.py tests/test_bg_itemlist_retail_metadata.py -q --tb=short -> 47 passed; python -m pytest tests/test_game_factory_regen_parity.py tests/test_game_weapon_parity.py -q --tb=short -> 38 passed; python -m pytest tests/test_game_ammopack_parity.py tests/test_bg_misc_validation_fixtures.py tests/test_bg_misc_helper_parity.py tests/test_bg_itemlist_indexes.py tests/test_bg_itemlist_retail_metadata.py -q --tb=short -> 51 passed.

Task A52: Re-audit cgame prediction pmove replay wiring [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_predict.c, src/code/cgame/cg_playerstate.c, src/code/cgame/cg_snapshot.c, tests/test_cgame_snapshot_parity.py, docs/reverse-engineering/cgame-mapping.md, docs/reverse-engineering/cgame-bg-parity-implementation-plan.md, references/symbol-maps/cgame.json Parity estimate: before 100% -> after 100% for the scoped cgame prediction/playerState replay seam: local pmove_t setup, usercmd replay, trigger/item prediction, mover correction, step smoothing, predictable-event handoff, and committed retail symbol evidence. No production source patch was required.

Completed work:

1. Rechecked the retail cgame prediction map for CG_PredictPlayerState, CG_TouchItem, CG_TouchTriggerPrediction, CG_UpdateStepChange, CG_UpdatePredictedRailFire, and CG_TransitionPlayerState against the committed symbol map and existing source reconstruction.
2. Expanded the cgame snapshot parity sentinel so the full prediction replay pipeline is pinned: projectile-nudge reset, local pmove_settings_t copy, trace/pointcontents callbacks, tracemask/body handling, command overflow guard, latest-command rail replay, next-snapshot seed selection, pmove_msec clamp, fixed-move view-angle update, Pmove, step smoothing, local projectile nudge, trigger prediction, final mover adjustment, and transition handoff.
3. Added structural coverage for item and trigger prediction sidecars, including retail server-authored pickup skips, grab/event/NODRAW/ammo side effects, no-predict and double-touch guards, box-trace trigger collision, teleport hyperspace, jump-pad activation, and jump-pad cache clearing.
4. Added a committed-evidence sentinel tying the source wiring back to references/symbol-maps/cgame.json and the cgame mapping ledgers.
5. Verification: python -m pytest tests/test_cgame_snapshot_parity.py -q --tb=short -> 11 passed; python -m pytest tests/test_cgame_snapshot_parity.py tests/test_cgame_playerstate_transition_parity.py tests/test_cgame_item_respawn_timer_parity.py -q --tb=short -> 23 passed; python -m pytest tests/test_game_active_pmove_wiring_parity.py tests/test_cgame_snapshot_parity.py tests/test_cgame_playerstate_transition_parity.py tests/test_cgame_item_respawn_timer_parity.py tests/test_pmove_validation_fixtures.py tests/test_pmove_air_control_runtime_parity.py tests/test_pmove_jump_timing_parity.py tests/test_pmove_movement_fixtures.py tests/test_pmove_helper_parity.py tests/test_pmove_acceleration_scope_parity.py tests/test_pmove_crouch_time_parity.py tests/test_pmove_crouch_slide_friction_parity.py tests/test_pmove_reload_fallback_parity.py tests/test_pmove_water_scale_parity.py tests/test_bg_playerstate_bridge_parity.py tests/test_playerstate_replication.py tools/tests/test_pmove_settings_configstring.py tests/test_pmove_selected_cvar_parity.py -q --tb=short -> 180 passed, 107 subtests passed; python -m pytest tests/test_game_module_retail_parity_gate.py tests/test_qcommon_full_parity_gate.py -q --tb=short -> 4 passed, 2 skipped.

Task A51: Reconstruct VM bounded cvar registration and integer cache wiring [COMPLETED]

Priority: High Primary areas: src/code/qcommon/cvar.c, src/code/qcommon/qcommon.h, src/code/game/q_shared.h, src/code/client/cl_cgame.c, tests/test_cvar_console_write_parity.py, tests/test_cgame_hud_parity.py, docs/reverse-engineering/quakelive_steam_mapping_round_294.md, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/quakelive_steam.exe_hlil_part04.txt, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/quakelive_steam.exe_hlil_part06.txt Parity estimate: before 97% -> after 98% for scoped core cvar handling/management, VM cvar shadows, and bounded range registration. Repo-wide remains 98% pending the active compatibility-only and runtime-evidence gaps.

Completed work:

1. Rechecked the retail sub_4CCCC0, sub_4CDD30, sub_4CE400, and sub_4CE460 evidence for hex integer parsing, bounded cvar metadata, VM cvar shadow flags, and ranged VM registration.
2. Added Cvar_ParseInteger() and routed cvar creation/mutation integer-cache updates through the retail "0x" / "0x%08x" fallback path.
3. Reconstructed bounded-cvar registration metadata by forcing CVAR_VM_CREATED in Cvar_GetBounded() and adding Cvar_RegisterBounded() for the native cgame range-register import.
4. Added the retail vmCvar_t.flags shadow field and copied cvar flags during VM registration.
5. Updated cgame range registration to call the recovered qcommon wrapper and refreshed focused cvar/HUD parity tests for that source-visible path.
6. Verification: python -m pytest tests/test_cvar_console_write_parity.py -q --tb=short -> 6 passed; python -m pytest tests/test_cvar_console_write_parity.py tests/test_cgame_displaycontext_parity.py::test_register_cvars_publishes_retail_version_and_vote_reset tests/test_cgame_hud_parity.py::test_crosshair_team_health_default_matches_retail_mode_cvar tests/test_qcommon_full_parity_gate.py -q --tb=short -> 10 passed, 1 skipped; python -m pytest tests/test_cvar_alias_console.py -q --tb=short -> 1 passed.

Task A50: Re-audit focused retail g_* vote and complaint cvar parity tranche [COMPLETED]

Priority: Medium Primary areas: src/code/game/g_main.c, src/game/g_config.c, docs/gameplay/cvars.md, tests/test_game_callvote_option_parity.py, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/qagamex86.dll.bndb_hlil_part03.txt, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil_split/qagamex86.dll.bndb_hlil_part02.txt Parity estimate: before 80% -> after 100% for the scoped ten-cvar vote/complaint tranche: g_allowSpecVote, g_allowVote, g_allowVoteMidGame, g_allowForfeit, g_allowKill, g_complaintLimit, g_complaintDamageThreshold, g_voteFlags, g_voteDelay, and g_voteLimit. Repo-wide remains 98% pending the active compatibility-only and runtime-evidence gaps.

Completed work:

1. Rechecked the selected qagame registration rows against the committed HLIL cvar slab and string/default evidence.
2. Corrected g_allowForfeit to the retail 1 default and recovered CVAR_GAMERULE flag, and corrected g_allowKill to the retail CVAR_GAMERULE flag.
3. Corrected the factory fallback constants for g_complaintDamageThreshold and g_complaintLimit to retail 400 and 5, matching the main registration table.
4. Confirmed the existing callvote, spectator vote, mid-game vote, vote-delay, vote-limit, vote-flag, forfeit, self-kill, complaint, and endvote wiring remains attached to the selected cvars.
5. Refreshed gameplay cvar notes and added focused regression coverage for the selected defaults, flags, retail evidence, and behavioral wiring.
6. Verification: python -m pytest tests/test_game_callvote_option_parity.py -q --tb=short -> 7 passed; python -m pytest tests/test_game_callvote_option_parity.py tests/test_game_forfeit_parity.py tests/test_game_factory_regen_parity.py tests/test_game_module_retail_parity_gate.py -q --tb=short -> 35 passed, 1 skipped; python -m pytest tests/test_cvar_console_write_parity.py tests/test_game_callvote_option_parity.py tests/test_game_forfeit_parity.py tests/test_game_factory_regen_parity.py tests/test_game_module_retail_parity_gate.py -q --tb=short -> 39 passed, 1 skipped.

Task A49: Restore retail key item mover wiring [COMPLETED]

Priority: High Primary areas: src/code/game/bg_misc.c, src/code/game/g_items.c, src/code/game/g_mover.c, src/code/game/g_target.c, src/code/cgame/cg_newdraw.c, tests/test_game_key_item_parity.py, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil.txt, references/symbol-maps/qagame.json, references/symbol-maps/cgame.json Parity estimate: before 96% -> after 100% for scoped silver/gold/master key item pickup, reset, keyed mover, remove-keys target, and HUD icon wiring; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked the retail qagame key item records, G_ResetKeyItem, Touch_DoorTriggerKeyed, Touch_ButtonKeyed, and Use_Target_RemoveKeys evidence against the committed HLIL and symbol maps.
2. Restored the missing keyed-button touch path: silver-key buttons now accept silver or master keys, gold-key buttons accept gold or master keys, and keyed buttons only fire from MOVER_POS1.
3. Kept the existing retail-aligned key pickup/reset surface: key pickup sets the carried mask/stat and does not auto-respawn, dropped copies are freed when keys are reset, and target_remove_keys clears carried keys through the reset helper.
4. Added structural coverage for key item metadata, pickup/reset/drop, keyed-door/keyed-button gates, and the CG_PLAYER_HASKEY ownerdraw path.

Task A48: Reconstruct core cvar handling and management wiring [COMPLETED]

Priority: High Primary areas: src/code/qcommon/cvar.c, src/code/game/q_shared.h, tests/test_cvar_console_write_parity.py, docs/reverse-engineering/quakelive_steam_mapping_round_293.md, references/analysis/quakelive_symbol_aliases.json, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/quakelive_steam.exe_hlil_part04.txt, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/quakelive_steam.exe_hlil_part06.txt Parity estimate: before 92% -> after 97% for scoped core cvar handling/management and source-visible WebView cvar-change wiring. Repo-wide remains 98% pending the active compatibility-only and runtime-evidence gaps.

Completed work:

1. Rechecked retail Cvar_Set2, Cvar_Get, Cvar_GetBounded, and QLWebView_PublishCvarChange ownership against the committed alias ledger and HLIL/string evidence.
2. Restored the retail 2048 static cvar slot cap, added the recovered CVAR_GAMERULE flag, and reconstructed the initialization-time GAMERULE console set refusal gate.
3. Added retail cvar flag conflict cleanup for CVAR_CHEAT combined with archive or replicated/protected flags, including the recovered warning strings.
4. Reconstructed the missing Cvar_Set2 publish call sites for user-created cvars and latched values while preserving the retail console write guard surface of ROM, INIT, LATCH, and CHEAT only.
5. Added mapping round 293 and focused regression coverage for the cvar core wiring, while leaving the repo-owned sets, setu, and setcloud compatibility helpers intact.
6. Verification: python -m pytest tests/test_cvar_console_write_parity.py tests/test_client_config_parity.py tests/test_engine_command_parity.py -q --tb=short -> 24 passed; python -m pytest tests/test_qcommon_full_parity_gate.py -q --tb=short -> 2 passed, 1 skipped. A broader tests/test_engine_cvar_retail_parity.py tests/test_qcommon_full_parity_gate.py run reached this cvar work but still failed two unrelated existing platform/Steam-service string checks.

Task A47: Active-client pmove and playerState wiring reconstruction [COMPLETED]

Priority: High Primary areas: src/code/game/g_main.c, src/code/game/g_active.c, src/code/game/g_local.h, tests/test_game_active_pmove_wiring_parity.py, docs/reverse-engineering/qagame-mapping.md Parity estimate: before 99% -> after 100% for the scoped qagame frame-step-to-ClientThink_real movement wiring edge. The broader shared pmove/playerState reconstruction remains 100% after this source-level call-site realignment and guard expansion.

Completed work:

1. Rechecked qagame mappings for G_RunFrame, ClientThink_real, SpectatorThink, ClientImpacts, G_TouchTriggers, SendPendingPredictableEvents, and related active-client movement wiring.
2. Reconstructed the frame entity-loop client slot so bot/synchronous client frames run scheduled thinks inline and dispatch directly to ClientThink_real, matching the recovered retail G_RunFrame shape instead of routing that call-site through the classic source G_RunClient wrapper.
3. Added ClientThink_real to the game-local prototype surface so the recovered direct frame-loop dispatch is explicit and compile-visible.
4. Added focused structural coverage for the full active-client pmove wiring: command-time clamping, spectator scoreboard PMF_NO_MOVE, pmove setup, tracemask selection, fixed-move transport, playerState-to-entityState projection, predictable-event flushing, trigger touch, client impacts, respawn checks, timer actions, and committed symbol-map evidence.
5. Verification: python -m pytest tests/test_game_active_pmove_wiring_parity.py -q --tb=short -> 4 passed; python -m pytest tests/test_game_active_pmove_wiring_parity.py tests/test_game_helper_seam_parity.py tests/test_game_spectator_connection_parity.py -q --tb=short -> 30 passed; python -m pytest tests/test_game_active_pmove_wiring_parity.py tests/test_pmove_validation_fixtures.py tests/test_pmove_air_control_runtime_parity.py tests/test_pmove_jump_timing_parity.py tests/test_pmove_movement_fixtures.py tests/test_pmove_helper_parity.py tests/test_pmove_acceleration_scope_parity.py tests/test_pmove_crouch_time_parity.py tests/test_pmove_crouch_slide_friction_parity.py tests/test_pmove_reload_fallback_parity.py tests/test_pmove_water_scale_parity.py tests/test_bg_playerstate_bridge_parity.py tests/test_cgame_playerstate_transition_parity.py tests/test_playerstate_replication.py tools/tests/test_pmove_settings_configstring.py tests/test_pmove_selected_cvar_parity.py -q --tb=short -> 162 passed, 107 subtests passed; python -m pytest tests/test_game_module_retail_parity_gate.py tests/test_qcommon_full_parity_gate.py -q --tb=short -> 4 passed, 2 skipped.

Task A46: Re-audit sixth focused retail cl_* cvar parity tranche [COMPLETED]

Priority: Medium Primary areas: src/code/client/cl_main.c, src/common/platform/platform_services.c, tests/test_engine_cvar_retail_parity.py, docs/client_cvars.md, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/quakelive_steam.exe_hlil_part04.txt, references/reverse-engineering/ghidra/quakelive_steam/decompile_top_functions.c Parity estimate: before 100% -> after 100% for the scoped ten-cvar service-disclosure tranche: cl_onlineServicesMode, cl_onlineServicesPolicy, cl_onlineServicesParityScope, cl_onlineServicesParityReason, cl_identityBootstrapMode, cl_identityBootstrapPolicy, cl_voiceServiceMode, cl_voiceServicePolicy, cl_workshopProvider, and cl_workshopPolicy. These are intentional non-retail ROM diagnostics for the documented online-services divergence; no production source patch was required. Repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked the selected cl_* service-disclosure cvars against the repository's online-services divergence policy: default disabled or unclassified values, CVAR_ROM flags, and refresh through CL_RefreshPlatformServiceCvars.
2. Confirmed the active labels are sourced from the platform-services descriptor helpers and workshop descriptor helpers rather than being hardcoded as live-service state.
3. Confirmed no selected disclosure cvar appears in the recovered retail CL_Init HLIL/Ghidra evidence, keeping this compatibility surface outside the strict retail cvar slab.
4. Added a focused regression test that pins the defaults, flags, refresh wiring, descriptor ownership, and retail-evidence absence for all ten cvars.
5. Refreshed docs/client_cvars.md with the sixth focused cvar-tranche note.
6. Verification: python -m pytest tests/test_engine_cvar_retail_parity.py::test_engine_cvar_thirtyninth_client_service_disclosure_tranche_matches_guarded_retail_divergence_contracts -q -> 1 passed.

Task A45: Re-audit fifth focused retail cl_* cvar parity tranche [COMPLETED]

Priority: Medium Primary areas: src/code/client/cl_main.c, src/code/client/cl_input.c, src/code/client/cl_parse.c, src/code/ui/ui_main.c, tests/test_engine_cvar_retail_parity.py, docs/client_cvars.md, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/quakelive_steam.exe_hlil_part04.txt, references/hlil/quakelive/uix86.all/uix86.dll_hlil_split/uix86.dll_hlil_part01.txt, references/reverse-engineering/ghidra/quakelive_steam/decompile_top_functions.c, references/reverse-engineering/ghidra/uix86/decompile_top_functions.c Parity estimate: before 100% -> after 100% for the scoped ten-cvar native-UI bridge tranche: cl_maxpackets, cl_packetdup, cl_serverStatusResendTime, cl_maxPing, cl_motdString, cl_downloadItem, cl_downloadName, cl_downloadTime, cl_downloadCount, and cl_downloadSize. No production source patch was required; this pass refreshed and pinned the retail/native UI evidence. Repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked retail registration/default/flag evidence for the selected engine-owned cvars against the committed Quake Live HLIL/Ghidra corpus.
2. Confirmed native uix86 bridge behavior for network preset writes, server-status resend publication, max-ping filtering reads, MOTD reads, and workshop progress reads.
3. Confirmed cl_downloadCount / cl_downloadSize remain UI-facing temp fallbacks for legacy download/QVM paths while the recovered native progress bridge stays keyed by cl_downloadItem plus GetItemDownloadInfo and cl_downloadTime.
4. Added a focused regression test that pins the retail evidence, source-visible registrations, flags, and wiring for all ten cvars.
5. Refreshed docs/client_cvars.md with the fifth focused cvar-tranche note.
6. Verification: python -m pytest tests/test_engine_cvar_retail_parity.py::test_engine_cvar_sixth_client_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_thirtyfourth_client_cl_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_thirtyseventh_client_lifecycle_workshop_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_thirtyeighth_client_native_ui_bridge_tranche_matches_retail_contracts tests/test_ui_menu_files.py::test_ui_retail_download_info_reads_item_progress_through_native_slot -q -> 5 passed. Broader workshop/platform-service verification was not runnable in this worktree because tests/test_client_workshop_bootstrap_parity.py and tests/test_platform_services.py currently fail collection on unrelated indentation errors before the selected tests can execute.

Task A44: Profile and utility pmove/playerState wiring sweep [COMPLETED]

Priority: High Primary areas: src/code/game/g_pmove.c, src/code/game/bg_pmove.c, src/code/cgame/cg_servercmds.c, src/game/g_config.c, src/code/game/g_main.c, tests/test_pmove_selected_cvar_parity.py, docs/reverse-engineering/qagame-mapping.md, docs/reverse-engineering/cgame-bg-parity-implementation-plan.md Parity estimate: before 100% -> after 100% for scoped shared pmove, playerState profile flag ownership, utility cvar transport, and custom-settings wiring. Production source was already aligned with the committed retail evidence; this pass added regression coverage for the remaining profile/utility movement slab.

Completed work:

1. Rechecked the remaining profile and utility pmove_* fields against the retail qagame cvar slab and shared cgame/qagame movement consumers: pmove_AirStepFriction, pmove_AirSteps, pmove_BunnyHop, pmove_CrouchSlide, pmove_CrouchSlideFriction, pmove_CrouchSlideTime, pmove_CrouchStepJump, pmove_DoubleJump, pmove_noPlayerClip, pmove_StepHeight, pmove_StepJump, pmove_velocity_gh, pmove_WaterSwimScale, pmove_WaterWadeScale, pmove_WeaponDropTime, and pmove_WeaponRaiseTime.
2. Expanded tests/test_pmove_selected_cvar_parity.py so this cvar group is pinned from retail registration defaults/flags through factory reset, refresh callback/no-callback ownership, g_pmoveSettings cache assignment, compact and JSON cgame parsing, default shared movement constants, and active movement consumers.
3. Added coverage for the linked playerState and wiring edges: spawn-time PMF_CROUCH_SLIDE / PMF_DOUBLE_JUMP / PMF_AIR_CONTROL seeding, crouch-slide friction/timer consumption, no-player-clip tracemask removal, step/crouch-step gates, water/step globals, grapple pull speed, weapon raise/drop timings, and custom-settings mask participation.
4. Refreshed the qagame and cgame/bg mapping notes with the new evidence lane.
5. Verification: python -m pytest tests/test_pmove_selected_cvar_parity.py -q --tb=short -> 12 passed; python -m pytest tests/test_pmove_validation_fixtures.py tests/test_pmove_air_control_runtime_parity.py tests/test_pmove_jump_timing_parity.py tests/test_pmove_movement_fixtures.py tests/test_pmove_helper_parity.py tests/test_pmove_acceleration_scope_parity.py tests/test_pmove_crouch_time_parity.py tests/test_pmove_crouch_slide_friction_parity.py tests/test_pmove_reload_fallback_parity.py tests/test_pmove_water_scale_parity.py tests/test_bg_playerstate_bridge_parity.py tests/test_cgame_playerstate_transition_parity.py tests/test_playerstate_replication.py tools/tests/test_pmove_settings_configstring.py tests/test_pmove_selected_cvar_parity.py -q --tb=short -> 158 passed, 107 subtests passed; python -m pytest tests/test_game_module_retail_parity_gate.py tests/test_qcommon_full_parity_gate.py -q --tb=short -> 4 passed, 2 skipped.

Task A43: Re-audit fourth focused retail cl_* cvar parity tranche [COMPLETED]

Priority: Medium Primary areas: src/code/client/cl_main.c, src/code/client/cl_input.c, src/code/client/cl_parse.c, src/code/client/cl_cgame.c, src/code/client/cl_scrn.c, src/code/client/cl_ui.c, src/code/qcommon/common.c, src/code/qcommon/cmd.c, src/code/ui/ui_main.c, tests/test_engine_cvar_retail_parity.py, docs/client_cvars.md, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/quakelive_steam.exe_hlil_part04.txt, references/hlil/quakelive/uix86.all/uix86.dll_hlil_split/uix86.dll_hlil_part01.txt, references/reverse-engineering/ghidra/quakelive_steam/decompile_top_functions.c, references/reverse-engineering/ghidra/uix86/decompile_top_functions.c Parity estimate: before 100% -> after 100% for the scoped ten-cvar lifecycle/MOTD/workshop tranche: cl_motd, cl_motdString, cl_timeout, cl_nodelta, cl_debugMove, cl_paused, cl_running, cl_downloadItem, cl_downloadName, and cl_downloadTime. The selected cvar registrations and source-visible owners were already aligned; the expanded workshop verification found and fixed one adjacent callback diagnostic string drift, bringing the scoped workshop handoff evidence from 99% -> 100%. Repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked retail registration/default/flag evidence for the selected CL_Init, CL_InitInput, and common runtime cl_* cvars against the committed HLIL/Ghidra corpus.
2. Confirmed source-visible owners already match retail: MOTD enable/ROM string publication, timeout disconnect checks, delta suppression, movement diagnostics, paused/running ROM state and command routing.
3. Confirmed the retained workshop request cvars stay aligned with the recovered retail/uix86 handoff: cl_downloadItem, cl_downloadName, and cl_downloadTime are seeded from the request owner and consumed by the native UI progress bridge.
4. Restored the retail ItemInstalled_t invalid-app diagnostic string, which reuses the recovered OnDownloadItemResult skip, invalid app id %d text.
5. Added a focused regression test that pins the retail evidence, source registrations, and source-visible wiring for all ten cvars.
6. Refreshed docs/client_cvars.md with the fourth focused cvar-tranche note.
7. Verification: python -m pytest tests/test_engine_cvar_retail_parity.py::test_engine_cvar_sixth_client_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_seventh_client_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_twentieth_client_debug_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_thirtyseventh_client_lifecycle_workshop_tranche_matches_retail_contracts tests/test_client_workshop_bootstrap_parity.py -q -> 8 passed.

Task A42: Deep pmove/playerState mapping sweep [COMPLETED]

Priority: High Primary areas: src/code/game/bg_pmove.c, src/code/game/g_pmove.c, src/code/cgame/cg_servercmds.c, src/code/game/bg_public.h, tests/pmove_validation_harness.c, tests/test_pmove_validation_fixtures.py, tests/test_pmove_selected_cvar_parity.py, tests/test_pmove_helper_parity.py, tests/test_playerstate_replication.py, references/symbol-maps/qagame.json, references/symbol-maps/cgame.json, docs/reverse-engineering/qagame-mapping.md, docs/reverse-engineering/cgame-bg-parity-implementation-plan.md Parity estimate: before 100% -> after 100% for scoped shared pmove, playerState command-gate management, pmove cvar transport, and qagame/cgame prediction wiring. No production source patch was required; this pass added retail-backed fixture and mapping coverage for the fragile remaining edges.

Completed work:

1. Rechecked qagame/cgame PmoveSingle symbol evidence for the playerState command gate: command time advances before PMF_NO_MOVE, while view-angle updates, command mirrors, trace dispatch, and movement side effects remain behind the no-move return.
2. Added an executable PMF_NO_MOVE fixture in tests/pmove_validation_harness.c and tests/test_pmove_validation_fixtures.py to pin that runtime behavior.
3. Validated the selected pmove cvar mapping sentinel, covering representative pmove_* retail names across registration defaults/flags, factory reset, refresh callback ownership, cached settings, compact/JSON cgame parsing, default shared movement constants, active bg_pmove.c consumers, and committed symbol-map evidence.
4. Re-ran the broad movement/playerState suite: python -m pytest tests/test_pmove_validation_fixtures.py tests/test_pmove_air_control_runtime_parity.py tests/test_pmove_jump_timing_parity.py tests/test_pmove_movement_fixtures.py tests/test_pmove_helper_parity.py tests/test_pmove_acceleration_scope_parity.py tests/test_pmove_crouch_time_parity.py tests/test_pmove_crouch_slide_friction_parity.py tests/test_pmove_reload_fallback_parity.py tests/test_pmove_water_scale_parity.py tests/test_bg_playerstate_bridge_parity.py tests/test_cgame_playerstate_transition_parity.py tests/test_playerstate_replication.py tools/tests/test_pmove_settings_configstring.py tests/test_pmove_selected_cvar_parity.py -q --tb=short -> 154 passed, 107 subtests passed.

Task A41: Re-audit third focused retail cl_* cvar parity tranche [COMPLETED]

Priority: Medium Primary areas: src/code/client/cl_main.c, src/code/client/cl_input.c, src/code/client/cl_parse.c, src/code/qcommon/msg.c, tests/test_engine_cvar_retail_parity.py, docs/client_cvars.md, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/quakelive_steam.exe_hlil_part04.txt, references/reverse-engineering/ghidra/quakelive_steam/decompile_top_functions.c Parity estimate: before 100% -> after 100% for the scoped ten-cvar debug/identity/mouse tranche: cl_shownet, cl_showSend, cl_anonymous, cl_platform, cl_maxPing, cl_mouseAccel, cl_mouseAccelDebug, cl_mouseAccelOffset, cl_mouseAccelPower, and cl_mouseSensCap. No source patch was required; this pass refreshed and pinned the retail evidence. Repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked retail CL_Init default values and flags against the committed HLIL/Ghidra evidence for the ten selected client cvars.
2. Confirmed the current source already matches the source-visible retail owners: network/message debug printing, packet-send diagnostics, anonymous auth userinfo, platform ROM marker publication, max-ping timeout filtering, mouse acceleration math, mouse acceleration debug logging, and sensitivity cap behavior.
3. Added a focused regression test that pins the retail evidence, source registrations, and source-visible wiring for all ten cvars.
4. Refreshed docs/client_cvars.md with the third focused cvar-tranche note.
5. Verification: python -m pytest tests/test_engine_cvar_retail_parity.py::test_engine_cvar_thirtysixth_client_debug_identity_tranche_matches_retail_contracts -q -> 1 passed.

Task A40: Re-audit second focused retail cl_* cvar parity tranche [COMPLETED]

Priority: Medium Primary areas: src/code/client/cl_main.c, src/code/client/cl_input.c, tests/test_engine_cvar_retail_parity.py, docs/client_cvars.md, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/quakelive_steam.exe_hlil_part04.txt, references/reverse-engineering/ghidra/quakelive_steam/decompile_top_functions.c Parity estimate: before 100% -> after 100% for the scoped ten-cvar demo/input tranche: cl_avidemo, cl_avidemo_latch, cl_avidemo_mintime, cl_avidemo_maxtime, cl_forceavidemo, cl_anglespeedkey, cl_yawspeed, cl_pitchspeed, cl_run, and cl_freelook. No source patch was required; this pass refreshed and pinned the retail evidence. Repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked retail CL_Init default values and flags against the committed HLIL/Ghidra evidence for the ten selected client demo/input cvars.
2. Confirmed the current source already matches the source-visible retail owners: avidemo latch/start/stop/screenshot/fixed-msec handling, keyboard and joystick angle-speed scaling, run/walk move-speed selection, and freelook center-view/mouse-pitch gating.
3. Added a focused regression test that pins the retail evidence, source registrations, and source-visible wiring for all ten cvars.
4. Refreshed docs/client_cvars.md with the second focused cvar-tranche note.
5. Verification: python -m pytest tests/test_engine_cvar_retail_parity.py::test_engine_cvar_thirtyfifth_client_demo_input_tranche_matches_retail_contracts -q -> 1 passed.

Task A39: Re-audit focused retail s_* sound cvar parity tranche [COMPLETED]

Priority: Medium Primary areas: src/code/client/snd_dma.c, tests/test_engine_cvar_retail_parity.py, tests/test_client_sound_voice_parity.py, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/quakelive_steam.exe_hlil_part04.txt Parity estimate: before 98% -> after 100% for the scoped ten-cvar tranche: s_announcerVolume, s_doppler, s_initsound, s_mixahead, s_mixPreStep, s_musicvolume, s_pvs, s_voiceVolume, s_voiceStep, and s_volume. Repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked retail S_Init default values, flags, and registration order against the committed HLIL evidence for the selected sound cvars.
2. Realigned the retail sound-registration cluster so s_show, s_testsound, and s_volume match the recovered order, while the legacy source-only s_separation/s_khz compatibility cvars now sit outside the retail early s_initsound gate.
3. Restored the retail S_Update_ mix-ahead owner for s_mixahead by removing the older source-only sane/op cap; s_mixPreStep remains wired through S_GetSoundtime and the voice-buffer scheduler.
4. Expanded the focused regression guard so default value, flag set, registration order, and source-visible wiring are pinned for the selected cvars, including s_pvs spatial culling and s_voiceStep voice scheduling.
5. Verification: python -m pytest tests/test_client_sound_voice_parity.py tests/test_engine_cvar_retail_parity.py::test_engine_cvar_eighteenth_sound_tranche_matches_retail_contracts -q --tb=short -> 9 passed.

Task A38: Re-audit focused retail cl_* cvar parity tranche [COMPLETED]

Priority: Medium Primary areas: src/code/client/cl_keys.c, src/code/client/cl_main.c, src/code/client/cl_input.c, src/code/client/cl_cgame.c, src/code/client/cl_scrn.c, tests/test_engine_cvar_retail_parity.py, docs/client_cvars.md, references/hlil/quakelive/quakelive_steam.exe/quakelive_steam.exe_hlil_split/quakelive_steam.exe_hlil_part04.txt, references/reverse-engineering/ghidra/quakelive_steam/decompile_top_functions.c Parity estimate: before 96% -> after 100% for the scoped ten-cvar tranche: cl_allowConsoleChat, cl_demoRecordMessage, cl_freezeDemo, cl_maxpackets, cl_packetdup, cl_timeNudge, cl_autoTimeNudge, cl_quitOnDemoCompleted, cl_serverStatusResendTime, and cl_showTimeDelta. Repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked retail CL_Init default values and flags against the committed HLIL/Ghidra evidence for the ten selected client cvars.
2. Restored the retail cl_allowConsoleChat console-enter owner so bare console text is forced to command form unless that cvar explicitly enables console chat.
3. Added a focused regression test that pins the retail evidence, source registrations, and source-visible wiring for all ten cvars.
4. Refreshed docs/client_cvars.md with the scoped cvar-tranche evidence note.
5. Verification: python -m pytest tests/test_engine_cvar_retail_parity.py::test_engine_cvar_second_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_sixth_client_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_seventh_client_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_eighth_client_input_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_ninth_client_misc_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_twentieth_client_debug_tranche_matches_retail_contracts tests/test_engine_cvar_retail_parity.py::test_engine_cvar_thirtyfourth_client_cl_tranche_matches_retail_contracts -q -> 7 passed.

Task A37: Re-audit pmove factory wiring [COMPLETED]

Priority: High Primary areas: src/code/game/g_pmove.c, src/code/game/g_factory.c, src/game/g_config.c, src/code/game/g_main.c, src/code/cgame/cg_servercmds.c, tools/tests/test_pmove_settings_configstring.py, tests/test_game_factory_regen_parity.py, tests/test_game_weapon_parity.py, tests/test_ui_menu_files.py, tests/test_cgame_displaycontext_parity.py, references/hlil/quakelive/qagamex86.dll/qagamex86.dll.bndb_hlil.txt, references/symbol-maps/qagame.json, docs/reverse-engineering/qagame-mapping.md Parity estimate: before 100% -> after 100% for scoped pmove factory cvar registration, factory reset/apply ordering, cached movement-settings refresh, custom-settings mask participation, and cgame/UI publication wiring.

Completed work:

1. Rechecked the 36-entry retail pmove_* cvar slab at 0x1008F7C4..0x1008FB20 against the current split g_pmove.c owner: default strings, flag groups, callback/no-callback slots, reset defaults, cache fields, and minimum-positive clamps still match the committed HLIL and symbol-map evidence.
2. Rechecked factory selection flow: Factory_Apply resets gameplay config and pmove-owned factory surfaces before overrides, refreshes all VM cvars, updates config mirrors, and only then rebuilds weapon/reload/knockback/ammo and match-factory mirrors. G_UpdateWeaponReloadConfig still pushes reload overrides into the pmove cache and forces a pmove settings refresh.
3. Rechecked custom-settings ownership: air-control, ramp-jump, physics, weapon-switching, no-player-clip, instagib, quad-hog, and grappling-hook bits still mirror the retail G_UpdateCustomSettingsMaskForCvar grouping while pmove_velocity_gh remains distinct from projectile g_velocity_gh.
4. Added a focused regression guard in tests/test_game_factory_regen_parity.py to pin the factory null/active reset sequence, post-override refresh order, and reload-to-pmove handoff.
5. Re-ran the focused wiring suite: python -m pytest tools/tests/test_pmove_settings_configstring.py tests/test_game_factory_regen_parity.py tests/test_game_weapon_parity.py::test_grappling_hook_full_server_and_cgame_wiring_matches_retail tests/test_ui_menu_files.py::test_ui_retail_server_settings_ownerdraw_restored tests/test_ui_menu_files.py::test_game_retail_weapon_reload_configstring_restored tests/test_cgame_displaycontext_parity.py::test_cgame_weapon_reload_configstring_bridge_restored tests/test_cgame_displaycontext_parity.py::test_cgame_server_settings_panel_reconstruction_uses_retail_custom_setting_configstrings -q --tb=short -> 32 passed, 107 subtests passed.

Task A36: Re-audit player movement, player states, and shared wiring [COMPLETED]

Priority: High Primary areas: src/code/game/bg_pmove.c, src/code/game/bg_slidemove.c, src/code/game/q_shared.h, src/code/qcommon/msg.c, src/code/game/g_pmove.c, src/code/cgame/cg_servercmds.c, tests/test_pmove_*.py, tests/test_playerstate_replication.py, references/symbol-maps/qagame.json, references/symbol-maps/cgame.json, docs/reverse-engineering/cgame-bg-parity-implementation-plan.md Parity estimate: before 100% -> after 100% for scoped player movement, playerState layout/replication, pmove settings transport, and cgame/qagame prediction wiring. No source patch was required; this pass refreshed the evidence that the existing retail reconstruction remains intact.

Completed work:

1. Rechecked the current source against the committed pmove/playerState mapping notes and symbol-map evidence for the shared qagame/cgame movement island.
2. Confirmed the retail playerState prefix remains intact, including clientNum at 0x88, location at 0x8c, weaponPrimary, fov, crouchSlideTime, and the signed command-byte mirrors at 0x1dc..0x1de.
3. Confirmed qcommon still delta-replicates the playerState command mirrors, holdable timer stats, movement flags, and jump/crouch timing fields used by retail prediction and demo/HUD consumers.
4. Confirmed qagame still publishes the compact retail pmove settings stream, cgame parses the 33-token retail core plus bounded reconstruction extension, and PmoveSingle consumes the server-seeded movement-profile flags rather than mutating them locally.
5. Re-ran the focused movement/playerState suites: python -m pytest tests/test_pmove_validation_fixtures.py tests/test_pmove_air_control_runtime_parity.py tests/test_pmove_jump_timing_parity.py tests/test_pmove_movement_fixtures.py tests/test_pmove_helper_parity.py tests/test_pmove_acceleration_scope_parity.py tests/test_pmove_crouch_time_parity.py tests/test_pmove_crouch_slide_friction_parity.py tests/test_pmove_reload_fallback_parity.py tests/test_pmove_water_scale_parity.py tests/test_bg_playerstate_bridge_parity.py tests/test_cgame_playerstate_transition_parity.py tests/test_playerstate_replication.py tools/tests/test_pmove_settings_configstring.py -q --tb=short -> 145 passed, 107 subtests passed.
6. Re-ran the shared projection and strict gate sentinels: python -m pytest tests/test_bg_misc_validation_fixtures.py tests/test_bg_misc_runtime_parity.py tests/test_bg_misc_helper_parity.py -q --tb=short -> 39 passed; python -m pytest tests/test_game_module_retail_parity_gate.py tests/test_qcommon_full_parity_gate.py -q --tb=short -> 4 passed, 2 skipped.

Task A33: Audit retail UI feeder parity and wiring [COMPLETED]

Priority: High Primary areas: src/code/ui/ui_main.c, src/code/cgame/cg_main.c, src/ui/*.menu (read-only evidence), references/hlil/quakelive/uix86.all/, references/symbol-maps/ui.json, tests/test_ui_menu_files.py, tests/test_cgame_displaycontext_parity.py Parity estimate: before 86% -> after 100% for scoped feeder ownership and callback wiring after the retail matrix, player-model feeder repair, and cgame scoreboard-family verification. Repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Feeder queue:

• Establish ownership buckets for the 0x00-0x13 Quake Live feeder ids: UI-owned listboxes, cgame-owned scoreboard/team listboxes, and dormant FEEDER_CLANS.
• Verify UI-owned callbacks for FEEDER_HEADS, FEEDER_MAPS, FEEDER_SERVERS, FEEDER_ALLMAPS, FEEDER_PLAYER_LIST, FEEDER_TEAM_LIST, FEEDER_MODS, FEEDER_DEMOS, FEEDER_Q3HEADS, FEEDER_SERVERSTATUS, FEEDER_FINDPLAYER, FEEDER_CINEMATICS, and FEEDER_CVMAPS.
• Verify cgame-owned callbacks for FEEDER_REDTEAM_LIST, FEEDER_BLUETEAM_LIST, FEEDER_SCOREBOARD, FEEDER_ENDSCOREBOARD, FEEDER_REDTEAM_STATS, and FEEDER_BLUETEAM_STATS.
• Keep FEEDER_CLANS as a defined-but-dormant retail id unless new committed evidence shows an active menu, host, or browser contract.
• Run focused feeder parity tests and patch any concrete mismatch surfaced by the matrix.
• Repair FEEDER_HEADS / FEEDER_Q3HEADS to use the validated retail player-model catalog, pass listbox item cvars through feederSelection, and drop the non-retail PunkBuster server-browser text column from the UI feeder.
• Verification: python -m pytest tests/test_ui_menu_files.py::test_ui_retail_feeder_matrix_matches_menu_consumers_and_callback_ownership tests/test_ui_menu_files.py::test_ui_retail_feeder_leaf_callbacks_match_remaining_ui_owned_ids tests/test_ui_menu_files.py::test_ui_retail_callvote_map_feeder_uses_active_map_slab tests/test_ui_menu_files.py::test_ui_retail_clan_feeder_scaffolding_is_removed tests/test_cgame_displaycontext_parity.py::test_cgame_scoreboard_feeder_matrix_owns_all_retail_scoreboard_feeders tests/test_cgame_displaycontext_parity.py::test_cgame_scoreboard_selection_callbacks_restore_cached_team_list_menu_seam
• Verification: Debug Win32 ui.vcxproj and cgamex86.vcxproj builds succeeded; ui.vcxproj still reports three pre-existing C4090 warnings in UI_RunOrdersScript / UI_RunVoiceOrdersScript.

Recent closure

Task A35: Re-audit match exit rules and round-mode wiring [COMPLETED]

Priority: High Primary areas: src/code/game/g_main.c, src/code/game/g_active.c, src/code/game/g_team.c, src/code/game/g_local.h, tests/test_game_exit_rules_parity.py, tests/test_game_round_controller_helper_parity.py, tests/test_game_attack_defend_parity.py, references/hlil/quakelive/qagamex86.dll/, references/symbol-maps/qagame.json, docs/reverse-engineering/qagame-mapping.md Parity estimate: before 96% -> after 100% for scoped match-exit rule, intermission, Red Rover tie, Freeze fallback, and A/D plus CA/Freeze mercy window wiring; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked qagame ScoreIsTied, CheckExitRules, G_ADCheckExitRules, G_CAFZCheckExitRules, and G_RRCheckExitRules against the committed Binary Ninja HLIL and symbol-map entries.
2. Restored Red Rover to the retail non-team leader-score tie path instead of the red/blue team-score tie path.
3. Restored the generic Freeze exit-rule fallback gate to use g_freezeRoundDelay, matching the retail branch in CheckExitRules.
4. Shared the overtime-adjusted exit-limit helper with the A/D and CA/Freeze mercy checks, and removed the source-only CA/Freeze roundlimit leader guard so the red-first/blue-second threshold order matches retail.
5. Refreshed focused structural coverage for the match-exit, round-controller, and Attack and Defend helper seams.

Task A34: Restore retail spectator player-name scorebox wiring [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_newdraw.c, tests/test_cgame_displaycontext_parity.py, references/symbol-maps/cgame.json, docs/reverse-engineering/cgame-mapping.md Parity estimate: before 88% -> after 100% for scoped cgame spectator player-name and scorebox score ownerdraw wiring, including follow-name anchor alignment, first/second-place configstring label ownership, fixed retail truncation thresholds, and missing/forfeit score fallbacks; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked the cgame ownerdraw corridor against the Binary Ninja HLIL for 0x10033BF0, 0x10035BD0, and 0x10035F30, plus the committed Ghidra switch and symbol-map entries.
2. Restored retail anchor alignment for CG_FOLLOW_PLAYER_NAME and CG_FOLLOW_PLAYER_NAME_EX instead of centering inside the authored rect.
3. Routed CG_1ST_PLYR / CG_2ND_PLYR through cached CS_FIRST_PLACE_NAME / CS_SECOND_PLACE_NAME labels, including the tournament - fallback and the fixed 140/132-pixel ellipsis truncation thresholds observed in HLIL.
4. Routed CG_1ST_PLYR_SCORE / CG_2ND_PLYR_SCORE through cached CS_SCORES1 / CS_SCORES2 values with the retail unavailable marker for SCORE_NOT_PRESENT and forfeit scores.
5. Refreshed cgame display-context parity coverage and reran the focused spectator/scorebox suites.

Task A32: Restore retail Flight refuel-rate registration boundary [COMPLETED]

Priority: High Primary areas: src/code/game/g_items.c, src/code/game/g_main.c, src/code/game/g_local.h, references/symbol-maps/qagame.json, tests/test_game_factory_regen_parity.py Parity estimate: before 99% -> after 100% for scoped Flight powerup pickup/fuel timer parity; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked qagame Pickup_Powerup at 0x1004DFE0 against the committed Ghidra decompile and the qagame symbol-map evidence.
2. Confirmed retail keeps g_flightThrust and g_flightRefuelRate registered for cvar/factory parity, but exposes no Flight-specific pickup multiplier.
3. Restored the registered g_flightRefuelRate storage/default while keeping direct quantity * 1000 timer addition for Flight fuel/refill.
4. Added structural coverage so Flight cannot gain a separate refuel-rate consumer while PM_FlyMove remains free of g_flightThrust movement input.

Task A31: Restore retail spectator follow command and fallback state [COMPLETED]

Priority: High Primary areas: src/code/game/g_cmds.c, src/code/game/g_active.c, src/code/game/g_local.h, tests/test_game_spectator_connection_parity.py, docs/reverse-engineering/qagame-spectator-follow-reconstruction-2026-05-22.md Parity estimate: before 88% -> after 96% for scoped qagame spectator follow/stop-follow command handling, HUD auto-follow command compatibility, saved-team restoration, and missing-target fallback; repo-wide remains 98% pending broader runtime and remaining subsystem gaps.

Completed work:

1. Rechecked the qagame symbol map entries for SpectatorThink, SpectatorClientEndFrame, StopFollowing, Cmd_Follow_f, and FollowCycle, plus the cgame symbol-map evidence for the retail follow %d%s HUD auto-follow command shape.
2. Added a shared qagame spectator fallback helper that preserves the local g_teamSpecFreeCam policy while giving stop-follow and missing-target paths a single state transition.
3. Restored StopFollowing to preserve sess.sessionTeam, restore ps.persistant[PERS_TEAM] from it, reset sess.spectatorClient to the caller's own slot, clear PMF_FOLLOW, and set scoreboard flags from the fallback state.
4. Restored Cmd_Follow_f player-string parsing and optional cgame powerup suffix tolerance. A later 2026-05-26 HLIL pass corrected the boundary: follow1 and follow2 are SetTeam tokens, not Cmd_Follow_f tokens.
5. Routed unresolved active-player follow targets, removed POIs, and invalid explicit follow targets through StopFollowing from SpectatorClientEndFrame.
6. Added focused static parity coverage for the stop-follow, follow-command, and end-frame fallback seams.

Task A30: Restore retail scoreboard feeder dispatch and team-row selection [COMPLETED]

Priority: High Primary areas: src/code/cgame/cg_main.c, tests/test_cgame_displaycontext_parity.py Parity estimate: before 94% -> after 100% for scoped scoreboard feeder dispatch, feeder-local row mapping, selection cursor synchronization, and invalid-row guard behavior; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked cgame CG_FeederCount, CG_SetScoreSelection, CG_InfoFromScoreIndex, CG_FeederItemText, and CG_FeederSelection against the Binary Ninja HLIL feeder block and the committed cgame symbol map.
2. Restored retail text-dispatch ownership: red/blue team-list feeders enter the team-list leaves, red/blue stats feeders enter the rich stats leaves, and all other text feeder requests fall back to the race or normal scoreboard leaf instead of being misrouted through team-list fallbacks.
3. Restored explicit team-local row mapping for team scoreboard cursors while keeping cg.selectedScore as the absolute cg.scores row used by cgame.
4. Hardened score-row lookup so invalid feeder rows return no row context instead of indexing past cg.scores.
5. Updated cgame display-context parity tests and reran the focused scoreboard suite.

Task A29: Restore retail pmove_* cvar registration and callback wiring [COMPLETED]

Priority: High Primary areas: src/code/game/g_pmove.c, src/code/game/g_main.c, src/game/g_config.c, tools/tests/test_pmove_settings_configstring.py, tests/test_game_factory_regen_parity.py, references/symbol-maps/qagame.json, docs/reverse-engineering/qagame-mapping.md, docs/reverse-engineering/cgame-bg-parity-implementation-plan.md Parity estimate: before 92% -> after 100% for scoped pmove_* cvar default, flag, callback, cache, and transport wiring; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked the qagame cvar-table slab at 0x1008F7C4..0x1008FB20, confirming the 36 retail pmove_* names, default strings, and four observed registration flag groups.
2. Restored those high-bit retail flags in the split g_pmove.c registration owner instead of registering every pmove cvar with flag 0.
3. Revalidated the surrounding cache, custom-setting, configstring, and cgame compact-parser wiring; no payload order change was needed.
4. Added structural coverage that pins each pmove_* registration's default and flag group against the recovered retail table.
5. Reconstructed the retail callback-backed update path for the pmove cvar slab: all callback-backed entries now update through named mirrors and the three callback-less profile toggles (pmove_AirControl, pmove_CrouchSlide, and pmove_DoubleJump) remain suppressed just like the retail table.
6. Restored the cached lower-bound behavior observed in G_InitPublishedCvarState and the pmove callbacks: pmove_velocity_gh, pmove_JumpVelocityTimeThreshold, and pmove_JumpVelocityTimeThresholdOffset now clamp to retail's 0.001 minimum before publication.
7. Rechecked the adjacent grapple-speed ownership: retail PM_GrappleMove reads the dedicated pmove_velocity_gh cache, while g_velocity_gh remains a separate projectile-speed cvar with the recovered 1800 default. Source now keeps those paths decoupled and updates the grappling-hook custom-settings comparison accordingly.
8. Re-ran the full-name pmove_* sweep and classified pmove_fixed / pmove_msec as legacy fixed-timestep controls outside the 36-entry QL factory-managed tuning slab. Coverage now pins their 0 / 8 defaults, server-side CVAR_SYSTEMINFO registration, cgame mirrors, 8..33 clamp, and shared Pmove chunking behavior.

Task A28: Re-audit retail step/crouch-step/chain-jump takeoff wiring [COMPLETED]

Priority: High Primary areas: src/code/game/bg_pmove.c, references/symbol-maps/qagame.json, references/symbol-maps/cgame.json, docs/reverse-engineering/qagame-mapping.md, docs/reverse-engineering/cgame-mapping.md, docs/reverse-engineering/cgame-bg-parity-implementation-plan.md, tests/test_pmove_helper_parity.py, tests/test_pmove_movement_fixtures.py Parity estimate: before 91% -> after 100% for scoped crouch-step, step-jump, and chain-jump takeoff wiring; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked qagame PM_StepSlideMove at 0x1002EFE0, cgame PM_StepSlideMove at 0x100034B0, and the shared qagame/cgame PM_ApplyJumpTakeoff leaves at 0x1002E2C0 / 0x10002790.
2. Restored the missing normal step-jump takeoff latch so the normal path selects pmove_StepJumpVelocity in the additive retail branches, while the crouch-step fallback continues to use pmove_ChainJumpVelocity and raises only the ramp-suppression latch.
3. Corrected chain-jump mode wiring so PMF_AIR_CONTROL overrides disabled pmove_ChainJump mode 0, the post-offset air-control addend divides by the retail offset threshold, and max clamping applies after additive and ramp-accumulated takeoff velocity.
4. Added executable fixtures that pin mode 0 air-control override, late-window air-control edge values, normal step additive takeoff, crouch-step chain-additive takeoff, step-jump scale-mode takeoff, step-jump air-control edge profiles, and non-ramp max clamping.

Task A27: Pin pmove command/vector math helper contracts [COMPLETED]

Priority: High Primary areas: tests/test_pmove_movement_fixtures.py, tests/test_pmove_helper_parity.py, references/symbol-maps/qagame.json, references/symbol-maps/cgame.json, docs/reverse-engineering/qagame-mapping.md, docs/reverse-engineering/cgame-mapping.md, docs/reverse-engineering/cgame-bg-parity-implementation-plan.md Parity estimate: before 95% -> after 100% for scoped executable evidence coverage of the retail PM_ClipVelocity, PM_Accelerate, PM_CmdScale, and PM_SetMovementDir math helpers; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked qagame PM_ClipVelocity, PM_Accelerate, PM_CmdScale, and PM_SetMovementDir against the cgame twins and the shared source helpers reached by walk, air, water, ladder, fly, and noclip movement.
2. Added executable fixtures for overbounce clip branches, acceleration clamp and overspeed no-op behavior, zero/axial/2D-diagonal/3D-diagonal command scaling, and the full movementDir 0..7 ring plus idle side-strafe snaps.
3. Updated structural tests, symbol-map comments, and mapping notes so the command/vector layer beneath the movement leaves is source-validated.

Task A26: Pin low-level pmove animation helper gates [COMPLETED]

Priority: High Primary areas: tests/test_pmove_movement_fixtures.py, tests/test_pmove_helper_parity.py, references/symbol-maps/qagame.json, references/symbol-maps/cgame.json, docs/reverse-engineering/qagame-mapping.md, docs/reverse-engineering/cgame-mapping.md, docs/reverse-engineering/cgame-bg-parity-implementation-plan.md Parity estimate: before 96% -> after 100% for scoped executable evidence coverage of the retail low-level pmove animation start/continue/force gates; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked qagame PM_StartTorsoAnim, PM_ContinueLegsAnim, and PM_ForceLegsAnim against the cgame twins plus the source-local PM_StartLegsAnim / PM_ContinueTorsoAnim call chain they expose.
2. Added executable fixtures for torso toggle-bit writes, PM_DEAD suppression, legs high-priority timer no-ops, current-animation no-ops, torso high-priority timer no-ops, and the PM_ForceLegsAnim timer-clear behavior before the live/dead start gate.
3. Updated structural tests, symbol-map comments, and mapping notes so the animation helpers beneath jump, footstep, weapon, and gesture paths are source-validated.

Task A25: Pin PM_AddEvent and PM_AddTouchEnt queue wiring [COMPLETED]

Priority: High Primary areas: tests/test_pmove_movement_fixtures.py, tests/test_pmove_helper_parity.py, references/symbol-maps/qagame.json, references/symbol-maps/cgame.json, docs/reverse-engineering/qagame-mapping.md, docs/reverse-engineering/cgame-mapping.md, docs/reverse-engineering/cgame-bg-parity-implementation-plan.md Parity estimate: before 97% -> after 100% for scoped executable evidence coverage of the retail pmove predictable-event and touch-entity queue helpers; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked qagame PM_AddEvent / PM_AddTouchEnt at 0x1002DAF0 / 0x1002DB20, cgame PM_AddEvent / PM_AddTouchEnt at 0x10001FC0 / 0x10001FF0, and their downstream use from trace/event leaves.
2. Added executable pmove fixtures for the predictable-event two-slot wrap with zero event parms and for the touch accumulator's ENTITYNUM_WORLD, duplicate-suppression, and MAXTOUCH cap gates.
3. Updated structural tests, symbol-map comments, and mapping notes so the low-level queues behind ground contacts, weapon events, water events, and prediction are source-validated.

Task A24: Pin PM_DropTimers misc and animation timer decay [COMPLETED]

Priority: High Primary areas: tests/test_pmove_movement_fixtures.py, tests/test_pmove_helper_parity.py, references/symbol-maps/qagame.json, references/symbol-maps/cgame.json, docs/reverse-engineering/qagame-mapping.md, docs/reverse-engineering/cgame-mapping.md, docs/reverse-engineering/cgame-bg-parity-implementation-plan.md Parity estimate: before 96% -> after 100% for scoped executable evidence coverage of the retail PM_DropTimers misc-time, animation-timer, and crouch-slide decay behavior; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked qagame PM_DropTimers at 0x10031E30, cgame PM_DropTimers at 0x10006300, and the PmoveSingle spectator/noclip/tail call sites against the committed pmove evidence.
2. Added executable fixture coverage for partial pm_time decay, exact/overrun misc timer expiry, PMF_ALL_TIMES clearing while preserving unrelated flags, independent legs/torso animation timer clamps, and the existing crouch-slide ground-plane decay gate.
3. Updated structural tests, symbol-map comments, and mapping notes so the timer helper is covered as a first-class reconstructed source leaf.

Task A23: Pin PM_TorsoAnimation ready-idle wiring [COMPLETED]

Priority: High Primary areas: tests/test_pmove_movement_fixtures.py, tests/test_pmove_helper_parity.py, references/symbol-maps/qagame.json, references/symbol-maps/cgame.json, docs/reverse-engineering/qagame-mapping.md, docs/reverse-engineering/cgame-mapping.md, docs/reverse-engineering/cgame-bg-parity-implementation-plan.md Parity estimate: before 97% -> after 100% for scoped executable evidence coverage of the retail PM_TorsoAnimation ready-state idle split and inherited timer/dead-player suppression behavior; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked qagame PM_TorsoAnimation at 0x1002DB80, cgame PM_TorsoAnimation at 0x10002050, and the surrounding PmoveSingle tail calls against the committed Ghidra function lattice.
2. Added executable pmove fixture coverage for the ready gauntlet TORSO_STAND2 branch, the default ready TORSO_STAND branch, the high-priority torso-timer no-op, the non-ready weaponstate no-op, and the inherited PM_DEAD animation-start suppression.
3. Updated structural tests, symbol-map comments, and pmove mapping notes so the post-weapon torso-idle helper is pinned by source behavior rather than only by normalized names.

Task A22: Pin PM_Animate command-button source behavior [COMPLETED]

Priority: High Primary areas: tests/test_pmove_movement_fixtures.py, references/symbol-maps/qagame.json, references/symbol-maps/cgame.json, docs/reverse-engineering/qagame-mapping.md, docs/reverse-engineering/cgame-mapping.md, docs/reverse-engineering/cgame-bg-parity-implementation-plan.md Parity estimate: before 96% -> after 100% for scoped executable evidence coverage of the retail PM_Animate command-button priority, torso-timer gate, and predictable-event behavior; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked qagame PM_Weapon, PM_Animate, and PmoveSingle tail ordering against the committed qagame/cgame Ghidra evidence and found no production-source mismatch in the dispatcher tail.
2. Added an executable pmove fixture for the shared PM_Animate leaf covering gesture priority over voice commands, the long gesture timer plus EV_TAUNT event, the retail voice-command priority order, the 600 ms voice timer, and the existing-torso-timer suppression gate.
3. Updated the symbol maps and reverse-engineering notes so the mapped animation/timer helper records the tested source behavior instead of relying only on structural matching.

Task A21: Pin 3D water and ladder pmove leaves with executable fixtures [COMPLETED]

Priority: High Primary areas: tests/test_pmove_movement_fixtures.py, references/symbol-maps/qagame.json, references/symbol-maps/cgame.json, docs/reverse-engineering/cgame-mapping.md, docs/reverse-engineering/cgame-bg-parity-implementation-plan.md Parity estimate: before 94% -> after 100% for scoped executable evidence coverage of the retail water-jump, water-idle, ladder-probe, and ladder-move leaves; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked qagame PM_CheckDuck, PM_Footsteps, and PM_Weapon against the committed qagame/cgame Ghidra decompiles and found no production-source mismatch in that tail corridor.
2. Extended the pmove movement fixture harness with deterministic trace and pointcontents probes for the adjacent 3D movement leaves.
3. Added executable coverage for the retail MASK_PLAYERSOLID + SURF_LADDER ladder probe, the 0.66 * speed ladder climb/descent clamp, the two-deep water-jump solid-lip/clearance sequence, and the idle -60 water sink fallback through PM_WaterMove.

Task A20: Rewire pmove_ChainJump as the retail jump velocity mode [COMPLETED]

Priority: High Primary areas: src/code/game/bg_public.h, src/code/game/bg_pmove.c, src/code/game/g_pmove.c, src/code/cgame/cg_servercmds.c, references/symbol-maps/qagame.json, references/symbol-maps/cgame.json, docs/reverse-engineering/qagame-mapping.md, docs/reverse-engineering/cgame-mapping.md, tests/test_pmove_helper_parity.py, tests/test_pmove_movement_fixtures.py, tools/tests/test_pmove_settings_configstring.py Parity estimate: before 97% -> after 100% for scoped pmove_ChainJump mode transport and takeoff-branch wiring; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked the qagame cvar refresh evidence where pmove_ChainJump's integer slot is copied into the cached jump velocity selector, then matched it against the recovered PM_ApplyJumpTakeoff velocity-mode branches.
2. Promoted pmove_settings_t.chainJump from a boolean to an integer mode and carried that value through server cvar caching, compact configstring serialization, cgame compact parsing, and the legacy JSON fallback parser.
3. Updated the shared takeoff selector so mode 0 disables chain-jump timing, mode 1 uses the gradient scaler, mode 2 reaches the additive branch, and PMF_AIR_CONTROL still selects the retail air-control additive branch.
4. Added executable pmove fixtures that pin the vertical takeoff results for modes 0, 1, and 2, plus the air-control override path.

Task A19: Restore retail pmove jump-takeoff velocity modes [COMPLETED]

Priority: High Primary areas: src/code/game/bg_pmove.c, src/code/game/bg_pmove_jump.h, references/symbol-maps/qagame.json, references/symbol-maps/cgame.json, docs/reverse-engineering/qagame-mapping.md, docs/reverse-engineering/cgame-mapping.md, docs/reverse-engineering/cgame-bg-parity-implementation-plan.md, tests/test_jump_velocity_scaling.py, tests/test_pmove_helper_parity.py, tests/test_pmove_jump_timing_parity.py, tests/test_crouch_step_prediction.py, tests/test_pmove_validation_fixtures.py, tests/pmove_validation_harness.c Parity estimate: before 96% -> after 100% for scoped PM_ApplyJumpTakeoff vertical velocity-mode reconstruction; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked qagame PM_ApplyJumpTakeoff at 0x1002E2C0, cgame PM_ApplyJumpTakeoff at 0x10002790, and PM_LoadMoveTuningConstants against the committed HLIL and Ghidra evidence.
2. Replaced the source-only wrapper-side pmove_StepJumpVelocity addition with the retail takeoff velocity-mode block: normal chain jumps use the gradient scaler, PMF_AIR_CONTROL switches to the additive chain/step branch, and ramp-jump accumulation applies to vertical velocity before the max clamp.
3. Updated structural, helper, and executable pmove fixtures to pin the gradient cmd.serverTime - jumpTime behavior and the crouch-step fallback velocity produced by the shared takeoff leaf.

Task A18: Align pmove playerState byte mirrors and item timer stats [COMPLETED]

Priority: High Primary areas: src/code/game/q_shared.h, src/code/qcommon/msg.c, src/code/game/bg_pmove.c, src/code/game/bg_public.h, src/code/game/g_active.c, src/code/game/g_team.c, src/code/cgame/cg_newdraw.c, references/symbol-maps/qagame.json, references/symbol-maps/cgame.json, docs/reverse-engineering/cgame-bg-parity-implementation-plan.md, tests/test_playerstate_replication.py, tests/test_invulnerability_move_parity.py, tests/test_pmove_movement_fixtures.py Parity estimate: before 92% -> after 100% for scoped pmove playerState sidecar layout, command-mirror transport, and progress-backed holdable timer wiring; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked the retail engine playerState netfield table plus cgame/qagame PmoveSingle decompiles, confirming clientNum at 0x88, location at 0x8c, and signed command mirrors at byte offsets 0x1dc..0x1de.
2. Rebuilt the shared playerState_t prefix to match those offsets, moved source-only local sidecars after the retail replicated prefix, and taught MSG_WriteDeltaPlayerstate / MSG_ReadDeltaPlayerstate to encode the command mirrors as one-byte fields without clobbering adjacent bytes.
3. Reconstructed the progress-backed holdable timer as retail stats slots STAT_PLAYER_ITEM_TIME_MAX, STAT_PLAYER_ITEM_TIME, and STAT_PLAYER_ITEM_RECHARGE, including the PmoveSingle recharge clamp and the invulnerability move stale/clear gates.
4. Added server location mirroring into ps.location, moved the CG_PLAYER_ITEM overlay to the stats-backed timer slots, and pinned the recovered offsets and round-trip behavior with executable replication tests.

Task A17: Remove source-only replicated ground-trace cache [COMPLETED]

Priority: High Primary areas: src/code/game/q_shared.h, src/code/qcommon/msg.c, src/code/game/bg_pmove.c, references/symbol-maps/qagame.json, references/symbol-maps/cgame.json, docs/reverse-engineering/qagame-mapping.md, docs/reverse-engineering/cgame-mapping.md, docs/reverse-engineering/cgame-bg-parity-implementation-plan.md, tests/test_pmove_helper_parity.py, tests/test_pmove_jump_timing_parity.py Parity estimate: before 99% -> after 100% for scoped playerState ground-trace layout and PM_GroundTrace storage parity; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked cgame PM_GroundTrace at 0x100053A0, qagame PM_GroundTrace at 0x10030ED0, and the adjacent PM_CheckDuck playerState offset evidence against the committed Ghidra decompile.
2. Removed the reconstruction-only groundTraceHistory* and groundTraceLatest* playerState fields plus their delta replication, leaving the retail groundEntityNum store as the only replicated ground-trace state.
3. Rewired ramp-jump reconstruction to consult the current frame-local pml.groundTrace normal instead of a non-retail playerState cache, and updated symbol-map/docs/tests to pin the layout boundary.

Task A16: Restore autoHop-only held-jump release wiring [COMPLETED]

Priority: High Primary areas: src/code/game/bg_pmove.c, src/code/game/bg_slidemove.c, references/symbol-maps/qagame.json, references/symbol-maps/cgame.json, docs/reverse-engineering/qagame-mapping.md, docs/reverse-engineering/cgame-mapping.md, docs/reverse-engineering/cgame-bg-parity-implementation-plan.md, tests/test_pmove_helper_parity.py, tests/test_step_jump_gate_parity.py, tests/test_pmove_movement_fixtures.py Parity estimate: before 99% -> after 100% for scoped PM_CheckJump / PM_CanStepJump held-release parity; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked qagame PM_CanStepJump at 0x1002E510 and PM_CheckJump at 0x1002E590, plus the cgame twins at 0x100029E0 and 0x10002A60, against the committed HLIL and compact pmove-setting parse order.
2. Restored the retail release-bypass split: autoHop can bypass the held-jump release requirement unless PMF_REQUIRE_JUMP_RELEASE is set, while bunnyHop remains separate movement tuning and no longer bypasses release.
3. Added structural and executable fixture coverage so both normal jump and step-jump gates reject held input when only bunny-hop tuning is enabled.

Task A15: Restore retail step-jump crouch fallback wiring [COMPLETED]

Priority: High Primary areas: src/code/game/bg_slidemove.c, src/code/game/bg_pmove.c, references/symbol-maps/qagame.json, references/symbol-maps/cgame.json, docs/reverse-engineering/qagame-mapping.md, docs/reverse-engineering/cgame-mapping.md, docs/reverse-engineering/cgame-bg-parity-implementation-plan.md, tests/test_step_jump_gate_parity.py, tests/test_pmove_validation_fixtures.py, tests/pmove_validation_harness.c, tests/test_crouch_step_prediction.py, tests/test_pmove_helper_parity.py Parity estimate: before 98% -> after 100% for scoped PM_StepSlideMove step-jump gate and crouch-fallback parity; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked qagame PM_StepSlideMove at 0x1002EFE0, cgame PM_StepSlideMove at 0x100034B0, and the adjacent PM_CanStepJump / PM_CanCrouchStepJump leaves against the committed HLIL plus qagame Ghidra decompile.
2. Replaced the source-only recent-ground-history step-jump delay helper with the retail cmd.serverTime - jumpTime >= jumpTimeDeltaMin gate.
3. Restored the retail branch split where the general PM_CanStepJump recheck launches the normal step jump, while the failed-general path can still run the crouch-step fallback through PM_CanCrouchStepJump and the shrunken-box clearance trace.
4. Added direct and executable fixtures for the crouch-step fallback, including the no-upmove case that retail permits after the general gate rejects the command, and corrected the cgame PM_StepSlideMove symbol signature back to the one-argument retail boundary.

Task A14: Restore exact invulnerability pmove timer gate [COMPLETED]

Priority: High Primary areas: src/code/game/bg_pmove.c, references/symbol-maps/qagame.json, references/symbol-maps/cgame.json, docs/reverse-engineering/cgame-bg-parity-implementation-plan.md, tests/test_invulnerability_move_parity.py, tests/test_pmove_movement_fixtures.py Parity estimate: before 99% -> after 100% for scoped PM_InvulnerabilityMove timer-gate and holdable-clear parity; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked qagame PM_InvulnerabilityMove at 0x1002FA80 and cgame PM_InvulnerabilityMove at 0x10003F50 against the committed HLIL, focusing on the held-item dispatch predicate and timer-collapse block.
2. Restored the exact retail stale-timer gate STAT_PLAYER_ITEM_RECHARGE != 0 && STAT_PLAYER_ITEM_TIME == 0, allowing negative countdown remnants to enter the active decay path instead of being filtered by a broader signed range test.
3. Moved the pending holdable-slot clear under the timer clamp and tied it to STAT_PLAYER_ITEM_RECHARGE == 0, matching the retail collapse point.
4. Added structural and executable pmove coverage for exact zero with a max timer, negative decay with a max timer, and zero with no max timer.

Task A13: Remove source-only flight-thrust pmove override [COMPLETED]

Priority: High Primary areas: src/code/game/bg_pmove.c, src/code/game/bg_public.h, src/code/game/g_main.c, src/code/game/g_pmove.c, src/code/cgame/cg_servercmds.c, references/symbol-maps/qagame.json, references/symbol-maps/cgame.json, docs/reverse-engineering/cgame-bg-parity-implementation-plan.md, tests/test_pmove_helper_parity.py, tests/test_game_factory_regen_parity.py, tools/tests/test_pmove_settings_configstring.py Parity estimate: before 99% -> after 100% for scoped PM_FlyMove and g_flightThrust wiring parity; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked qagame PM_FlyMove at 0x1002FA20 and cgame PM_FlyMove at 0x10003EF0 against the committed HLIL; both retail bodies run PM_Friction, PM_BuildWishMove3D, PM_Accelerate(..., 8), and PM_StepSlideMove(qfalse) without a flight-thrust branch.
2. Confirmed g_flightThrust is still a retail qagame cvar table entry with default 1200, but its vmCvar storage is not referenced by the shared pmove leaf or pmove settings transport.
3. Removed the source-only flightThrust field from pmove_settings_t, the compact/JSON pmove settings payload extension, and PM_FlyMove, while keeping the retail cvar registration.
4. Updated symbol-map/docs evidence and regression tests to pin the no-override movement leaf and cvar registration boundary.

Task A12: Restore invulnerability-aware dead pmove collision gate [COMPLETED]

Priority: High Primary areas: src/code/game/bg_pmove.c, references/symbol-maps/qagame.json, references/symbol-maps/cgame.json, docs/reverse-engineering/qagame-mapping.md, docs/reverse-engineering/cgame-mapping.md, docs/reverse-engineering/cgame-bg-parity-implementation-plan.md, tests/test_pmove_helper_parity.py, tests/test_pmove_movement_fixtures.py Parity estimate: before 99% -> after 100% for scoped PmoveSingle dead-player trace-mask parity; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked qagame PmoveSingle at 0x10031FA0 and cgame PmoveSingle at 0x10006470 against the committed Ghidra evidence, focusing on the top tracemask gate before command-time handling.
2. Restored the retail condition that clears CONTENTS_BODY from the pmove trace mask only when STAT_HEALTH <= 0 and PW_INVULNERABILITY is not active, preserving body collision for invulnerable dead-player states.
3. Updated symbol-map and mapping documentation to record the paired health/powerup gate.
4. Added structural and executable pmove coverage for the dispatcher ordering and both dead-player tracemask outcomes.

Task A11: Move pmove profile-bit ownership back to spawn wiring [COMPLETED]

Priority: High Primary areas: src/code/game/g_client.c, src/code/game/g_pmove.c, src/code/game/bg_pmove.c, src/code/game/g_local.h, references/symbol-maps/qagame.json, references/symbol-maps/cgame.json, docs/reverse-engineering/qagame-mapping.md, docs/reverse-engineering/cgame-bg-parity-implementation-plan.md, tests/test_pmove_helper_parity.py, tests/test_pmove_crouch_time_parity.py, tests/test_pmove_crouch_slide_friction_parity.py, tests/pmove_validation_harness.c Parity estimate: before 99% -> after 100% for scoped pmove profile-flag ownership and prediction wiring parity; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked qagame ClientSpawn at 0x1003BC30 and shared PmoveSingle at qagame 0x10031FA0 / cgame 0x10006470 against the committed Ghidra evidence.
2. Restored the retail ownership split where spawn seeds PMF_CROUCH_SLIDE, PMF_DOUBLE_JUMP, and PMF_AIR_CONTROL from the active movement profile, while PmoveSingle consumes those replicated bits instead of deriving them from settings every frame.
3. Rewired crouch-slide friction and air double-jump acceptance to respect the profile flags directly, preserving the existing tuning fields for friction strength, jump velocity, and prediction-side settings.
4. Updated symbol-map/docs evidence and regression fixtures to lock the spawn profile helper, shared pmove dispatcher order, crouch-slide friction gate, and double-jump latch behavior.

Task A10: Reconstruct compact pmove settings configstring wiring [COMPLETED]

Priority: High Primary areas: src/code/game/g_pmove.c, src/code/cgame/cg_servercmds.c, references/symbol-maps/cgame.json, docs/reverse-engineering/cgame-mapping.md, docs/reverse-engineering/cgame-bg-parity-implementation-plan.md, tests/test_pmove_helper_parity.py, tools/tests/test_pmove_settings_configstring.py Parity estimate: before 99% -> after 100% for scoped pmove settings transport/source wiring parity; repo-wide remains 98% pending the active portability/runtime-evidence gaps.

Completed work:

1. Rechecked CG_ParsePmoveConfigString at 0x10048F30 against the committed cgame Ghidra/HLIL evidence and mapped the retail 33-token compact numeric payload order, including the non-negative clamps on the recovered threshold and velocity fields.
2. Replaced the reconstruction-only JSON CS_PMOVE_SETTINGS server payload with the compact retail token stream from G_PmoveSerializeSettings, while retaining current source-only fields as a trailing extension after the retail core.
3. Rewired cgame parsing so compact payloads take the retail path, JSON remains only as a backward-compatible fallback for older local artifacts, and reload timing stays owned by the dedicated compact reload configstring.
4. Updated symbol-map/docs evidence and regression tests to pin the recovered token order, clamp behavior, extension boundary, and JSON fallback.

Task A9: Contain inherited Quake III service endpoints behind the online-services policy [COMPLETED]

Priority: High Primary areas: src/code/qcommon/qcommon.h, src/code/client/cl_main.c, src/code/server/, src/common/platform/platform_config.h, docs/reverse-engineering/network-handshake.md, docs/reverse-engineering/quakelive_steam_mapping_round_73.md, docs/steam_platform_abstraction.md, tests/test_platform_services.py Parity estimate: before 98% -> after 98% repo-wide; scoped network-service policy parity improves from 82% -> 94%.

Completed work:

1. Confirmed from the retail Quake Live evidence that the retained game port remains 27960, while the inherited Quake III update, master, and authorize hostnames are not retail QL Steam service evidence.
2. Added QL_ENABLE_LEGACY_Q3_SERVICES, defaulted it off, and forced it off whenever QL_BUILD_ONLINE_SERVICES=0, so default builds no longer resolve the retired *.quake3arena.com hosts.
3. Kept PORT_SERVER and the LAN server-port scan range unconditional, while quarantining UPDATE_SERVER_NAME, MASTER_SERVER_NAME, AUTHORIZE_SERVER_NAME, PORT_MASTER, PORT_UPDATE, and PORT_AUTHORIZE behind the legacy-service gate.
4. Restored the retail sv_master heartbeat gate as sv_masterAdvertise and left the older sv_master1 through sv_master5 array as an explicitly configured compatibility lane.
5. Replaced default build calls into the legacy Q3 MOTD, master query, authorize, getIpAuthorize, and remote-ban flows with policy-visible stubs or local challenge responses, and covered the boundary with focused tests.

Task A8: Re-audit renderer source and host wiring, then remove the non-retail FontStash prebuild [COMPLETED]

Priority: High Primary areas: src/code/renderer/tr_font.c, docs/reverse-engineering/renderer-wiring-reverse-engineering-round-2026-05-20.md, docs/reverse-engineering/renderer-full-parity-audit-and-implementation-plan-2026-04-09.md, docs/reverse-engineering/renderer-host-text-core-ownership-2026-04-10.md, docs/platform/retail-font-stack.md, tests/test_renderer_host_text_core_parity.py Parity estimate: before 99% -> after 100% for scoped renderer source/wiring parity; repo-wide remains 98% pending fresh runtime evidence.

Completed work:

1. Re-ran the renderer source/wiring evidence walk against the committed quakelive_steam.exe Ghidra/HLIL corpus, current symbol aliases, renderer parity gates, and the tracked renderer/module runtime artifacts.
2. Confirmed that scene, world/model, shader, post-process, memory-image, export ABI, Win32 host glue, and native UI/cgame host-text import ownership still do not need a new renderer file-level gap note.
3. Isolated one hidden retained-host-text mismatch: source eagerly prebuilt every byte glyph for every retained FontStash face during R_InitFontStash, while retail HLIL creates the atlas, installs R_fonsErrorCallback, loads the five faces, and leaves glyph population lazy through the draw/measure cache path.
4. Removed the eager R_PrebuildFontStashAtlas startup sweep from tr_font.c, added focused source/docs coverage so the lazy retained-atlas ownership stays pinned, and documented that the older R_fonsErrorCallback retail-module artifact is now stale RW-G04 evidence until the module runtime probe is rerun.
5. Followed up on the same FontStash overflow lane by preserving old atlas pixels and cached glyph UVs during R_ResizeFontStashAtlas expansion, leaving full glyph-cache clearing reserved for the retail max-size flush path.
6. Matched the retail *fontstash texture callback by uploading the retained one-byte atlas through GL_ALPHA texture storage instead of expanding it into RGBA for every atlas refresh.
7. Restored the retail renderer export ABI shape around GetRefAPI: source now uses REF_API_VERSION == 9, keeps the legacy post-ModelBounds font slot as a no-op, preserves SetColor immediately after the loading-view bridge, and leaves postprocess_restart at the private tail offset recovered from the 0x9c retail export table.

Task A7: Run the source-file parity audit campaign and isolate the remaining file-level gaps [COMPLETED]

Priority: High Primary areas: docs/reverse-engineering/source-file-parity-ledger-2026-04-22.md, docs/reverse-engineering/source-file-parity-audit-plan-2026-04-22.md, docs/reverse-engineering/source-file-gap-notes/, tools/reverse-engineering/generate_source_file_audit.py, all tracked src/ source trees Parity estimate: before 96% -> after 96%

Completed work:

1. Closed the full source-file parity audit campaign by walking all 567 tracked source entries across the primary runtime, compatibility-only, and secondary support trees in the generated plan order.
2. Kept inherited subsystem closure as the baseline while isolating concrete file-level ownership only where the current evidence supported it, which left the repo-wide source-file gap register concentrated in the seeded RW-G01 and RW-G02 note set instead of reopening already-bounded surfaces.
3. Extended the source-file audit generator and its regression coverage so completed tranche metadata, compatibility wording, function-count fixes, and Phase 4 secondary-source summaries all survive regeneration on the current worktree.
4. Regenerated the source-file ledger, source-file audit plan, gap-note index/readme, and historical audit index so the campaign snapshot is now self-consistent and no pending tranche remains in the generated plan.
5. Returned the top-level queue to the remaining repo-wide workstreams, with the online-service compatibility boundary now back at the head of the active task list.

Task A7p: Audit src/q3radiant function-by-function and isolate file-level gap ownership [COMPLETED]

Priority: High Primary areas: docs/reverse-engineering/source-file-parity-ledger-2026-04-22.md, docs/reverse-engineering/source-file-parity-audit-plan-2026-04-22.md, src/q3radiant/, tests/test_source_file_audit_generator.py, tools/reverse-engineering/generate_source_file_audit.py Parity estimate: before 96% -> after 96%

Completed work:

1. Completed the final Phase 4 secondary-source tranche in the source-file campaign by walking all 97 tracked files under src/q3radiant against the current generated ledger and the repo-wide parity register.
2. Confirmed that this tranche does not open any new file-level gap notes: the retained Radiant editor shell, plugin bridge, OpenGL host glue, and bundled spline/editor helper sources remain bounded secondary support trees on current evidence rather than concrete repo-wide gap owners.
3. Extended the source-file audit generator and its regression coverage so completed metadata for src/q3radiant and the checked final Phase 4 plan item now survive regeneration too.
4. Re-ran the refreshed generator suite (20 passed) and regenerated the source-file campaign docs so the final Phase 4 tranche is now recorded as complete in the generated outputs.
5. Closed the currently enumerated secondary-source file-walk queue in the generated source-file audit plan, leaving no remaining pending Phase 4 tranche in that campaign snapshot.

Task A7o: Audit src/lcc and src/libs function-by-function and isolate file-level gap ownership [COMPLETED]

Priority: High Primary areas: docs/reverse-engineering/source-file-parity-ledger-2026-04-22.md, docs/reverse-engineering/source-file-parity-audit-plan-2026-04-22.md, src/lcc/, src/libs/, tests/test_source_file_audit_generator.py, tools/reverse-engineering/generate_source_file_audit.py Parity estimate: before 96% -> after 96%

Completed work:

1. Completed the third Phase 4 secondary-source tranche in the source-file campaign by walking all 99 tracked files across src/lcc and src/libs against the current generated ledger and the repo-wide parity register.
2. Confirmed that this tranche does not open any new file-level gap notes: the retained LCC compiler/preprocessor/code-generator and bundled test sources under src/lcc, plus the tracked command-line, JPEG, and pak helper sources under src/libs, remain bounded secondary support trees on current evidence rather than concrete repo-wide gap owners.
3. Extended the source-file audit generator and its regression coverage so completed metadata for src/lcc, src/libs, and the checked third Phase 4 plan item now survive regeneration too.
4. Re-ran the refreshed generator suite (19 passed) and regenerated the source-file campaign docs so the third Phase 4 tranche is now recorded as complete in the generated outputs.
5. Advanced the file-walk queue so the remaining Phase 4 target is now the secondary-source tranche covering src/q3radiant/.

Task A7n: Audit src/game, src/q3asm, and src/q3map function-by-function and isolate file-level gap ownership [COMPLETED]

Priority: High Primary areas: docs/reverse-engineering/source-file-parity-ledger-2026-04-22.md, docs/reverse-engineering/source-file-parity-audit-plan-2026-04-22.md, src/game/, src/q3asm/, src/q3map/, tests/test_source_file_audit_generator.py, tools/reverse-engineering/generate_source_file_audit.py Parity estimate: before 96% -> after 96%

Completed work:

1. Completed the second Phase 4 secondary-source tranche in the source-file campaign by walking all 40 tracked files across src/game, src/q3asm, and src/q3map against the current generated ledger and the repo-wide parity register.
2. Confirmed that this tranche does not open any new file-level gap notes: the retained gameplay config helpers and standalone fixture/support sources under src/game, the q3asm bytecode assembler sources, and the q3map compile/light/vis toolchain all remain bounded secondary support trees on current evidence rather than concrete repo-wide gap owners.
3. Extended the source-file audit generator and its regression coverage so completed metadata for src/game, src/q3asm, src/q3map, and the checked second Phase 4 plan item now survive regeneration too.
4. Re-ran the refreshed generator suite plus focused src/game helper coverage (41 passed across the source-audit, factory-config, scoreboard-helper, and tournament-queue suites) so this tranche stays recorded as complete on the current worktree.
5. Regenerated the source-file campaign docs so the second Phase 4 tranche is now recorded as complete in the generated outputs, and the next file-walk target is now the secondary-source tranche covering src/lcc/ and src/libs/.

Task A7m: Audit src/code/bspc, src/code/jpeg-6, and src/code/splines function-by-function and isolate file-level gap ownership [COMPLETED]

Priority: High Primary areas: docs/reverse-engineering/source-file-parity-ledger-2026-04-22.md, docs/reverse-engineering/source-file-parity-audit-plan-2026-04-22.md, src/code/bspc/, src/code/jpeg-6/, src/code/splines/, tests/test_source_file_audit_generator.py, tools/reverse-engineering/generate_source_file_audit.py Parity estimate: before 96% -> after 96%

Completed work:

1. Completed the first Phase 4 secondary-source tranche in the source-file campaign by walking all 106 tracked files across src/code/bspc, src/code/jpeg-6, and src/code/splines against the current repo-wide ledger and the generated source-file register.
2. Corrected the audit generator’s function extractor so the tranche now counts libjpeg macro-style definitions and retained splines C++ method definitions instead of leaving most of jpeg-6 and parts of the splines tree as misleading zero-function placeholders in the ledger.
3. Confirmed that this tranche does not open any new file-level gap notes: the retained BSPC compiler/AAS toolchain, bundled jpeg-6 support sources, and legacy splines helper/editor sources remain bounded secondary trees on current evidence rather than concrete repo-wide gap owners.
4. Extended the source-file audit generator and its regression coverage so completed secondary-section metadata and the checked first Phase 4 plan item now survive regeneration too; the refreshed generator suite passes on the current worktree (17 passed).
5. Regenerated the source-file campaign docs so the first Phase 4 tranche is now recorded as complete in the generated outputs, and the next file-walk target is now the secondary-source tranche covering src/game/, src/q3asm/, and src/q3map/.

Task A7l: Audit src/code/null function-by-function and isolate file-level gap ownership [COMPLETED]

Priority: High Primary areas: docs/reverse-engineering/source-file-parity-ledger-2026-04-22.md, docs/reverse-engineering/source-file-parity-audit-plan-2026-04-22.md, docs/reverse-engineering/source-file-gap-notes/, src/code/null/, docs/reverse-engineering/repo-wide-parity-audit-2026-04-21.md, tests/test_non_windows_portability.py, tools/reverse-engineering/generate_source_file_audit.py Parity estimate: before 96% -> after 96%

Completed work:

1. Completed the next compatibility tranche in the source-file campaign by walking all 7 tracked src/code/null entries against the current repo-wide parity ledger, the existing RW-G02 portability notes, and the focused non-Windows portability regression lane.
2. Isolated src/code/null/null_glimp.c as an additional concrete RW-G02 owner because the null renderer host still consists of no-op GLimp_* entry points plus a QGL_Init() path that returns success without loading a real GL backend, while null_main.c, null_client.c, null_input.c, and null_snddma.c remain the other null-runtime owners already tracked in dedicated notes.
3. Confirmed that no further null file needs a new file-level gap note in this tranche: null_net.c and mac_net.c remain bounded loopback helper surfaces on current evidence rather than newly isolated repo-wide blockers, so they are now explicitly recorded as current compatibility function-walk completions.
4. Extended the source-file audit generator and its regression coverage so a completed null compatibility tranche, the newly isolated null_glimp.c note, and the walked-complete helper rows all survive regeneration; the refreshed generator plus non-Windows portability suites now pass together (20 passed).
5. Regenerated the source-file campaign docs so Phase 3 is now fully closed in the generated outputs, and the next file-walk target is now the first Phase 4 secondary-source tranche covering src/code/bspc/, src/code/jpeg-6/, and src/code/splines/.

Task A7k: Audit src/code/unix function-by-function and isolate file-level gap ownership [COMPLETED]

Priority: High Primary areas: docs/reverse-engineering/source-file-parity-ledger-2026-04-22.md, docs/reverse-engineering/source-file-parity-audit-plan-2026-04-22.md, src/code/unix/, docs/reverse-engineering/repo-wide-parity-audit-2026-04-21.md, tests/test_non_windows_portability.py, tools/reverse-engineering/generate_source_file_audit.py Parity estimate: before 96% -> after 96%

Completed work:

1. Completed the next compatibility tranche in the source-file campaign by walking all 10 tracked src/code/unix entries against the current repo-wide parity ledger, the existing RW-G02 portability notes, and the focused non-Windows portability regression lane.
2. Confirmed that no additional src/code/unix file needs a new file-level gap note: the existing RW-G02 ownership remains concentrated in unix_main.c, linux_glimp.c, linux_snd.c, and linux_joystick.c, while linux_common.c, linux_qgl.c, linux_signals.c, unix_net.c, unix_shared.c, and vm_x86.c are now explicitly recorded as current function-walk completions on present evidence.
3. Extended the source-file audit generator and its regression coverage so a completed compatibility tranche now survives regeneration too, including the checked Phase 3 Unix plan item, the compatibility-specific ledger row wording, and the retained RW-G02 note ownership inside the still-open portability lane.
4. Re-ran the generator and the focused portability coverage (18 passed across the generator and non-Windows suites) so the Unix tranche now stays recorded as complete in the generated docs, and the next file-walk target is now the compatibility-only src/code/null tranche at the head of the remaining queue.

Task A7j: Audit src/code/ui function-by-function and isolate file-level gap ownership [COMPLETED]

Priority: High Primary areas: docs/reverse-engineering/source-file-parity-ledger-2026-04-22.md, docs/reverse-engineering/source-file-parity-audit-plan-2026-04-22.md, src/code/ui/, docs/reverse-engineering/ui-full-parity-audit-and-implementation-plan-2026-04-05.md, tools/reverse-engineering/generate_source_file_audit.py Parity estimate: before 96% -> after 96%

Completed work:

1. Completed the next source-file campaign tranche by walking all 9 tracked src/code/ui entries against the refreshed strict-retail UI audit, the focused UI parity lane, and the current repo-wide gap ledger.
2. Confirmed that no src/code/ui file needs a new file-level gap note: the current UI closure still holds on current evidence, the focused UI suite is green (56 passed, 2 skipped), the clean read-only src/ui runtime-panel parity proof still bounds the menu/data payload, and the read-only tree therefore did not require or permit a source correction in this tranche.
3. Extended the source-file audit generator and its regression coverage so completed read-only tranche metadata now survives regeneration too, including the UI checklist state, the read-only ledger row wording, and the UI tranche summary in the generated audit docs.
4. Regenerated the source-file campaign docs from the hardened generator so the UI tranche is now recorded as complete beside the earlier runtime walks, and the next file-walk target is now the compatibility-only src/code/unix tranche at the head of the remaining queue.

Task A7i: Audit src/code/cgame function-by-function and isolate file-level gap ownership [COMPLETED]

Priority: High Primary areas: docs/reverse-engineering/source-file-parity-ledger-2026-04-22.md, docs/reverse-engineering/source-file-parity-audit-plan-2026-04-22.md, src/code/cgame/, docs/reverse-engineering/game-module-parity-audit-and-implementation-plan-2026-04-10.md, tools/reverse-engineering/generate_source_file_audit.py Parity estimate: before 96% -> after 96%

Completed work:

1. Completed the next module/gameplay tranche in the source-file campaign by walking all 22 tracked src/code/cgame entries against the refreshed strict-retail module audit, the focused cgame parity lane, and the current repo-wide gap ledger.
2. Confirmed that no src/code/cgame file needs a new file-level gap note: the current module closure still holds on current evidence, the focused cgame suite is green (199 passed, 1 skipped), and the shared native export-helper certification still bounds the tree without needing a source correction in this tranche.
3. Hardened the source-file audit generator so completed tranche checkmarks, section result notes, and current-walk ledger rows now survive regeneration instead of reverting already-audited runtime sections back to the default pending wording; --help is also now a safe no-write path.
4. Regenerated the source-file campaign docs from the hardened generator so the corrected cgame function counts and the previously completed runtime tranches live in one durable output again, and the next file-walk target is now the read-only src/code/ui tranche at the head of the remaining module/gameplay queue.

Task A7h: Audit src/code/game function-by-function and isolate file-level gap ownership [COMPLETED]

Priority: High Primary areas: docs/reverse-engineering/source-file-parity-ledger-2026-04-22.md, docs/reverse-engineering/source-file-parity-audit-plan-2026-04-22.md, src/code/game/, docs/reverse-engineering/game-module-parity-audit-and-implementation-plan-2026-04-10.md, tools/reverse-engineering/generate_source_file_audit.py Parity estimate: before 96% -> after 96%

Completed work:

1. Hardened the source-file audit generator again so it now ignores control-flow keywords such as if, for, while, and switch, which keeps the source-file campaign counts aligned with real Quake-style C definitions in the src/code/game tranche instead of overcounting control blocks.
2. Completed the next module/gameplay tranche in the source-file campaign by walking all 45 tracked src/code/game entries against the refreshed strict-retail module audit, the qagame retail gate, and the current repo-wide gap ledger.
3. Closed the concrete issues surfaced during the walk without opening a new persistent gap note: BG_CanGrabWeaponItem() now matches the retained retail world-weapon regrab gate, the Attack & Defend lifecycle/frame hooks are explicit in the module-side source layout again, and the standalone auto-shuffle countdown harness now uses the shared compiler-discovery path on Windows instead of assuming gcc.
4. Confirmed that no src/code/game file needs a new file-level gap note: the current gameplay/module closure still holds on current evidence, and the newer auto-shuffle/countdown, Clan Arena shuffle, Race, ready-up, and shared pmove coverage all remain bounded inside the already-closed module register rather than opening a new repo-wide gap family.
5. Recorded the result directly in the source-file campaign docs so the game rows now distinguish current function-walk completion from the still-pending module trees, and the next file-walk target is now the src/code/cgame tranche at the head of the remaining module/gameplay queue.

Task A7g: Audit src/code/botlib function-by-function and isolate file-level gap ownership [COMPLETED]

Priority: High Primary areas: docs/reverse-engineering/source-file-parity-ledger-2026-04-22.md, docs/reverse-engineering/source-file-parity-audit-plan-2026-04-22.md, src/code/botlib/, docs/reverse-engineering/botlib-internal-parity-audit-and-implementation-plan-2026-04-10.md, tools/reverse-engineering/generate_source_file_audit.py Parity estimate: before 96% -> after 96%

Completed work:

1. Fixed the source-file audit generator so it now counts Quake-style brace-on-next-line C function definitions, which restores accurate function-count coverage for the src/code/botlib tranche instead of leaving most files reported as zero-function placeholders.
2. Completed the next primary-runtime tranche in the source-file campaign by walking all 28 tracked src/code/botlib entries against the dedicated botlib internal audit, the retained mapping-round 61 bridge/import ownership, and the current repo-wide gap ledger.
3. Confirmed that no src/code/botlib file needs a new file-level gap note: the retained bridge/import closures and the deterministic AAS/reachability/goal-state proof lane still bound the tree on current evidence, while the remaining live-map or gameplay-tuning nuance stays explicitly outside the repo-wide/file-level gap register.
4. Recorded the result directly in the source-file campaign docs so the botlib rows now distinguish current function-walk completion from the still-pending module and compatibility trees, and the next file-walk target is now the src/code/game tranche at the head of the remaining module/gameplay queue.

Task A7f: Audit src/code/win32 function-by-function and isolate file-level gap ownership [COMPLETED]

Priority: High Primary areas: docs/reverse-engineering/source-file-parity-ledger-2026-04-22.md, docs/reverse-engineering/source-file-parity-audit-plan-2026-04-22.md, src/code/win32/, docs/reverse-engineering/platform-specific-engine-parity-audit-and-implementation-plan-2026-04-16.md Parity estimate: before 96% -> after 96%

Completed work:

1. Completed the next primary-runtime tranche in the source-file campaign by walking all 11 tracked src/code/win32 entries against the closed strict-retail Windows platform audit, renderer host-glue evidence, and the current repo-wide gap ledger.
2. Confirmed that no src/code/win32 file needs a new file-level gap note: the retained clipboard, raw-input, loading-window, renderer-host glue, and win_glimp.c pixel-format closures still hold on current evidence, and no file in this tree currently owns a repo-wide gap family.
3. Recorded the result directly in the source-file campaign docs so the Win32 rows now distinguish current function-walk completion from the still-pending runtime trees, without reopening the closed strict-retail Windows platform register.
4. Narrowed A7 procedurally again: the next file-walk target is now the src/code/botlib tranche at the head of the remaining strict-retail engine-core queue.

Task A7e: Audit src/code/renderer function-by-function and isolate file-level gap ownership [COMPLETED]

Priority: High Primary areas: docs/reverse-engineering/source-file-parity-ledger-2026-04-22.md, docs/reverse-engineering/source-file-parity-audit-plan-2026-04-22.md, src/code/renderer/, docs/reverse-engineering/renderer-full-parity-audit-and-implementation-plan-2026-04-09.md Parity estimate: before 96% -> after 96%

Completed work:

1. Completed the next primary-runtime tranche in the source-file campaign by walking all 23 tracked src/code/renderer entries against the closed strict-retail renderer audit, mapping notes, and the current repo-wide gap ledger.
2. Confirmed that no src/code/renderer file needs a new file-level gap note: the retained export, image, post-process, scene/runtime, font, and host-text closures still hold on current evidence, and the then-bounded R_fonsErrorCallback module-runtime blocker remained classified under RW-G04 evidence freshness rather than as a new renderer source-gap owner. The later 2026-05-20 renderer wiring pass patched the source-side eager FontStash prebuild behind that stale artifact.
3. Recorded the result directly in the source-file campaign docs so the renderer rows now distinguish current function-walk completion from the still-pending runtime trees, without reopening the closed strict-retail renderer register.
4. Narrowed A7 procedurally again: the next file-walk target is now the src/code/win32 tranche at the head of the remaining strict-retail engine-core queue.

Task A7d: Audit src/code/server function-by-function and isolate file-level gap ownership [COMPLETED]

Priority: High Primary areas: docs/reverse-engineering/source-file-parity-ledger-2026-04-22.md, docs/reverse-engineering/source-file-parity-audit-plan-2026-04-22.md, src/code/server/, docs/reverse-engineering/server-full-parity-audit-and-implementation-plan-2026-04-10.md Parity estimate: before 96% -> after 96%

Completed work:

1. Completed the next primary-runtime tranche in the source-file campaign by walking all 11 tracked src/code/server entries against the closed strict-retail server audit, mapping notes, and the current repo-wide gap ledger.
2. Confirmed that no new src/code/server file-level owners need opening beyond the already-seeded RW-G01 note for sv_rankings.c; the retained Steam GameServer/auth/stats and idZMQ publication/runtime owners remain closed on current evidence.
3. Recorded the result directly in the source-file campaign docs so the server rows now distinguish current function-walk completion from the still-pending runtime trees, while keeping the bounded rankings ownership explicit in the existing sv_rankings.c gap note instead of reopening the closed strict-retail server register.
4. Narrowed A7 procedurally again: the next file-walk target is now the src/code/renderer tranche at the head of the remaining strict-retail engine-core queue.

Task A7c: Audit src/code/client function-by-function and isolate file-level gap ownership [COMPLETED]

Priority: High Primary areas: docs/reverse-engineering/source-file-parity-ledger-2026-04-22.md, docs/reverse-engineering/source-file-parity-audit-plan-2026-04-22.md, src/code/client/, docs/reverse-engineering/client-full-parity-audit-and-implementation-plan-2026-04-09.md Parity estimate: before 96% -> after 96%

Completed work:

1. Completed the next primary-runtime tranche in the source-file campaign by walking all 22 tracked src/code/client entries against the closed strict-retail client audit, mapping notes, and the current repo-wide gap ledger.
2. Confirmed that no new src/code/client file-level owners need opening beyond the already-seeded RW-G01 notes for ql_auth.c and cl_steam_resources.c.
3. Recorded the result directly in the source-file campaign docs so the client rows now distinguish current function-walk completion from the still-pending runtime trees, while keeping the bounded online-services ownership explicit in the two existing client gap notes instead of reopening the closed strict-retail client register.
4. Narrowed A7 procedurally again: the next file-walk target is now the src/code/server tranche at the head of the remaining strict-retail engine-core queue.

Task A7b: Audit src/code/qcommon function-by-function and isolate file-level gap ownership [COMPLETED]

Priority: High Primary areas: docs/reverse-engineering/source-file-parity-ledger-2026-04-22.md, docs/reverse-engineering/source-file-parity-audit-plan-2026-04-22.md, src/code/qcommon/, docs/reverse-engineering/qcommon-full-parity-audit-and-implementation-plan-2026-04-10.md Parity estimate: before 96% -> after 96%

Completed work:

1. Completed the next primary-runtime tranche in the source-file campaign by walking all 19 tracked src/code/qcommon entries against the closed strict-retail qcommon audit, mapping rounds, and the current repo-wide ledger.
2. Confirmed that no new src/code/qcommon file-level owners need opening: the current strict-retail qcommon register remains closed, and no additional repo-wide gap family is concretely owned by a qcommon source file on the current worktree.
3. Recorded the result directly in the source-file campaign docs so the qcommon rows now distinguish current function-walk completion from the still-pending runtime trees, while keeping vm_x86.c classified as the already-bounded compatibility carry beneath the closed vm.c host-selection seam.
4. Narrowed A7 procedurally again: the next file-walk target is now the src/code/client tranche at the head of the remaining strict-retail engine-core queue.

Task A7a: Audit src/common function-by-function and isolate file-level gap ownership [COMPLETED]

Priority: High Primary areas: docs/reverse-engineering/source-file-parity-ledger-2026-04-22.md, docs/reverse-engineering/source-file-parity-audit-plan-2026-04-22.md, src/common/, src/common/platform/ Parity estimate: before 96% -> after 96%

Completed work:

1. Completed the first primary-runtime tranche in the new source-file campaign by walking all 18 tracked src/common entries against the current engine-wide closure and repo-wide gap ledgers.
2. Confirmed that no new src/common file-level owners need opening beyond the already-seeded RW-G01 notes for platform_config.h, platform_services.c, and the bounded auth backend shims.
3. Recorded the result directly in the source-file campaign docs so the src/common rows now distinguish current function-walk completion from the still-pending primary-runtime trees, while keeping platform_steamworks.c classified as a retained Steamworks wrapper surface gated by policy rather than as a newly opened repo-wide gap owner.
4. Narrowed A7 procedurally: the next file-walk target is now the src/code/qcommon tranche at the head of the remaining strict-retail engine-core queue.

Task A4i: Replace the Unix Sys_CheckCD() unconditional pass with a bounded data-root probe [COMPLETED]

Priority: Medium Primary areas: src/code/unix/unix_main.c, tests/test_non_windows_portability.py, docs/platform/toolchain-matrix.md, docs/build/linux-glibc-32bit.md Parity estimate: before 96% -> after 96%

Completed work:

1. Replaced the Unix Sys_CheckCD() return qtrue; placeholder with a bounded Quake-data probe that now scans the configured fs_basepath, fs_cdpath, default install roots, and current working directory for baseq3/default.cfg, pak00.pk3, or pak0.pk3 before reporting success.
2. Kept the reconstructed lane deliberately compatibility-scoped rather than over-claiming retail parity: the probe is a coarse host-side data-root guard for existing engine callers such as SV_BotInitBotLib(), not a full filesystem bootstrap replacement.
3. Expanded the focused non-Windows portability suite so the new root list, accepted asset markers, and baseq3 path contract are source-pinned alongside the previously restored low-memory, symbol-compare, release-marker, profiling, and clipboard seams.
4. Narrowed RW-G02 again: the remaining portability debt is now even more cleanly concentrated in the absent real Unix renderer/audio/input host modernization and broader non-Windows validation breadth, not in an unconditional Unix content-root success stub.

Task A4h: Restore a bounded Unix clipboard compatibility path [COMPLETED]

Priority: Medium Primary areas: src/code/unix/unix_main.c, tests/test_non_windows_portability.py, docs/platform/toolchain-matrix.md, docs/build/linux-glibc-32bit.md Parity estimate: before 96% -> after 96%

Completed work:

1. Replaced the bare Unix Sys_GetClipboardData() NULL stub with a bounded compatibility path that now reads clipboard text through wl-paste, xclip, or xsel when the surrounding Wayland/X11 environment and helper binaries are available, while still returning NULL cleanly on unsupported hosts.
2. Bounded the new lane explicitly so it mirrors the existing host contract instead of over-claiming parity: clipboard text is trimmed at the first newline/control break, capped to a fixed maximum size, and copied into the engine allocator before it reaches the existing client, UI, and browser clipboard consumers.
3. Expanded the focused non-Windows portability suite so the Unix clipboard probe chain, PATH lookup, command gating, and Wayland/X11 fallback order are source-pinned alongside the earlier low-memory, symbol-compare, monkey-marker, profiling, and null-runtime compatibility seams.
4. Narrowed RW-G02 again: the remaining portability debt is now even more clearly in the missing real Unix renderer/audio/input host modernization and broader non-Windows validation breadth, not in an unimplemented Unix clipboard host hook.

Task A2c: Refresh the repo-wide parity audit evidence on the current worktree [COMPLETED]

Priority: Medium Primary areas: AUDIT.md, IMPLEMENTATION_PLAN.md, docs/reverse-engineering/repo-wide-parity-audit-2026-04-21.md Parity estimate: before 96% -> after 96%

Completed work:

1. Re-ran a broad current-worktree parity sweep across the top-level strict-retail gates, gameplay validation fixtures, portability suite, and staged runtime-audit lane, producing 72 passed, 7 skipped.
2. Rebased the top-level ledgers on that sweep so the current repo-wide gap register is explicitly backed by one aggregated evidence pass instead of only by the individual runtime-probe refresh notes.
3. Clarified that the remaining repo-wide gap families still collapse to RW-G01, RW-G02, and RW-G04, with the staged retail-runtime DLL guard now treated as closed and the remaining RW-G04 debt narrowed to fresh artifact regeneration/publication.
4. Cleaned up stale date language in the historical top-level audit anchors so the repo-wide ledgers no longer point readers at the older 2026-04-17 top-level state while describing the current 2026-04-21 worktree.

Task A6f: Add a strict staged retail-runtime DLL audit for native Windows validation [COMPLETED]

Priority: Medium Primary areas: tools/ci/audit-retail-dependencies.ps1, tools/ci/validate-windows-native.ps1, Windows/toolchain docs Parity estimate: before 96% -> after 96%

Completed work:

1. Extended the retail dependency auditor so it can now audit either the local Steam install or an explicit runtime root, checking for missing expected retail DLLs, extra non-retail DLLs, and byte-hash mismatches instead of only diffing the Steam tree opportunistically.
2. Added a strict retail-runtime staging step to tools/ci/validate-windows-native.ps1: the retail validation lane now assembles build\win32\<Config>\retail-runtime\ from the rebuilt host executables and gameplay/UI DLLs plus the exact retail launcher DLL payload from assets\quakelive\, then fails fast if the staged root contains extra DLLs or is missing any retail payload DLLs.
3. Rebased the Windows/toolchain documentation on that explicit staging boundary so the mixed developer build root build\win32\<Config>\bin\ is no longer described as the strict retail payload surface.
4. Narrowed RW-G04 again: the remaining freshness debt is now no longer the previously documented local/runtime DLL-guard follow-up, and is instead concentrated in refreshed native build outputs plus the broader glibc and self-hosted publication follow-ups.

Task A6e: Re-promote the retail module runtime alias through an explicit renderer-text bounded blocker [COMPLETED]

Priority: Medium Primary areas: tools/modules/run_retail_module_runtime_probe.ps1, tests/test_game_module_retail_parity_gate.py, runtime-evidence ledgers Parity estimate: before 96% -> after 96%

Completed work:

1. Expanded the retail-module probe's bounded-owner classification so the current 2026-04-21 rerun can distinguish two renderer-owned live-map blocker shapes outside module scope: the older fatal R_LoadMD3 drop path and the current R_fonsErrorCallback font-atlas saturation lane that keeps retail uix86.dll, cgamex86.dll, and qagamex86.dll loaded but prevents CS_ACTIVE.
2. Rebased the module runtime probe on the current map-launch contract again by threading the same com_maxfps 30 / no-warmup assumptions used by the other runtime probes into the retail-module lane, while keeping the stable alias guarded so only sufficient reruns are promoted.
3. Reran tools/modules/run_retail_module_runtime_probe.ps1 on 2026-04-21 and refreshed artifacts/module_validation/logs/retail_module_runtime_evidence_latest.json to the current bounded 20260421 bundle instead of leaving it pinned to the older 2026-04-09 artifact.
4. Narrowed RW-G04 again: the remaining freshness debt is now no longer a non-promotable retail-module alias question, and is instead concentrated in fresh native build outputs plus the broader glibc/self-hosted publication follow-ups.

Task A4g: Restore a bounded Unix profiling control path [COMPLETED]

Priority: Medium Primary areas: src/code/unix/unix_main.c, src/code/unix/Makefile, tests/test_non_windows_portability.py, portability ledgers Parity estimate: before 96% -> after 96%

Completed work:

1. Replaced the empty Unix profiling hooks with a bounded gprof-compatible runtime path: Sys_Init() now pauses profiling early when moncontrol is available, Sys_BeginProfiling() re-enables it on first active play, and Sys_Exit() now flushes _mcleanup() through Sys_EndProfiling().
2. Added an explicit Unix build toggle QL_ENABLE_GPROF=1 so profiling is a reproducible opt-in lane rather than an undocumented linker accident. The Unix makefile now threads -pg through the relevant engine compile/link paths when that toggle is enabled.
3. Rebased the focused non-Windows portability checks and ledgers so they now record a bounded profiling lane instead of an empty Unix no-op while still keeping the broader Unix renderer/audio/input host gap open.
4. Narrowed RW-G02 again: the remaining portability debt is now even more clearly in the modern Unix runtime boundary itself, not in stale host-side syscall placeholders.

Task A6d: Refresh renderer runtime evidence and stabilize alias promotion for the renderer/module probes [COMPLETED]

Priority: Medium Primary areas: tools/renderer/run_renderer_runtime_probe.ps1, tools/modules/run_retail_module_runtime_probe.ps1, tests/test_renderer_full_parity_gate.py, tests/test_renderer_text_runtime_validation_parity.py, tests/test_game_module_retail_parity_gate.py, runtime-evidence docs Parity estimate: before 96% -> after 96%

Completed work:

1. Repaired the renderer runtime probe contract so the current map <name> ffa path, quoted launch arguments, and normalized path handling work reliably again, then reran the probe on 2026-04-21 to refresh the clean dated bundle artifacts/renderer_validation/logs/renderer_runtime_evidence_20260421.json.
2. Added guarded *_latest.json promotion to both the renderer and retail module probes so only sufficient reruns replace the stable tracked alias. The renderer alias now points at the clean 2026-04-21 bundle, and the retail-module lane gained the guarded promotion rules that later enabled A6e to refresh the stable alias to the current bounded 2026-04-21 module artifact instead of silently accepting degraded reruns.
3. Rebased the renderer/module parity gates, the font audit script, and the active runtime-evidence docs on those stable aliases plus the current map <name> ffa launch contract.
4. Narrowed RW-G04 again: the remaining freshness debt is now concentrated in fresh native build outputs and the broader glibc/self-hosted publication follow-ups.

Task A4f: Classify the Linux glibc preset as server-only portability evidence [COMPLETED]

Priority: Medium Primary areas: docs/build/linux-glibc-32bit.md, docs/platform/toolchain-matrix.md, tests/test_non_windows_portability.py, portability ledgers Parity estimate: before 96% -> after 96%

Completed work:

1. Updated docs/build/linux-glibc-32bit.md so the helper is explicitly framed as a qagamei386.so server-module reconstruction preset rather than as evidence of a retail-equivalent Linux client/runtime.
2. Mirrored that portability boundary in the native toolchain matrix so the top-level Linux lane description now points contributors at the correct server-only interpretation immediately.
3. Expanded the focused non-Windows portability regression coverage so the server-module-only classification and the still-open Unix profiling or renderer/audio/input host gaps are pinned in the docs.
4. Narrowed RW-G02 procedurally: the remaining work is now more cleanly about whether to modernise the real Unix host boundary, not about what the existing Linux build helper is supposed to prove.

Task A4e: Modernise the null input compatibility contract [COMPLETED]

Priority: Medium Primary areas: src/code/null/null_input.c, tests/test_non_windows_portability.py, tests/test_engine_host_support_full_parity_gate.py, portability ledgers Parity estimate: before 95% -> after 96%

Completed work:

1. Modernised src/code/null/null_input.c so the dedicated/null host now carries the current input bootstrap cvars (in_mouse, in_nograb, in_joystick, in_debugjoystick, and joy_threshold) plus an explicit compatibility-state refresh that keeps ui_joyavail pinned to 0 instead of leaving the file as a bare inert no-op.
2. Reconstructed the null input frame pump as an explicit no-device contract: IN_Frame() now maintains the compatibility cvar surface and consumes joystick modification latches, while Sys_SendKeyEvents() now refreshes the same no-device state without emitting real input events.
3. Expanded both the focused non-Windows portability suite and the engine-host/support compatibility gate so the current null input lane is pinned alongside the earlier null host/client/audio modernization work.
4. Narrowed RW-G02 again: the remaining portability debt is now concentrated even more tightly in the Unix profiling placeholders and the absence of a real modern Unix client/renderer/audio/input host, rather than in stale null input bootstrap/event-pump seams.

Task A4d: Modernise the null client/audio compatibility contracts [COMPLETED]

Priority: Medium Primary areas: src/code/null/null_client.c, src/code/null/null_snddma.c, tests/test_non_windows_portability.py, portability ledgers Parity estimate: before 94% -> after 95%

Completed work:

1. Modernised src/code/null/null_client.c so the null compatibility lane now exposes the current browser/advert/input-facing client.h contract instead of only the older partial web-host shim set: the null client now carries CL_RefreshOnlineServicesBridgeState(), the current browser cursor and app activation hooks, mouse/button/wheel webview input stubs, and the retained advertisement-bridge entry points while explicitly forcing the browser cvars back to the non-live state.
2. Promoted src/code/null/null_snddma.c from a narrow legacy dummy into an explicit current silent-audio compatibility seam by adding the current SNDDMA_Activate() and S_AddVoiceSamples() entry points alongside the existing null DMA and local-sound hooks.
3. Expanded the focused non-Windows portability regression coverage so it now asserts the current null client/audio contract surfaces directly instead of only pinning the older null host/network/glimp tranche.
4. Narrowed RW-G02 again: the remaining non-Windows debt is now concentrated in the profiling hooks plus the absence of a real modern Unix client/renderer/audio/input host, rather than in stale or missing null browser/advert/audio contract entry points.

Task A4c: Modernise the null host contracts and reconstruct the Unix monkey-marker hook [COMPLETED]

Priority: Medium Primary areas: src/code/unix/unix_main.c, src/code/null/null_main.c, src/code/null/null_net.c, src/code/null/mac_net.c, src/code/null/null_glimp.c, tests/test_non_windows_portability.py, portability ledgers Parity estimate: before 93% -> after 94%

Completed work:

1. Reconstructed the Unix Sys_MonkeyShouldBeSpanked() compatibility hook as a retained q3monkeyid release-marker probe rooted in the historical Construct script plus the committed retail string corpus, instead of leaving it as an unconditional placeholder.
2. Modernised the stale null host shims so they now match the current qcommon contracts: null_main.c now carries current command-line, executable-name, timer, path, and homepath scaffolding; null_net.c and mac_net.c now expose Sys_StringToAdr() plus the current loopback packet stub signatures; and null_glimp.c no longer carries the stale int GLimp_Init() signature.
3. Added focused non-Windows portability coverage for the new Unix marker probe and the null-host contract cleanup.
4. Narrowed RW-G02 again: the remaining open portability debt is now concentrated in the profiling hooks plus the broader Unix/client/audio/input and still-stubbed null runtime surfaces, rather than in stale host-contract mismatches or the old q3monkeyid placeholder.

Task A6c: Refresh the qcommon, server, and WOW64 evidence contracts on the current machine [COMPLETED]

Priority: Medium Primary areas: tools/qcommon/run_qcommon_runtime_probe.ps1, tools/server/run_server_runtime_probe.ps1, tools/ci/wow64-smoketest.ps1, their tracked artifacts, and repo-wide validation ledgers Parity estimate: before 93% -> after 93%

Completed work:

1. Repaired the stale qcommon and dedicated-server probe contracts so both runtime lanes now use the current map <name> ffa command shape, then re-ran them on 2026-04-21 to refresh artifacts/qcommon_validation/logs/qcommon_runtime_evidence_20260410.json and artifacts/server_validation/logs/server_runtime_evidence_20260410.json.
2. Rebased the qcommon and server gate/documentation expectations on the current runtime markers, including the qcommon game.error plus native stopRefresh fallback evidence and the current dedicated getstatus server-type field spelling.
3. Updated the retained WOW64 smoke harness so it now prefers retail DLLs from assets/quakelive/baseq3/ when present but falls back to the current Win32 build outputs under build/win32/Debug/bin/baseq3/, then captured a successful 32-bit PowerShell run at artifacts/wow64-smoketest/wow64-smoketest.log.
4. Narrowed RW-G04 again: the remaining freshness debt is now concentrated in fresh native build outputs, the still-archived runtime bundles outside the refreshed client/qcommon/server lanes, and the broader glibc or self-hosted validation publication follow-ups.

Task A4b: Restore bounded Unix function compare/checksum coverage on the compatibility host lane [COMPLETED]

Priority: Medium Primary areas: src/code/unix/unix_main.c, tests/test_non_windows_portability.py, docs/platform/toolchain-matrix.md, top-level portability ledgers Parity estimate: before 93% -> after 93%

Completed work:

1. Restored Unix Sys_FunctionCmp() and Sys_FunctionCheckSum() with a Linux/glibc symbol-backed path that resolves function byte ranges through dladdr1(..., RTLD_DL_SYMENT) and hashes them with Com_BlockChecksum() when symbol sizes are available.
2. Added focused portability coverage that now checks the new source-side helper path directly instead of only tracking the old placeholder story in documentation.
3. Narrowed RW-G02 again: the remaining open Unix helper debt is now concentrated in the historical profiling lane plus the broader non-Windows client/audio/input/toolchain boundary.

Task A6b: Re-baseline the src/ui runtime-panel contract to the clean current state [COMPLETED]

Priority: Medium Primary areas: tests/test_ui_src_panel_parity.py, tests/test_ui_full_parity_gate.py, docs/quakelive_asset_audit.md, docs/ui/hud-audit.md, docs/ui/scripting-guide.md, top-level ledgers Parity estimate: before 93% -> after 93%

Completed work:

1. Tightened the repo/UI parity gates so the current worktree now expects the src/ui runtime-panel compare to be clean (65 / 65, 0 content diffs) instead of tolerating the older historical non-zero drift contract.
2. Re-based the UI documentation and the repo-wide ledgers so they now distinguish clean checked-in runtime-panel parity from the separate bundle-staged retail art path.
3. Documented that explicit runtime-root refreshes always write the runtime package manifest, but only emit pak_ui_src_retail_overlay.pk3 when drift_files is non-empty.

Task A6a: Refresh the tracked client runtime evidence bundle after UI package-proof hardening [COMPLETED]

Priority: Medium Primary areas: tools/client/run_client_runtime_probe.ps1, tests/test_client_full_parity_gate.py, artifacts/client_validation/logs/client_runtime_evidence_20260410.json, client/runtime audit docs Parity estimate: before 93% -> after 93%

Completed work:

1. Tightened the client parity gate so the tracked runtime bundle must now prove the emitted UI runtime-package manifest plus the refreshed main package hashes instead of relying only on the older screenshot/log bundle.
2. Rebased the main-menu offline-browser proof on the current retained runtime behavior, using the observed game.error publication and UI-side stopRefresh fallback markers instead of the older stale probe expectations.
3. Re-ran tools/client/run_client_runtime_probe.ps1 on 2026-04-21 and refreshed artifacts/client_validation/logs/client_runtime_evidence_20260410.json with fresh screenshots, logs, demo output, and UI package evidence.
4. Narrowed RW-G04 by refreshing the client runtime bundle while leaving the remaining archived runtime/build artifacts as the still-open repo-wide freshness debt.

Task A5: Harden the read-only src/ui parity path with runtime package proof [COMPLETED]

Priority: High Primary areas: tools/build_ui_bundle.py, tests/test_ui_src_panel_parity.py, tools/client/run_client_runtime_probe.ps1, docs/quakelive_asset_audit.md, docs/ui/hud-audit.md, docs/ui/scripting-guide.md Parity estimate: before 92% -> after 93%

Completed work:

1. Added an explicit runtime-package manifest to tools/build_ui_bundle.py for --runtime-root runs so emitted pak_uiql.pk3 and pak_ui_src_retail_overlay.pk3 outputs are now machine-described instead of assumed from console output alone.
2. Added focused UI bundle coverage that proves the explicit runtime-root path emits the expected PK3s, includes the required runtime entries and compatibility aliases, and keeps any emitted overlay package in lockstep with the source-vs-retail drift contract.
3. Tightened the client runtime probe so it now verifies the refreshed runtime package manifest and fails fast if the required UI runtime package was not emitted into the writable homepath.
4. Refreshed the UI packaging docs so they consistently describe the current contract: staging/overlay manifests by default, explicit package emission only for runtime-root refreshes or probe flows, with the overlay PK3 remaining optional when the drift contract is empty.

Task A4a: Restore Unix low-memory detection on the compatibility host lane [COMPLETED]

Priority: Medium Primary areas: src/code/unix/unix_main.c, docs/platform/toolchain-matrix.md Parity estimate: before 92% -> after 92%

Completed work:

1. Replaced the unconditional Unix Sys_LowPhysicalMemory() placeholder with a native sysconf()-backed physical page-count query so the renderer/cgame low-memory heuristic no longer hardcodes the non-low-memory path on Unix.
2. Refreshed the platform note so it now records the restored low-memory probe separately from the still-bounded checksum/comparison/profiling helpers.
3. Narrowed the active RW-G02 description to the remaining portability debt instead of continuing to list Sys_LowPhysicalMemory() as an unresolved placeholder.

Task A2: Repo-wide parity re-baseline and documentation convergence [COMPLETED]

Priority: Critical Primary areas: AUDIT.md, IMPLEMENTATION_PLAN.md, docs/reverse-engineering/repo-wide-parity-audit-2026-04-21.md, docs/parity-plan.md, docs/ui_deltas.md, docs/ui_followup_issues.md, docs/quakelive_asset_audit.md Parity estimate: before 92% -> after 92%

Completed work:

1. Re-ran the repo's top-level parity-gate suites on 2026-04-21 and confirmed that the strict-retail Windows gate set remains green on the current worktree.
2. Re-based the top-level ledger so it now distinguishes the strict-retail Windows closure state from repo-wide parity across compatibility-only, portability, and packaging-dependent lanes.
3. Published a dedicated repo-wide gap register in docs/reverse-engineering/repo-wide-parity-audit-2026-04-21.md.
4. Demoted the most misleading broad-planning documents to explicit historical references so they no longer conflict with the current reverse-engineering ledgers.

Task A1: Client parity-gate revalidation and workflow restoration [COMPLETED]

Priority: Critical Primary areas: .github/workflows/client-validation.yml, tools/client/run_client_runtime_probe.ps1, artifacts/client_validation/logs/*, client ledger docs Parity estimate: before 99% -> after 100%

Completed work:

1. Restored .github/workflows/client-validation.yml so the client validation lane again runs the focused platform, Steamworks, config, workshop, UI-menu, and unified parity-gate suites expected by the audited CL-P6 closure.
2. Repaired tools/client/run_client_runtime_probe.ps1 so it now:
   • runs against the default build-disabled client configuration
   • uses the valid map bloodrun ffa command shape for the local-map probe
   • avoids helper-stdout pollution in the structured probe return path
   • retries transient pak_uiql.pk3 rewrite locks on Windows
   • caps the probe at com_maxfps 30, which lets the map pass reach CS_ACTIVE before the queued demo/screenshot/disconnect commands fire
3. Refreshed artifacts/client_validation/logs/client_runtime_evidence_20260410.json and artifacts/client_validation/logs/client_full_parity_gate.json with a clean runtime bundle.
4. Revalidated the lane with:
   • pytest tests/test_client_full_parity_gate.py -q --tb=no
   • pytest tests/test_platform_services.py tests/test_steamworks_harness.py tests/test_client_config_parity.py tests/test_client_workshop_bootstrap_parity.py tests/test_ui_menu_files.py tests/test_client_full_parity_gate.py -q --tb=no

Task A3a: Expose explicit compatibility-only provider/policy labels through auth, workshop, and browser-overlay surfaces [COMPLETED]

Priority: Critical Primary areas: src/common/platform/, src/code/client/, docs/platform/authentication.md, docs/steam_platform_abstraction.md Parity estimate: before 96% -> after 96%

Completed work:

1. Added QL_DescribePlatformFeaturePolicy(...) as the shared companion label for platform-service descriptors and threaded that label through the client auth dispatcher so build-disabled, runtime-disabled, and provider-unavailable lanes stay explicit in auth lifecycle logs.
2. Mirrored the browser/advert overlay descriptor through the ROM cvars ui_browserAwesomiumProvider and ui_browserAwesomiumPolicy, and routed the blocked browser commands through provider-aware compatibility logging instead of generic “provider unavailable” messages.
3. Bounded the retained workshop bootstrap to the Steam UGC owner lane, documented provider/policy-aware fallback logging for non-Steam/bootstrap paths, and refreshed the focused platform/workshop regression coverage to keep those compatibility-only labels visible.

Task A3b: Expose explicit compatibility-only provider/policy labels through server auth and Steam GameServer surfaces [COMPLETED]

Priority: Critical Primary areas: src/code/server/, src/code/qcommon/, docs/platform/authentication.md, docs/steam_platform_abstraction.md Parity estimate: before 96% -> after 96%

Completed work:

1. Mirrored the retained server auth and matchmaking descriptors through the ROM cvars sv_platformAuthProvider, sv_platformAuthPolicy, sv_steamServerProvider, and sv_steamServerPolicy so dedicated-server diagnostics expose the active provider/policy pair just as explicitly as the client-side browser, workshop, and auth surfaces.
2. Bounded the Steam GameServer bootstrap/publication lane with provider/policy-aware fallback and callback logging, including the compatibility-only dedicated-server publication fallback when the retained GameServer owner is unavailable.
3. Updated server-side auth telemetry so it now appends provider=<...> policy=<...> to the message payload while intentionally preserving the legacy credential=steam field, then refreshed the focused platform-service regression coverage to keep that compatibility labeling explicit.

Task A3c: Expose explicit compatibility-only provider/policy labels through client live-resource bridge surfaces [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_steam_resources.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-client-steam-resources.md Parity estimate: before 96% -> after 96%

Completed work:

1. Threaded the overlay service descriptor’s provider/policy labels into the retained live-resource bridge so the steam:// avatar/resource path has the same explicit compatibility labeling as the browser overlay and workshop seams.
2. Replaced the generic Steam-resource and launcher-backend diagnostics with provider-aware logs that now spell out when the Steam resource bridge is disabled, when the retained SteamDataSource owner cannot satisfy a request, and when the launcher/web fallback owner is the remaining compatibility path.
3. Refreshed the focused platform-service regression coverage and the file-level RW-G01 gap note so the live-resource bridge now documents the bounded compatibility/fallback story it still implements instead of the older generic “backend unavailable” wording.

Task A3d: Expose explicit compatibility-only provider/policy labels through client matchmaking, stats, and social-overlay command surfaces [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md Parity estimate: before 96% -> after 96%

Completed work:

1. Mirrored the retained client matchmaking, stats, and social-overlay descriptors through the ROM cvars cl_matchmakingProvider, cl_matchmakingPolicy, cl_statsProvider, cl_statsPolicy, cl_socialOverlayProvider, and cl_socialOverlayPolicy so the active provider/policy pair remains visible at bootstrap and callback-init time.
2. Routed the retained stats_clear, connect_lobby, clientviewprofile, clientfriendinvite, callback-bundle fallback, and main-menu rich-presence failure paths through provider-aware compatibility diagnostics instead of generic Steam-only wording.
3. Refreshed the focused platform-service regression coverage and the platform abstraction notes so this client command surface now documents the bounded compatibility lane it still implements.

Task A3e: Expose explicit compatibility-only provider/policy labels through the default-disabled server rankings surface [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_rankings.c, src/code/server/sv_init.c, tests/test_platform_services.py, docs/reverse-engineering/source-file-gap-notes/rw-g01-server-rankings.md Parity estimate: before 96% -> after 96%

Completed work:

1. Added explicit rankings provider/policy helpers and mirrored them through the ROM cvars sv_rankingsProvider and sv_rankingsPolicy so the default-disabled rankings lane no longer hides behind the generic sv_rankingsActive flag alone.
2. Threaded the provider/policy pair into the disabled rankings bootstrap diagnostics so the forced sv_enableRankings 0 fallback and the retained compatibility-only rankings surface stay explicit at runtime.
3. Refreshed the focused server/platform-service parity checks and the RW-G01 rankings gap note so this retained GPL-era rankings lane is now documented using the same compatibility-labeling story as the other active repo-wide online-service seams.

Task A3f: Expose explicit heuristic compatibility labeling through auth backend responses and fallback traces [COMPLETED]

Priority: Critical Primary areas: src/common/platform/backends/, src/code/client/ql_auth.c, tools/integration/auth_flow_trace.py, docs/platform/authentication.md, docs/steam_platform_abstraction.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the retained Steamworks and open-adapter auth backends so their response payloads explicitly identify themselves as heuristic compatibility backends instead of reading like retail live-service verdicts.
2. Threaded an explicit hybrid-fallback stage plus fallback-dispatch log through the hybrid auth path so the open-adapter handoff is visible in lifecycle traces whenever Steamworks returns a retry-eligible result.
3. Refreshed the scripted auth-flow trace, the focused platform-service regression coverage, and the auth/backend gap notes so the documented QA evidence now matches the bounded compatibility story carried by the current source.

Task A3g: Expose explicit overall online-services mode/policy labels through the structural service-table surface [COMPLETED]

Priority: Critical Primary areas: src/common/platform/platform_services.c, src/code/client/cl_main.c, src/code/server/sv_init.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md Parity estimate: before 96% -> after 96%

Completed work:

1. Added common helpers that collapse the cached service table into overall online-services mode/policy labels, keeping the build-disabled, externally-disabled, Steamworks-only, open-adapter-only, and hybrid compatibility lanes explicit at the structural layer itself.
2. Mirrored those summary labels through the new ROM cvars cl_onlineServicesMode, cl_onlineServicesPolicy, sv_onlineServicesMode, and sv_onlineServicesPolicy so client/server diagnostics no longer require reading each per-feature provider cvar to understand the current repo-wide online-service boundary.
3. Refreshed the focused platform-service probes plus the RW-G01 platform-services/platform-config notes so the structural service-table and build-flag story now matches the bounded compatibility surface described by the current docs and tests.

Task A3h: Expose explicit overall online-services mode/policy labeling through auth policy-block and ticket-request failure paths [COMPLETED]

Priority: Critical Primary areas: src/code/client/ql_auth.c, tests/test_platform_services.py, docs/platform/authentication.md, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-client-auth.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the early Steam and standalone auth policy gates through a shared helper that now emits explicit policy-blocked lifecycle logs and response text naming the active overall online-services mode/policy lane instead of the older generic build/runtime wording.
2. Threaded that same structural mode/policy context through Steam ticket-request failure and backend-uninitialised responses so the bounded auth surface stays explicit even when no backend dispatch occurs.
3. Refreshed the focused auth/platform-service regression coverage plus the client-auth gap note and supporting docs so the documented RW-G01 compatibility story now matches those early auth exit paths too.

Task A3i: Expose explicit provider/policy labeling through the retained advert bridge surface [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_cgame.c, src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Mirrored the retained advert bridge through the new ROM cvars ui_advertisementBridgeProvider and ui_advertisementBridgePolicy so the active advert bridge lane is visible independently of the older browser overlay cvars.
2. Added bounded provider-aware advert lifecycle diagnostics for the UI/cgame bridge transitions (init-ui, activate, set-active, and shutdown-cgame) instead of leaving that seam implicit behind generic overlay state refreshes.
3. Refreshed the focused platform-service assertions plus the platform-service abstraction/gap-note docs so the retained advert flow now matches the same explicit compatibility-labeling story as the other active RW-G01 surfaces.

Task A3j: Expose explicit overall mode/policy labeling through the retained client voice fallback lane [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Mirrored the retained client voice compatibility lane through the new ROM cvars cl_voiceServiceMode and cl_voiceServicePolicy so the local speaking-state fallback no longer hides behind the broader online-services summary cvars alone.
2. Added explicit voice fallback diagnostics for +voice and -voice, naming the active overall online-services mode/policy lane whenever Steam voice is unavailable and the local speaking-state bridge is the remaining compatibility path.
3. Refreshed the focused platform-service assertions plus the platform-service abstraction/gap-note docs so the retained voice surface now matches the same explicit compatibility-labeling story as the other active RW-G01 seams.

Task A3k: Expose explicit overall mode/policy labeling through the retained client identity bootstrap and UI subscription lanes [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, src/code/client/cl_ui.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Mirrored the retained client identity/bootstrap and UI app-subscription lanes through the new ROM cvars cl_identityBootstrapMode, cl_identityBootstrapPolicy, ui_subscriptionBridgeMode, and ui_subscriptionBridgePolicy so these remaining Steam-owned shims no longer rely on the broader summary cvars alone for compatibility context.
2. Added explicit identity/subscription diagnostics for Steam persona-name seeding, Steam country seeding, and the UI IsSubscribedApp bridge, naming the active overall online-services mode/policy lane whenever the compatibility path falls back or the current provider is unavailable.
3. Refreshed the focused platform-service assertions plus the platform-service abstraction/gap-note docs so the retained client bootstrap and UI subscription surface now matches the same explicit compatibility-labeling story as the other active RW-G01 seams.

Task A3l: Expose explicit provider/policy labeling through the retained live-resource and avatar bridge lane [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_steam_resources.c, src/code/client/cl_main.c, src/code/client/cl_cgame.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Mirrored the retained live-resource bridge through the new ROM cvars ui_resourceBridgeProvider and ui_resourceBridgePolicy so the steam:///launcher-backed image lane is visible independently of the older browser overlay and advert bridge cvars.
2. Routed the cgame avatar import through the existing provider-aware resource stub path instead of returning before the compatibility diagnostics fired, which keeps disabled steam://avatar/... lookups explicit at runtime.
3. Refreshed the focused platform-service assertions plus the platform-service abstraction/gap-note docs so the retained live-resource and avatar lane now matches the same explicit compatibility-labeling story as the other active RW-G01 seams.

Task A3m: Expose explicit provider/policy labeling through the retained client workshop lifecycle [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, tests/test_client_workshop_bootstrap_parity.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Mirrored the retained client workshop seam through the ROM cvars cl_workshopProvider and cl_workshopPolicy so the active workshop owner is visible alongside the other client online-service descriptors.
2. Routed the workshop bootstrap, download, callback-ignore, queue-complete, and filesystem-restart diagnostics through a shared provider-aware helper, which keeps the current compatibility-only workshop lane explicit instead of falling back to generic Steamworks-only wording.
3. Refreshed the focused workshop/platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the documented workshop lifecycle now matches the same explicit compatibility-labeling story as the other active online-service seams.

Task A3n: Expose explicit provider/policy labeling through the retained dedicated-server P2P callback lane [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_client.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Added a shared dedicated-server P2P callback logger that reports the active Steam GameServer provider/policy pair for the retained session-request surface instead of leaving those diagnostics as raw Steam-only wording.
2. Routed the server-side P2P request callback through that helper for missing client, unauthenticated client, and accept-failure paths so the bounded compatibility lane stays explicit during those callback exits too.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the dedicated-server P2P callback path now matches the same explicit compatibility-labeling story as the other active online-service seams.

Task A3o: Expose explicit provider/policy labeling through the retained dedicated-server stats lane [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_init.c, src/code/server/sv_client.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Mirrored the retained dedicated-server stats seam through the ROM cvars sv_statsProvider and sv_statsPolicy so the active GameServerStats owner is visible alongside the other server online-service descriptors.
2. Added a shared provider-aware server-stats logger and routed the request-current-values, session-bootstrap P2P hello failure, and stat-delta diagnostics through it so this compatibility-only lane no longer hides behind raw Steam-only wording.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the dedicated-server stats path now matches the same explicit compatibility-labeling story as the other active online-service seams.

Task A3p: Expose explicit provider/policy labeling through the retained client rich-presence lifecycle [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, src/code/client/cl_cgame.c, src/code/client/client.h, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Exported the shared matchmaking-lane logger through client.h so the retained first-snapshot rich-presence owner can report the active provider/policy pair instead of silently depending on a Steam-only path.
2. Routed the main-menu and first-snapshot rich-presence writes through explicit provider-aware fallback diagnostics for provider-unavailable and update-failed exits while preserving the existing retail status strings.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained client rich-presence lifecycle now matches the same explicit compatibility-labeling story as the other active online-service seams.

Task A3q: Expose explicit provider/policy labeling through the retained dedicated-server auth rejection lane [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_client.c, tests/test_platform_services.py, docs/platform/authentication.md, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Added a shared dedicated-server auth rejection logger so the retained auth-session bootstrap reports the active provider/policy pair when session startup fails instead of falling back to raw Steam-only debug wording.
2. Routed both auth-session bootstrap failure exits through that helper while preserving the legacy outward Failed to authenticate with Steam: ... message contract for stable client-facing and telemetry consumers.
3. Refreshed the focused platform-service regression coverage plus the auth and platform-service docs so the retained dedicated-server auth rejection path now matches the same explicit compatibility-labeling story as the other active online-service seams.

Task A3r: Expose explicit provider/policy labeling through the retained client callback and stats-registration bootstrap gates [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the disabled-services early exit in CL_Steam_InitCallbacks() through the same provider-aware callback-bundle fallback diagnostic used by the registration-failure path, so the retained browser-event compatibility lane no longer stays silent when callbacks never become eligible.
2. Added explicit stats provider/policy-aware skip diagnostics to the stats_clear registration gate for provider-unavailable, initialisation- failed, and unsupported-app-id exits, which keeps that bootstrap boundary explicit even when the command is never registered.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained client callback/stats bootstrap path now matches the same explicit compatibility-labeling story as the other active online-service seams.

Task A3s: Expose explicit provider/policy labeling through the retained dedicated-server callback bootstrap stub [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_client.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the build-disabled SV_SteamServerInitCallbacks() stub through the same provider-aware callback-registration fallback diagnostic used by the active registration failure path, so the dedicated-server callback lane no longer disappears silently during startup.
2. Extended the focused platform-service regression coverage to pin that build-disabled callback-bootstrap fallback against the existing provider and policy labels.
3. Refreshed the platform-service abstraction and RW-G01 gap note so the retained dedicated-server callback bootstrap now matches the same explicit compatibility-labeling story as the other active online-service seams.

Task A3t: Expose explicit mode/policy labeling through the retained client voice transport lifecycle [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Added a shared voice-transport lifecycle logger so the retained Steam voice send/receive lane now reports the active overall mode/policy pair instead of falling back to raw generic trace wording when the transport fails.
2. Routed voice packet send failure, packet read failure, decompress failure, and zero-byte-decompress diagnostics through that helper so the retained voice transport now matches the same explicit compatibility story as the +voice / -voice command surface.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained client voice transport lifecycle now matches the same explicit compatibility-labeling story as the other active online-service seams.

Task A3u: Expose explicit provider/policy labeling through the retained dedicated-server stats-owner stubs [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_client.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Added a shared provider-aware stub logger for the build-disabled dedicated- server stats owner path so the retained SV_SteamStats_* surface no longer disappears silently when the current stats lane is unavailable.
2. Routed the build-disabled stat-delta, achievement-unlock, and achievement- query stubs through that helper with call-site detail strings, keeping the compatibility-only fallback explicit even when the owner functions still return without touching a live backend.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained dedicated-server stats-owner stubs now match the same explicit compatibility-labeling story as the other active online-service seams.

Task A3v: Expose explicit provider/policy labeling through the retained dedicated-server networking maintenance lane [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Added a shared provider-aware networking logger for the retained Steam GameServer maintenance owner so keepalive and packet-relay failures no longer collapse into silent compatibility-only behavior.
2. Routed keepalive send failure, inbound relay read failure, unknown relay sender, and relay-send failure exits through that helper, keeping the bounded Steam GameServer networking lane explicit whenever those paths fail.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained dedicated-server networking maintenance lane now matches the same explicit compatibility-labeling story as the other active online-service seams.

Task A3w: Expose explicit provider/policy labeling through the retained dedicated-server published-state owner [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Added a shared provider-aware published-state logger for the retained Steam GameServer metadata owner so publication writes no longer fail silently behind the bounded compatibility lane.
2. Routed max-player, password, hostname, map, game-description, game-tags, score key-value, player-data, and bot-count publication failures through that helper, keeping the retained GameServer metadata owner explicit when those writes cannot be applied.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained dedicated-server published-state owner now matches the same explicit compatibility-labeling story as the other active online-service seams.

Task A3x: Expose explicit provider/policy labeling through the retained client P2P callback lane [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Added a shared matchmaking callback logger for the retained client P2P session-request owner so this callback lane no longer falls back to a raw generic Steam trace.
2. Routed both the accept-failure and accepted-session callback exits through that helper with the requested remote SteamID detail, which keeps the bounded client matchmaking callback surface explicit whenever the callback fires.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained client P2P callback lane now matches the same explicit compatibility-labeling story as the other active online-service seams.

Task A3y: Expose explicit provider/policy labeling through the retained client browser-event publish lane [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Added a shared browser-event lifecycle logger for the retained client browser-event owner so this queue lane no longer falls back to generic PublishEvent failed or raw steam_event traces.
2. Routed the missing-view, missing-window-object, and queued-event exits through that helper with explicit queue metadata, keeping the retained overlay/browser seam honest about the current provider/policy pair whenever browser events are published.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained client browser-event publish lane now matches the same explicit compatibility labeling story as the other active online-service seams.

Task A3z: Expose explicit provider/policy labeling through the retained client microtransaction callback lane [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Added a shared microtransaction callback logger for the retained client purchase-authorization lane so this callback no longer falls back to a raw generic payload dump.
2. Routed the retained microtransaction authorization callback through that helper before it forwards the purchase update into the browser-event queue, keeping the overlay/browser compatibility owner explicit whenever this callback fires.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained client microtransaction callback lane now matches the same explicit compatibility-labeling story as the other active online-service seams.

Task A3aa: Expose explicit provider/policy labeling through the retained dedicated-server workshop operator surface [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_init.c, src/code/server/sv_ccmds.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Mirrored the retained dedicated-server workshop seam through the ROM cvars sv_workshopProvider and sv_workshopPolicy so the active workshop owner is visible alongside the other server online-service descriptors.
2. Added a shared provider-aware workshop operator logger plus a bounded Steam-UGC support gate for steam_downloadugc, steam_subscribeugc, and steam_unsubscribeugc, which keeps those Steam-specific operator commands explicit about the current compatibility lane instead of pretending that every workshop provider owns the same surface.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained server workshop operator surface now matches the same explicit compatibility-labeling story as the other active online-service seams.

Task A3ab: Expose explicit provider/policy labeling through the retained client web-host social and UGC export lanes [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_cgame.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Added retained web-host helpers for the overall online-service mode/policy plus matchmaking and workshop provider/policy labels, keeping the browser social/UGC export seams aligned with the repo-wide compatibility boundary.
2. Updated GetConfig, friend-list export, UGC export, and web.ugc.failed publication so browser callers see explicit provider/policy metadata while disabled or unavailable Steam-backed exports log bounded fallback diagnostics without changing the legacy array payload shape.
3. Refreshed focused regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained client web-host export lane now matches the same explicit compatibility-labeling story as the other active online-service seams.

Task A3ac: Expose explicit provider/policy labeling through the retained client lobby callback lane [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Extended the retained matchmaking callback diagnostics so the client lobby callback owner now reports provider/policy-aware lifecycle detail for lobby create, enter, membership, chat, metadata, game-created, kicked, and join-requested events instead of only making the bounded queue visible once the browser-event publication layer runs.
2. Kept the existing browser-event payload shapes intact while making the callback-origin seam explicit, which preserves the compatibility-only browser contract without letting those lobby callbacks read like silent Steam-owned behavior.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained client lobby callback lane now matches the same explicit compatibility-labeling story as the other active online-service seams.

Task A3ad: Expose explicit provider/policy labeling through the retained client callback connect-handoff lane [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained rich-presence join and server-change callback handoff through the shared matchmaking callback logger so the immediate join/connect command path now reports the active provider/policy pair instead of silently executing callback payloads.
2. Added bounded callback detail for missing join commands, missing server targets, and password-seeded server-connect handoff without changing the legacy immediate-command behavior or the outward connect payload shape.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained client callback connect-handoff lane now matches the same explicit compatibility-labeling story as the other active online-service seams.

Task A3ae: Expose explicit provider/policy labeling through the retained client stats callback lane [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Added a shared stats callback logger so the retained user-stats callback owner now reports the active stats provider/policy pair instead of only becoming visible once the browser-event queue publishes the resulting payload.
2. Routed the users.stats.*.received callback path through that helper with bounded SteamID, game-ID, and result detail while preserving the legacy browser-event payload shape and queue semantics.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained client stats callback lane now matches the same explicit compatibility-labeling story as the other active online-service seams.

Task A3af: Expose explicit provider/policy labeling through the retained client social-presence and UGC callback lanes [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained client persona-state and friend-rich-presence callback owners through the shared matchmaking callback logger so those browser-facing presence updates now report the active provider/policy pair instead of only becoming visible when the shared queue publishes them.
2. Routed the retained client UGC query-complete callback through the workshop lifecycle logger with bounded call/query/result detail while preserving the legacy web.ugc.results / web.ugc.failed payload shapes and browser event names.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained client social-presence and UGC callback lanes now match the same explicit compatibility-labeling story as the other active online-service seams.

Task A3ag: Expose explicit provider/policy labeling through the retained client workshop callback lane [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained client workshop ItemInstalled and DownloadItemResult callback owners through provider/policy-aware lifecycle diagnostics so tracked completion, untracked-item ignore, and result-failure exits are explicit instead of only becoming visible through the downstream queue helpers.
2. Kept the existing workshop queue behavior intact: item completion still flows through CL_Workshop_FinalizeInstalledItem(), failures still flow through CL_Workshop_FailActiveDownload(), and the legacy queue-advance behavior remains unchanged.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained client workshop callback lane now matches the same explicit compatibility-labeling story as the other active online-service seams.

Task A3ah: Expose explicit provider/policy labeling through the retained dedicated-server connection callback lane [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_client.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained Steam GameServer connected, connect-failure, and disconnected callbacks through a shared provider/policy-aware callback lifecycle logger so those server callback-owner state changes no longer appear as isolated one-off print lines.
2. Kept the existing server behavior intact: successful connects still publish identity and refreshed published state, failure/disconnect callbacks still clear the connected flag, and no auth/session routing changed.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained dedicated-server connection callback lane now matches the same explicit compatibility-labeling story as the other active online-service seams.

Task A3ai: Expose explicit provider/policy labeling through the retained client workshop callback bootstrap gate [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained client workshop callback registration-failure path through the shared workshop lifecycle logger so the polling-only fallback is explicit about the active provider/policy pair instead of appearing as a one-off raw debug line.
2. Kept the existing bootstrap behavior intact: the client still enables the main callback bundle, leaves workshop progress on the polling fallback when workshop callback registration fails, and does not change queue or browser event behavior.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained client workshop callback bootstrap gate now matches the same explicit compatibility-labeling story as the other active online-service seams.

Task A3aj: Expose explicit provider/policy labeling through the retained dedicated-server identity publication lane [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_init.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained Steam GameServer identity-publication owner through a shared provider/policy-aware lifecycle logger so both the unavailable and successful publish paths are explicit instead of only surfacing a raw identity-unavailable debug line.
2. Kept the existing identity behavior intact: the server still publishes the 0x2ca SteamID configstring, mirrors sv_referencedSteamworks, and writes the 0x2cb referenced-workshop configstring without changing the outward contract.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained dedicated-server identity publication lane now matches the same explicit compatibility-labeling story as the other active online-service seams.

Task A3ak: Expose explicit provider/policy labeling through the retained dedicated-server callback bootstrap gate [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_client.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained Steam GameServer callback-registration fallback through a provider/policy-aware bootstrap lifecycle logger in both the live registration-failure path and the build-disabled stub path instead of leaving those exits as raw one-off debug lines.
2. Kept the existing callback-registration behavior intact: successful registration still clears sv_steamServerConnected, while the unavailable paths still return immediately after surfacing the explicit compatibility diagnostic.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained dedicated-server callback bootstrap gate now matches the same explicit compatibility-labeling story as the other active server-owned seams.

Task A3al: Expose explicit provider/policy labeling through the retained client callback bootstrap gate [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained client callback-bundle fallback through a shared provider-aware bootstrap logger so the services-disabled and registration-failure exits now spell out the active matchmaking, stats, and social-overlay provider/policy pairs instead of staying as raw one-off debug lines.
2. Kept the existing callback bootstrap behavior intact: the client still clears the browser-event and lobby state before registration, still returns early when online services stay disabled, and still unregisters the partial client/lobby/micro callback bundle before falling back.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained client callback bootstrap gate now matches the same explicit compatibility-labeling story as the other active client-owned seams.

Task A3am: Expose explicit provider/policy labeling through the retained client workshop required-items bootstrap gate [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained client workshop required-items bootstrap fallback through the shared workshop lifecycle logger so the non-Steam/bootstrap- unavailable exit no longer stays as a raw one-off print line.
2. Kept the existing workshop bootstrap behavior intact: the client still logs the server-published required item list, still returns immediately when the current workshop lane cannot own Steam bootstrap, and still preserves the existing queue setup once a Steam UGC owner is available.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained client workshop required-items bootstrap gate now matches the same explicit compatibility-labeling story as the other active workshop seams.

Task A3an: Expose explicit provider/policy labeling through the retained dedicated-server stats request/session lifecycle lane [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_client.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed successful stats request issue plus the active-session reuse, fresh session-bootstrap, and backend reconnect requery transitions through the shared stats lifecycle logger, so those retained GameServerStats success paths now surface the same provider/policy pair as the existing failure diagnostics.
2. Kept the retained stats owner behavior intact: existing request issuance, session reset/bootstrap, P2P hello delivery, and reconnect-driven requery semantics all remain unchanged apart from the new explicit compatibility labels.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained dedicated-server stats owner now documents the same explicit compatibility-labeling story on both its request and session lifecycle paths.

Task A3ao: Expose explicit provider/policy labeling through the retained dedicated-server stats publish/store lane [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_client.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained stats pending-value flush path through the shared stats lifecycle logger so stat publish failures, achievement publish failures, store-request failure, and successful store completion now all surface the same provider/policy pair instead of failing silently inside the GameServerStats owner.
2. Kept the retained stats owner behavior intact: dirty stat/achievement accumulation, backend store submission, and dirty-bit clearing still follow the same recovered control flow, with the new compatibility labels only documenting the existing publish/store decisions.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained dedicated-server stats owner now documents the same explicit compatibility-labeling story across its request, session, and publish/store lifecycle paths.

Task A3ap: Expose explicit provider/policy labeling through the retained dedicated-server stats query/load lane [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_client.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained GetUserStatInt and GetUserAchievement fetch paths through the shared stats lifecycle logger so stat/achievement query success and failure now surface the same provider/policy pair instead of silently returning through the dedicated-server GameServerStats owner.
2. Kept the retained stats owner behavior intact: request issuance, cached value reuse, pending-delta accumulation, and achievement unlock decisions still follow the same recovered control flow, with the new compatibility labels only documenting the existing query/load outcomes.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained dedicated-server stats owner now documents the same explicit compatibility-labeling story across its request, session, query, and publish/store lifecycle paths.

Task A3aq: Expose explicit provider/policy labeling through the retained dedicated-server stats achievement owner/query lane [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_client.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained achievement-owner entry points through the shared stats lifecycle logger so invalid-achievement, gameplay-gated, unavailable, session-unavailable, already-held, queued-unlock, and ownership-result outcomes now surface the same provider/policy pair instead of returning silently through the active GameServerStats owner.
2. Kept the retained stats owner behavior intact: achievement gate checks, session bootstrap, cached ownership reuse, and pending unlock/store flows still follow the same recovered control flow, with the new compatibility labels only documenting the existing owner/query decisions.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained dedicated-server stats owner now documents the same explicit compatibility-labeling story across its request, session, query, achievement-owner, and publish/store lifecycle paths.

Task A3ar: Expose explicit provider/policy labeling through the retained dedicated-server stats field-owner lane [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_client.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained stat-delta owner entry point through the shared stats lifecycle logger so invalid/no-op, unavailable, session-unavailable, and baseline-unavailable queue decisions now surface the same provider/policy pair instead of returning silently through the active GameServerStats owner.
2. Kept the retained stats owner behavior intact: stat lookup, session bootstrap, pending-delta accumulation, and dirty-bit publication still follow the same recovered control flow, with the new compatibility labels only documenting the existing field-owner decisions.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained dedicated-server stats owner now documents the same explicit compatibility-labeling story across its request, session, query, field-owner, achievement-owner, and publish/store lifecycle paths.

Task A3as: Expose explicit provider/policy labeling through the retained dedicated-server stats session-teardown lane [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_client.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained stats session-teardown helper through the shared stats lifecycle logger so inactive-session skips and completed session clears now surface the same provider/policy pair instead of returning silently during auth/session shutdown.
2. Kept the retained teardown behavior intact: pending stat/achievement flush, session reset, and caller-owned auth shutdown sequencing still follow the same recovered control flow, with the new compatibility labels only documenting the existing session-clear decisions.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained dedicated-server stats owner now documents the same explicit compatibility-labeling story across its request, session, query, field-owner, achievement-owner, session-teardown, and publish/store lifecycle paths.

Task A3at: Expose explicit provider/policy labeling through the retained dedicated-server stats session-bootstrap gate [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_client.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained stats session-bootstrap gate through the shared stats lifecycle logger so null, out-of-range, zombie, missing-gentity, missing-SteamID, bot-owned, and invalid-SteamID skips now surface the same provider/policy pair instead of returning silently before session creation.
2. Kept the retained bootstrap behavior intact: valid clients still reuse live sessions, still create/reset sessions the same way, still request current values, and still send the Steam P2P hello packet with the same recovered control flow.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained dedicated-server stats owner now documents the same explicit compatibility-labeling story across its request, session-bootstrap, query, field-owner, achievement-owner, session-teardown, and publish/store lifecycle paths.

Task A3au: Expose explicit provider/policy labeling through the retained dedicated-server stats client-slot gate [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_client.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained stats client-slot helper through the shared stats lifecycle logger so out-of-range, inactive, zombie, missing-gentity, missing-SteamID, bot-owned, and invalid-SteamID gating now surface the same provider/policy pair instead of returning silently before the field and achievement owners can explain why the request stopped.
2. Kept the retained owner behavior intact: valid stat/achievement callers still resolve the same live client slot, still bootstrap or reuse sessions the same way, and still apply the same field/achievement control flow once a valid slot is available.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained dedicated-server stats owner now documents the same explicit compatibility-labeling story across its request, session-bootstrap, client-slot, query, field-owner, achievement-owner, session-teardown, and publish/store lifecycle paths.

Task A3av: Expose explicit provider/policy labeling through the retained dedicated-server stats request gate [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_client.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained SV_SteamStats_RequestCurrentValues gate through the shared stats lifecycle logger so null-session, inactive-session, and missing-SteamID request skips now surface the same provider/policy pair instead of returning silently before the existing request issue/failure diagnostics run.
2. Kept the retained request behavior intact: valid sessions still issue the same SteamGameServerStats request, still mark backendAvailable and requestIssued, and still preserve the same success/failure control flow once the request gate is satisfied.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained dedicated-server stats owner now documents the same explicit compatibility-labeling story across its request gate, session-bootstrap, client-slot, query, field-owner, achievement-owner, session-teardown, and publish/store lifecycle paths.

Task A3aw: Expose explicit provider/policy labeling through the retained dedicated-server stats value-query gate [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_client.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained stat/achievement value-query gate through the shared stats lifecycle logger so null-session, inactive-session, invalid-id, unmapped-id, and already-cached paths now surface the same provider/policy pair instead of returning silently before the existing query result diagnostics run.
2. Kept the retained query behavior intact: valid uncached stat and achievement requests still issue the same backend reads, still apply the same pending-delta merge or cached ownership update, and still preserve the same success/failure control flow once the query gate is satisfied.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained dedicated-server stats owner now documents the same explicit compatibility-labeling story across its request gate, session-bootstrap, client-slot, value-query gate, field-owner, achievement-owner, session-teardown, and publish/store lifecycle paths.

Task A3ax: Expose explicit provider/policy labeling through the retained dedicated-server stats value-flush gate [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_client.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained SV_SteamStats_FlushPendingValues gate through the shared stats lifecycle logger so null-session, inactive-session, missing-SteamID, and no-pending-update skips now surface the same provider/policy pair instead of returning silently before the existing publish/store diagnostics run.
2. Kept the retained flush behavior intact: dirty stat/achievement scans, backend publish attempts, store submission, and dirty-bit clearing still follow the same recovered control flow once the flush gate is satisfied.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained dedicated-server stats owner now documents the same explicit compatibility-labeling story across its request gate, session-bootstrap, client-slot, value-query gate, value-flush gate, field-owner, achievement-owner, session-teardown, and publish/store lifecycle paths.

Task A3ay: Expose explicit provider/policy labeling through the retained dedicated-server stats session-reset helper [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_client.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained SV_SteamStats_ResetSession helper through the shared stats lifecycle logger so null-session skips and retained-session clears now surface the same provider/policy pair instead of clearing state silently inside the dedicated-server GameServerStats owner.
2. Kept the retained reset behavior intact: session memory still clears with the same recovered Com_Memset flow, while the new compatibility labels only document when the helper is skipping or clearing retained state.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained dedicated-server stats owner now documents the same explicit compatibility-labeling story across its request gate, session-bootstrap, client-slot, value-query gate, value-flush gate, session-reset helper, field-owner, achievement-owner, session-teardown, and publish/store lifecycle paths.

Task A3az: Expose explicit provider/policy labeling through the retained dedicated-server stats descriptor lookup helpers [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_client.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained SV_SteamStats_GetFieldName and SV_SteamStats_GetAchievementName helpers through the shared stats lifecycle logger so invalid and unmapped descriptor lookups now surface the same provider/policy pair instead of returning silently beneath the dedicated-server GameServerStats owner.
2. Kept the retained stats behavior intact: the same caller-specific query, flush, stat-delta, and achievement owner paths still short-circuit on missing descriptors, while the helper signatures now accept the active lifecycle stage so those compatibility labels stay contextual.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so the retained dedicated-server stats owner now documents the same explicit compatibility-labeling story across its descriptor lookup helpers, request gate, session-bootstrap, client-slot, value-query gate, value-flush gate, session-reset helper, field-owner, achievement-owner, session-teardown, and publish/store lifecycle paths.

Task A3ba: Expose explicit provider/policy labeling through the retained client rich-presence join callback null-payload guard [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained CL_Steam_Client_OnRichPresenceJoinRequested null-event guard through the shared matchmaking callback logger so the compatibility lane now labels ignored null join payloads instead of silently returning.
2. Kept the retained callback handoff intact: non-null join payloads still flow through CL_Steam_OnRichPresenceJoinRequested without behavioral changes beyond the new compatibility labeling.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so this client callback owner now documents the same explicit compatibility-labeling story.

Task A3bb: Expose explicit provider/policy labeling through the retained client user-stats callback null-payload guard [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained CL_Steam_Client_OnUserStatsReceived null-event guard through the shared stats callback logger so ignored null payloads now surface the active provider/policy pair instead of dropping silently.
2. Kept the retained stats browser-event behavior intact for non-null payloads: user-stat JSON publishing still follows the same recovered control flow.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so this client callback owner now documents the same explicit compatibility-labeling story.

Task A3bc: Expose explicit provider/policy labeling through the retained client persona callback null-payload guard [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained CL_Steam_Client_OnPersonaStateChange null-event guard through the shared matchmaking callback logger so ignored null persona payloads no longer return silently.
2. Kept the retained persona browser-event behavior intact for non-null payloads: the same summary formatting and publish path still applies.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so this client callback owner now documents the same explicit compatibility-labeling story.

Task A3bd: Expose explicit provider/policy labeling through the retained client P2P-session callback null-payload guard [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained CL_Steam_Client_OnP2PSessionRequest null-event guard through the shared matchmaking callback logger so ignored null session payloads no longer return silently.
2. Kept the retained accept/fail logging and QL_Steamworks_AcceptP2PSession behavior intact for non-null payloads.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so this client callback owner now documents the same explicit compatibility-labeling story.

Task A3be: Expose explicit provider/policy labeling through the retained client server-change callback null-payload guard [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained CL_Steam_Client_OnGameServerChangeRequested null-event guard through the shared matchmaking callback logger so ignored null server-change payloads no longer return silently.
2. Kept the retained connect handoff intact: non-null payloads still flow through CL_Steam_OnGameServerChangeRequested without behavioral changes.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so this client callback owner now documents the same explicit compatibility-labeling story.

Task A3bf: Expose explicit provider/policy labeling through the retained client friend rich-presence callback null-payload guard [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained CL_Steam_Client_OnFriendRichPresenceUpdate null-event guard through the shared matchmaking callback logger so ignored null rich-presence payloads no longer return silently.
2. Kept the retained friend-summary formatting and browser-event publishing intact for non-null payloads.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so this client callback owner now documents the same explicit compatibility-labeling story.

Task A3bg: Expose explicit provider/policy labeling through the retained client UGC-query callback null-payload guard [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained CL_Steam_Client_OnUGCQueryCompleted null-event guard through the shared workshop lifecycle logger so ignored null query payloads no longer return silently.
2. Kept the retained UGC result/failure browser-event publishing intact for non-null payloads.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so this client callback owner now documents the same explicit compatibility-labeling story.

Task A3bh: Expose explicit provider/policy labeling through the retained client lobby-created callback null-payload guard [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained CL_Steam_Lobby_OnLobbyCreated null-event guard through the shared matchmaking callback logger so ignored null lobby-create payloads no longer return silently.
2. Kept the retained success/error lobby browser-event publishing intact for non-null payloads.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so this client callback owner now documents the same explicit compatibility-labeling story.

Task A3bi: Expose explicit provider/policy labeling through the retained client lobby-enter callback null-payload guard [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained CL_Steam_Lobby_OnLobbyEnter null-event guard through the shared matchmaking callback logger so ignored null lobby-enter payloads no longer return silently.
2. Kept the retained lobby-enter success/error publishing and current-lobby tracking intact for non-null payloads.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so this client callback owner now documents the same explicit compatibility-labeling story.

Task A3bj: Expose explicit provider/policy labeling through the retained client lobby-chat-update callback null-payload guard [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained CL_Steam_Lobby_OnLobbyChatUpdate null-event guard through the shared matchmaking callback logger so ignored null lobby-chat payloads no longer return silently.
2. Kept the retained user-summary formatting and join/leave browser-event publishing intact for non-null payloads.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so this client callback owner now documents the same explicit compatibility-labeling story.

Task A3bk: Expose explicit provider/policy labeling through the retained client lobby-chat-message callback null-payload guard [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained CL_Steam_Lobby_OnLobbyChatMessage null-event guard through the shared matchmaking callback logger so ignored null lobby-chat message payloads no longer return silently.
2. Kept the retained friend-summary lookup, JSON escaping, and browser-event publishing intact for non-null payloads.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so this client callback owner now documents the same explicit compatibility-labeling story.

Task A3bl: Expose explicit provider/policy labeling through the retained client lobby-data-update callback null-payload guard [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained CL_Steam_Lobby_OnLobbyDataUpdate null-event guard through the shared matchmaking callback logger so ignored null lobby-data update payloads no longer return silently.
2. Kept the retained lobby/member detail formatting and browser-event publishing intact for non-null payloads.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so this client callback owner now documents the same explicit compatibility-labeling story.

Task A3bm: Expose explicit provider/policy labeling through the retained client lobby-game-created callback null-payload guard [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained CL_Steam_Lobby_OnLobbyGameCreated null-event guard through the shared matchmaking callback logger so ignored null lobby-game created payloads no longer return silently.
2. Kept the retained game-server detail formatting and browser-event publishing intact for non-null payloads.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so this client callback owner now documents the same explicit compatibility-labeling story.

Task A3bn: Expose explicit provider/policy labeling through the retained client lobby-kicked callback null-payload guard [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained CL_Steam_Lobby_OnLobbyKicked null-event guard through the shared matchmaking callback logger so ignored null lobby-kicked payloads no longer return silently.
2. Kept the retained current-lobby clearing, detail formatting, and browser-event publishing intact for non-null payloads.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so this client callback owner now documents the same explicit compatibility-labeling story.

Task A3bo: Expose explicit provider/policy labeling through the retained client lobby-join-requested callback null-payload guard [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained CL_Steam_Lobby_OnGameLobbyJoinRequested null-event guard through the shared matchmaking callback logger so ignored null join-requested payloads no longer return silently.
2. Kept the retained lobby/friend detail formatting and browser-event publishing intact for non-null payloads.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so this client callback owner now documents the same explicit compatibility-labeling story.

Task A3bp: Expose explicit provider/policy labeling through the retained client microtransaction authorization callback null-payload guard [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained CL_Steam_Micro_OnAuthorizationResponse null-event guard through the shared microtransaction callback logger so ignored null authorization payloads no longer return silently.
2. Kept the retained purchase payload formatting and browser-event publishing intact for non-null payloads while extending the logger to spell out the null-payload compatibility path.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so this client callback owner now documents the same explicit compatibility-labeling story.

Task A3bq: Expose explicit provider/policy labeling through the retained client workshop item-installed callback null-payload guard [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained CL_Steam_Workshop_OnItemInstalled null-event guard through the shared workshop lifecycle logger so ignored null item-installed payloads no longer return silently.
2. Kept the retained app-id validation, item lookup, and install-finalization behavior intact for non-null payloads.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so this client callback owner now documents the same explicit compatibility-labeling story.

Task A3br: Expose explicit provider/policy labeling through the retained client workshop item-installed inactive-download-state guard [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained CL_Steam_Workshop_OnItemInstalled inactive-download guard through the shared workshop lifecycle logger so compatibility-owned installed callbacks now explain why no active download state can consume them.
2. Kept the retained unexpected-app-id, untracked-item, and install-complete handling intact once a live download state exists.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so this workshop callback guard now documents the same explicit compatibility-labeling story.

Task A3bs: Expose explicit provider/policy labeling through the retained client workshop download-result callback null-payload guard [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained CL_Steam_Workshop_OnDownloadItemResult null-event guard through the shared workshop lifecycle logger so ignored null download-result payloads no longer return silently.
2. Kept the retained app-id validation, active-item matching, and success/failure completion handling intact for non-null payloads.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so this workshop callback owner now documents the same explicit compatibility-labeling story.

Task A3bt: Expose explicit provider/policy labeling through the retained client workshop download-result inactive-lane guards [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained CL_Steam_Workshop_OnDownloadItemResult inactive-download-state and active-item-index guards through the shared workshop lifecycle logger so compatibility-owned download callbacks now explain why no active lane can consume them.
2. Kept the retained unexpected-app-id, inactive-item, and success/failure-completion handling intact once a live download lane exists.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so these workshop guard exits now document the same explicit compatibility-labeling story.

Task A3bu: Expose explicit provider/policy labeling through the retained dedicated-server connect-failure callback null-payload guard [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_client.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained SV_SteamServerConnectFailureCallback null-event guard through the shared server-callback lifecycle logger so ignored null connect-failure payloads no longer return silently.
2. Kept the retained connected-flag clear and failure-result formatting intact for non-null payloads.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so this dedicated-server callback owner now documents the same explicit compatibility-labeling story.

Task A3bv: Expose explicit provider/policy labeling through the retained dedicated-server disconnected callback null-payload guard [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_client.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained SV_SteamServerDisconnectedCallback null-event guard through the shared server-callback lifecycle logger so ignored null disconnect payloads no longer return silently.
2. Kept the retained connected-flag clear and disconnect-result formatting intact for non-null payloads.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so this dedicated-server callback owner now documents the same explicit compatibility-labeling story.

Task A3bw: Expose explicit provider/policy labeling through the retained dedicated-server auth-ticket callback null-payload guard [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_client.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained SV_SteamServerValidateAuthTicketResponseCallback null-event guard through the shared server-callback lifecycle logger so ignored null auth-ticket payloads no longer return silently.
2. Kept the retained fake-VAC override, auth-state publication, and accept/drop handling intact for non-null payloads.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so this dedicated-server callback owner now documents the same explicit compatibility-labeling story.

Task A3bx: Expose explicit provider/policy labeling through the retained dedicated-server auth-ticket missing-client guard [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_client.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained SV_SteamServerValidateAuthTicketResponseCallback missing-client guard through the shared server-callback lifecycle logger so auth responses for no-longer-tracked clients no longer return silently.
2. Kept the retained client lookup and finalise/drop behavior intact once one live client session still owns the callback.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so this dedicated-server callback owner now documents the same explicit compatibility-labeling story.

Task A3by: Expose explicit provider/policy labeling through the retained dedicated-server P2P-session-request callback null-payload guard [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_client.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained SV_SteamServerP2PSessionRequestCallback null-event guard through the shared P2P-session lifecycle logger so ignored null session-request payloads no longer return silently.
2. Kept the retained client lookup, auth gate, and accept-call failure handling intact for non-null payloads.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so this dedicated-server callback owner now documents the same explicit compatibility-labeling story.

Task A3bz: Expose explicit provider/policy labeling through the Steamworks servers-connected dispatch guard [COMPLETED]

Priority: Critical Primary areas: src/common/platform/platform_steamworks.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Added a shared QL_Steamworks_LogServerCallbackDispatch helper and routed the retained QL_Steamworks_DispatchServersConnected missing-state and missing-binding guards through it so that dispatcher no longer returns silently.
2. Kept the retained zeroed event bootstrap and callback invocation intact once one registered dispatch target exists.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so this dispatcher guard now documents the same explicit compatibility-labeling story.

Task A3ca: Expose explicit provider/policy labeling through the Steamworks connect-failure dispatch guards [COMPLETED]

Priority: Critical Primary areas: src/common/platform/platform_steamworks.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained QL_Steamworks_DispatchServerConnectFailure missing-state, missing-binding, and missing-payload guards through the shared server-dispatch logger so that dispatcher no longer returns silently.
2. Kept the retained raw-event translation and callback invocation intact once one valid dispatch surface and payload exist.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so these dispatcher guards now document the same explicit compatibility-labeling story.

Task A3cb: Expose explicit provider/policy labeling through the Steamworks disconnected dispatch guards [COMPLETED]

Priority: Critical Primary areas: src/common/platform/platform_steamworks.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained QL_Steamworks_DispatchServersDisconnected missing-state, missing-binding, and missing-payload guards through the shared server-dispatch logger so that dispatcher no longer returns silently.
2. Kept the retained raw-event translation and callback invocation intact once one valid dispatch surface and payload exist.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so these dispatcher guards now document the same explicit compatibility-labeling story.

Task A3cc: Expose explicit provider/policy labeling through the Steamworks auth-ticket dispatch guards [COMPLETED]

Priority: Critical Primary areas: src/common/platform/platform_steamworks.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained QL_Steamworks_DispatchValidateAuthTicketResponse missing-state, missing-binding, and missing-payload guards through the shared server-dispatch logger so that dispatcher no longer returns silently.
2. Kept the retained raw auth-ticket response translation and callback invocation intact once one valid dispatch surface and payload exist.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so these dispatcher guards now document the same explicit compatibility-labeling story.

Task A3cd: Expose explicit provider/policy labeling through the Steamworks P2P-session-request dispatch guards [COMPLETED]

Priority: Critical Primary areas: src/common/platform/platform_steamworks.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Routed the retained QL_Steamworks_DispatchServerP2PSessionRequest missing-state, missing-binding, and missing-payload guards through the shared server-dispatch logger so that dispatcher no longer returns silently.
2. Kept the retained raw P2P-session translation and callback invocation intact once one valid dispatch surface and payload exist.
3. Refreshed the focused platform-service regression coverage plus the platform-service abstraction and RW-G01 gap note so these dispatcher guards now document the same explicit compatibility-labeling story.

Task A3ce: Restore the retail direct-request workshop detail string [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the retained client workshop direct-request lane so it now reuses the retail-observed Workshop item %llu: requesting download. detail.
2. Kept the existing compatibility wrapper ownership and queue-state updates intact around that retail string restoration.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the workshop request lane.

Task A3cf: Restore the retail queued-request workshop detail string [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the retained client workshop bootstrap queueing lane so it now reuses the retail-observed Workshop item %llu: queueing download. detail.
2. Kept the existing deferred-download state model intact around that retail string restoration.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the queued-request lane.

Task A3cg: Restore the retail workshop cache-hit detail string [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the retained client workshop bootstrap cache-hit lane so it now reuses the retail-observed Workshop item %llu: in cache. detail.
2. Kept the existing cached-item completion handling intact around that retail string restoration.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the cache-hit lane.

Task A3ch: Restore the retail queued-handoff workshop detail string [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the retained client workshop queue-advance lane so it now reuses the retail-observed Workshop item %llu: was queued, requesting download. detail.
2. Kept the existing queue-advance ownership and state updates intact around that retail string restoration.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the queued handoff lane.

Task A3ci: Restore the retail workshop completion detail string [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the retained client workshop active-item completion lane so it now reuses the retail-observed Steamworks download complete: %llu detail.
2. Kept the existing active-download clear path intact around that retail string restoration.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the workshop completion lane.

Task A3cj: Restore the retail workshop queue-complete detail string [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the retained client workshop queue-settled lane so it now reuses the retail-observed Download completed for all steamworks items detail.
2. Kept the existing queue-active reset behavior intact around that retail string restoration.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the queue-complete lane.

Task A3ck: Restore the retail workshop invalid-app skip detail in the item-installed callback [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the retained client workshop item-installed invalid-app guard so it now reuses the retail-observed OnDownloadItemResult skip, invalid app id %d detail.
2. Kept the existing compatibility wrapper ownership intact around that observed retail string restoration.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the item-installed invalid-app guard.

Task A3cl: Restore the retail workshop invalid-app skip detail in the download-result callback [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the retained client workshop download-result invalid-app guard so it now reuses the retail-observed OnDownloadItemResult skip, invalid app id %d detail.
2. Kept the existing callback wrapper ownership intact around that retail string restoration.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the download-result invalid-app guard.

Task A3cm: Restore the retail workshop active-download skip detail [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the retained client workshop active-download mismatch guard so it now reuses the retail-observed OnDownloadItemResult skip, not the active download %llu detail.
2. Kept the existing active-item comparison and early-return behavior intact around that retail string restoration.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the active-download skip lane.

Task A3cn: Restore the retail workshop failure-result detail string [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the retained client workshop download-result failure lane so it now reuses the retail-observed Download item %llu failed with EResult code %i detail.
2. Kept the existing failure-state and queue-advance behavior intact around that retail string restoration.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the failure-result lane.

Task A3co: Restore the retail workshop missing-pak00.pk3 warning [COMPLETED]

Priority: Critical Primary areas: src/code/qcommon/files.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the filesystem-side workshop startup lane so it now reuses the retail-observed WARNING: Skipping workshop PK3s since pak00 doesn't exist. detail.
2. Kept the retained workshop mount pass active after that warning, matching the retail mount-mode toggle instead of turning the warning into a hard return.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the missing-pak00.pk3 guard.

Task A3cp: Restore the retail fs_skipWorkshop skip detail [COMPLETED]

Priority: Critical Primary areas: src/code/qcommon/files.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the filesystem-side workshop startup lane so the fs_skipWorkshop gate now reuses the retail-observed Skipping workshop since fs_skipWorkshop is set. detail.
2. Kept the existing early-return ownership intact around that exact retail string restoration.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the fs_skipWorkshop gate.

Task A3cq: Restore the retail workshop build-mode skip detail [COMPLETED]

Priority: Critical Primary areas: src/code/qcommon/files.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the filesystem-side workshop startup lane so the retained com_build/com_buildScript gate now reuses the retail-observed Skipping workshop since running in build mode. detail.
2. Kept the existing build-mode early return intact around that retail string restoration.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the build-mode gate.

Task A3cr: Restore the retail null-ISteamUGC skip detail [COMPLETED]

Priority: Critical Primary areas: src/code/qcommon/files.c, src/common/platform/platform_steamworks.c, src/common/platform/platform_steamworks.h, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the filesystem-side workshop startup lane so it now reuses the retail-observed WARNING: Skipping workshop, ISteamUGC is NULL. detail.
2. Kept the existing early-return behavior intact while splitting the null interface case away from the zero-subscribed-items case.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the null-ISteamUGC guard.

Task A3cs: Expose the exact workshop UGC-availability helper [COMPLETED]

Priority: Critical Primary areas: src/common/platform/platform_steamworks.c, src/common/platform/platform_steamworks.h, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Added QL_Steamworks_HasUGCInterface() so the retained workshop startup lane can distinguish a null ISteamUGC owner from an empty subscription list.
2. Kept the helper thin by routing it straight through the retained QL_Steamworks_GetUGCInterface() owner.
3. Refreshed the focused regression coverage for the new availability helper.

Task A3ct: Restore the retail basepath-only pak00.pk3 probe [COMPLETED]

Priority: Critical Primary areas: src/code/qcommon/files.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated FS_HasBasePak0 so it now mirrors the retail fs_basepath/baseq3/pak00.pk3 probe instead of scanning the broader local root set.
2. Kept the helper bounded to the workshop startup gate, matching the retail SteamWorkshop_Init ownership.
3. Refreshed the focused regression coverage for the basepath-only probe.

Task A3cu: Restore the retail workshop raw-path mount toggle [COMPLETED]

Priority: Critical Primary areas: src/code/qcommon/files.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the workshop mount pass so it now forwards the retained mountRawPath flag into FS_AddGameDirectoryInternal instead of always forcing the raw-path lane.
2. Mirrored the retail rule where the flag depends on whether pak00.pk3 was present when workshop startup began.
3. Refreshed the focused regression coverage for the mount-mode toggle.

Task A3cv: Restore the retail manual workshop direct-download detail [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_ccmds.c, tests/test_platform_services.py, tests/test_engine_operator_command_parity.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the retained steam_downloadugc immediate-request lane so it now reuses the retail-observed Workshop item %llu: download detail.
2. Kept the existing provider/policy-aware operator logger intact around that exact retail string restoration.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the direct-download lane.

Task A3cw: Restore the retail manual workshop cache-hit detail [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_ccmds.c, tests/test_platform_services.py, tests/test_engine_operator_command_parity.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the retained steam_downloadugc cache-hit lane so it now reuses the retail-observed Workshop item %llu: in cache. detail.
2. Kept the existing immediate-return ownership intact around that retail string restoration.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the manual cache-hit lane.

Task A3cx: Remove the non-retail manual download failure-only diagnostic [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_ccmds.c, tests/test_platform_services.py, tests/test_engine_operator_command_parity.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the retained steam_downloadugc immediate-request lane so it no longer emits the extra compatibility-only download request failed diagnostic that retail never printed.
2. Kept the existing command registration and provider-gated command surface intact while mirroring the retail immediate-request call shape more closely.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the removed non-retail failure-only diagnostic.

Task A3cy: Remove the non-retail client workshop request-failure gate [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated CL_Workshop_StartDownload so it now ignores the raw QL_Steamworks_DownloadItem return value like the retail SteamWorkshop_RequestDownload owner.
2. Kept the existing queue-state and active-item updates intact around that control-flow restoration.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the removed client request-failure gate.

Task A3cz: Remove the non-retail client workshop request-failure detail [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Removed the extra item %llu download request failed detail from CL_Workshop_StartDownload, which retail never printed.
2. Kept the retained retail request strings unchanged around that removal.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the removed request-failure detail.

Task A3da: Remove the duplicate client workshop failure trace [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated CL_Workshop_FailActiveDownload so it no longer emits a second item %llu failed with EResult code %i trace after the callback owner has already logged the retail failure detail.
2. Kept the existing completed-state and active-download clear behavior intact.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the removed duplicate failure trace.

Task A3db: Restore the thin retail steam_subscribeugc command shape [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_ccmds.c, tests/test_platform_services.py, tests/test_engine_operator_command_parity.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated SV_SteamCmd_SubscribeUGC_f so it now mirrors the retail thin wrapper shape instead of stopping early behind the compatibility-only workshop-support gate.
2. Kept the existing one-argument item-ID parse surface intact.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the subscribe command shape.

Task A3dc: Remove the non-retail subscribe-request trace [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_ccmds.c, tests/test_platform_services.py, tests/test_engine_operator_command_parity.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Removed the extra subscribe requested diagnostic from SV_SteamCmd_SubscribeUGC_f, which retail never printed.
2. Kept the retained command registration and parse surface unchanged.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the removed subscribe-request trace.

Task A3dd: Remove the non-retail subscribe-failure trace [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_ccmds.c, tests/test_platform_services.py, tests/test_engine_operator_command_parity.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Removed the extra subscribe request failed diagnostic from SV_SteamCmd_SubscribeUGC_f, which retail never printed.
2. Kept the retained direct wrapper call intact around that removal.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the removed subscribe-failure trace.

Task A3de: Restore the retail post-subscribe item-state refresh [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_ccmds.c, tests/test_platform_services.py, tests/test_engine_operator_command_parity.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated SV_SteamCmd_SubscribeUGC_f so it now re-reads QL_Steamworks_GetItemState after the subscribe call, matching the retail SteamWorkshop_SubscribeItem wrapper behavior.
2. Kept the existing one-argument item-ID parse surface intact.
3. Refreshed the focused regression coverage for the post-subscribe state refresh.

Task A3df: Restore the retail post-subscribe filesystem restart fast path [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_ccmds.c, tests/test_platform_services.py, tests/test_engine_operator_command_parity.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated SV_SteamCmd_SubscribeUGC_f so it now calls FS_Restart when the newly subscribed item is already installed, matching the retail installed- item fast path recovered from SteamWorkshop_SubscribeItem.
2. Kept the retained operator command surface intact around that restart hook.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the post-subscribe restart path.

Task A3dg: Restore the thin retail steam_unsubscribeugc command shape [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_ccmds.c, tests/test_platform_services.py, tests/test_engine_operator_command_parity.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated SV_SteamCmd_UnsubscribeUGC_f so it now mirrors the retail thin wrapper shape instead of stopping early behind the compatibility-only workshop-support gate.
2. Kept the existing one-argument item-ID parse surface intact.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the unsubscribe command shape.

Task A3dh: Remove the non-retail unsubscribe request/failure traces [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_ccmds.c, tests/test_platform_services.py, tests/test_engine_operator_command_parity.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Removed the extra unsubscribe requested and unsubscribe request failed diagnostics from SV_SteamCmd_UnsubscribeUGC_f, which retail never printed.
2. Kept the retained direct wrapper call intact around those removals.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the removed unsubscribe traces.

Task A3di: Restore the thin retail steam_downloadugc command helper ownership [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_ccmds.c, tests/test_platform_services.py, tests/test_engine_operator_command_parity.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated SV_SteamCmd_DownloadUGC_f so it now mirrors the retail thin wrapper shape by handing the parsed item ID off to SV_SteamWorkshop_RequestDownload instead of inlining the workshop control-flow directly in the operator command.
2. Kept the existing one-argument item-ID parse surface intact.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the recovered command-helper ownership split.

Task A3dj: Restore the local SteamWorkshop_RequestDownload helper ownership [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_ccmds.c, tests/test_platform_services.py, tests/test_engine_operator_command_parity.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Added SV_SteamWorkshop_RequestDownload so the compatibility-gated cache hit, restored retail detail strings, and QL_Steamworks_DownloadItem dispatch now sit under the same local helper boundary that the retail steam_downloadugc command calls into.
2. Kept the restored Workshop item %llu: in cache. and Workshop item %llu: download details intact around that helper recovery.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the recovered request-download helper ownership.

Task A3dk: Restore the thin retail steam_subscribeugc command helper ownership [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_ccmds.c, tests/test_platform_services.py, tests/test_engine_operator_command_parity.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated SV_SteamCmd_SubscribeUGC_f so it now mirrors the retail thin wrapper shape by handing the parsed item ID off to SV_SteamWorkshop_SubscribeItem instead of owning the workshop control flow directly.
2. Kept the existing one-argument item-ID parse surface intact.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the recovered subscribe-helper ownership split.

Task A3dl: Restore the thin retail steam_unsubscribeugc command helper ownership [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_ccmds.c, tests/test_platform_services.py, tests/test_engine_operator_command_parity.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated SV_SteamCmd_UnsubscribeUGC_f so it now mirrors the retail thin wrapper shape by handing the parsed item ID off to SV_SteamWorkshop_UnsubscribeItem instead of owning the unsubscribe call path directly.
2. Kept the existing one-argument item-ID parse surface intact.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the recovered unsubscribe-helper ownership split.

Task A3dm: Move the recovered post-subscribe item-state refresh under the local helper owner [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_ccmds.c, tests/test_platform_services.py, tests/test_engine_operator_command_parity.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated SV_SteamWorkshop_SubscribeItem so the recovered QL_Steamworks_GetItemState re-read now sits under the same local helper owner that the thin steam_subscribeugc command calls into.
2. Kept the direct QL_Steamworks_SubscribeItem wrapper call intact around that ownership correction.
3. Refreshed the focused regression coverage for the helper-owned post-subscribe state refresh.

Task A3dn: Move the recovered post-subscribe filesystem restart under the local helper owner [COMPLETED]

Priority: Critical Primary areas: src/code/server/sv_ccmds.c, tests/test_platform_services.py, tests/test_engine_operator_command_parity.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated SV_SteamWorkshop_SubscribeItem so the recovered installed-item FS_Restart fast path now sits under the same local helper owner that the thin steam_subscribeugc command calls into.
2. Kept the retained operator command surface intact around that helper-owned restart path.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the helper-owned post-subscribe restart path.

Task A3do: Restore the retail platform subscribe-wrapper item-state refresh [COMPLETED]

Priority: Critical Primary areas: src/common/platform/platform_steamworks.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated QL_Steamworks_SubscribeItem so it now re-reads QL_Steamworks_GetItemState after the raw subscribe vtable call, matching the retail SteamWorkshop_SubscribeItem helper shape.
2. Kept the recovered 0xc0 / 4 subscribe slot ownership intact around that state refresh.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the platform subscribe-wrapper state refresh.

Task A3dp: Remove the unconditional platform subscribe success return [COMPLETED]

Priority: Critical Primary areas: src/common/platform/platform_steamworks.c, tests/test_platform_services.py, tests/test_steamworks_harness.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Replaced the unconditional return qtrue; tail in QL_Steamworks_SubscribeItem with the recovered itemState != 0u result, matching the retail helper's post-subscribe return contract.
2. Kept the wrapper ABI and surrounding UGC interface ownership unchanged.
3. Refreshed the focused regression coverage and the matching abstraction/gap notes for the removed unconditional subscribe success return.

Task A3dq: Add harness coverage for the zero-state subscribe result [COMPLETED]

Priority: Critical Primary areas: tests/test_steamworks_harness.py, tests/steamworks_harness.c Parity estimate: before 96% -> after 96%

Completed work:

1. Extended the Steamworks harness so the enabled workshop helper test now asserts that QLR_Steamworks_SubscribeItem returns false when the post- subscribe item state remains zero.
2. Kept the existing disabled-lane and installed-item success expectations intact around that new regression.
3. Refreshed the harness support stubs so platform_steamworks.c continues to compile directly under the test fixture after the newer platform-services diagnostics were added.

Task A3dr: Add harness coverage for dedicated-server installed-state subscribe success [COMPLETED]

Priority: Critical Primary areas: tests/test_steamworks_harness.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the dedicated-server UGC ownership harness so it now primes the installed-item state before asserting subscribe success, matching the restored retail post-subscribe return contract.
2. Kept the existing dedicated-server UGC call-count and item-ID assertions intact.
3. Refreshed the focused harness coverage for the server-owned workshop lane.

Task A3ds: Restore the client bootstrap helper name to the retail request owner [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Renamed the retained client bootstrap helper to CL_Workshop_RequestDownload so the source-side owner now mirrors the retail SteamWorkshop_RequestDownload helper boundary again.
2. Dropped the synthetic queued-branch parameter from that helper.
3. Refreshed the focused regression coverage for the recovered helper owner.

Task A3dt: Remove the non-retail queued-handoff branch from the request helper [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Removed the queued-handoff branch from CL_Workshop_RequestDownload so it now owns only the initial Workshop item %llu: requesting download. detail recovered from retail SteamWorkshop_RequestDownload.
2. Kept the retained request-state updates intact around that narrower helper contract.
3. Refreshed the focused regression coverage for the removed queued branch.

Task A3du: Restore the retail queued-handoff detail under SteamWorkshop_AdvanceDownloadQueue [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Moved the retail Workshop item %llu: was queued, requesting download. detail under CL_Workshop_AdvanceQueue, matching the recovered SteamWorkshop_AdvanceDownloadQueue string owner.
2. Removed that detail from the initial-request helper owner.
3. Refreshed the focused regression coverage for the recovered queued-handoff detail owner.

Task A3dv: Restore the queued download dispatch under the queue-pop helper [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Moved the queued QL_Steamworks_DownloadItem dispatch under CL_Workshop_AdvanceQueue, matching the recovered retail queue-pop helper flow instead of reusing the initial-request helper.
2. Kept the retained queue-state bookkeeping intact around that owner shift.
3. Refreshed the focused regression coverage for the recovered dispatch owner.

Task A3dw: Restore the queued active-item handoff under the queue-pop helper [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Moved the queued CL_Workshop_SetActiveItem handoff under CL_Workshop_AdvanceQueue, matching the retail helper that promotes the next queued item before issuing its download request.
2. Kept the retained progress-cvar updates intact around that recovered owner split.
3. Refreshed the focused regression coverage for the queued active-item handoff owner.

Task A3dx: Restore the queue-advance owner split under SteamWorkshop_FinalizeItem [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated CL_Workshop_FinalizeInstalledItem so it now returns the CL_Workshop_AdvanceQueue result when the active item completes, matching the recovered retail SteamWorkshop_FinalizeItem helper flow.
2. Kept the retained completion log and active-download clear intact around that owner split.
3. Refreshed the focused regression coverage for the finalize-owned queue advance.

Task A3dy: Remove the explicit queued-handoff from the item-installed callback owner [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Removed the extra CL_Workshop_AdvanceQueue call from CL_Steam_Workshop_OnItemInstalled so the callback now leaves queued handoff ownership with CL_Workshop_FinalizeInstalledItem, matching retail.
2. Kept the retained null/app-id/tracked-item guard surface intact.
3. Refreshed the focused regression coverage for the callback-owner cleanup.

Task A3dz: Remove the explicit queued-handoff from the download-result success owner [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Removed the extra CL_Workshop_AdvanceQueue call from the successful CL_Steam_Workshop_OnDownloadItemResult lane so the completion path now mirrors the retail finalize-helper ownership split.
2. Kept the recovered invalid-app, active-download, and failure-result details intact.
3. Refreshed the focused regression coverage for the success-owner cleanup.

Task A3ea: Route bootstrap cache hits through the finalize helper owner [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the bootstrap cache-hit path so it now calls CL_Workshop_FinalizeInstalledItem after the retail Workshop item %llu: in cache. detail instead of marking the item complete inline.
2. Kept the retained cache-hit logging intact around that recovered owner split.
3. Refreshed the focused regression coverage for the finalize-owned cache-hit path.

Task A3eb: Update the bootstrap caller to use the recovered request helper split [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the client workshop bootstrap caller so the first request now goes through CL_Workshop_RequestDownload, while queued handoffs and completion handoffs stay with the recovered queue/finalize owners.
2. Refreshed the focused regression coverage plus the matching abstraction and gap-note documentation for the recovered helper ownership split.
3. Kept the bounded compatibility-only workshop provider/policy diagnostics intact around that tighter retail flow.

Task A3ec: Add explicit retained state for queued workshop items [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Added an explicit queued flag to the retained client workshop item state so the source can distinguish queued-for-later items from items that have already issued DownloadItem, matching the retail queue-container separation more closely.
2. Kept the retained requestNumber, downloadRequested, and completed fields intact around that narrower queue-state split.
3. Refreshed the focused regression coverage for the new queued-item state.

Task A3ed: Restore bootstrap queueing ownership under SteamWorkshop_RequestDownload [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated CL_Workshop_RequestDownload so it now owns the retail Workshop item %llu: queueing download. branch when a bootstrap download is already active, instead of leaving that detail under the caller.
2. Kept the recovered initial requesting download branch intact in the same helper.
3. Refreshed the focused regression coverage for the restored queueing-detail owner.

Task A3ee: Route every uncached bootstrap item through the recovered request helper [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the bootstrap loop so every uncached item now flows through CL_Workshop_RequestDownload, matching the retail CL_InitDownloads caller shape instead of splitting the queueing branch out in the caller.
2. Kept the retained cache-hit finalize path intact around that caller tightening.
3. Refreshed the focused regression coverage for the unified request-helper routing.

Task A3ef: Mark queued bootstrap items explicitly when the request helper defers them [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated CL_Workshop_RequestDownload so it now marks later bootstrap items as explicitly queued when the active download already exists, matching the recovered retail helper's queue-container behavior more closely.
2. Kept the retained active-download issue path intact for the first item.
3. Refreshed the focused regression coverage for the explicit queued-item mark.

Task A3eg: Make the queue-pop helper consume explicit queued items only [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated CL_Workshop_AdvanceQueue so it now consumes only items marked as queued, instead of inferring queue membership from !completed && !downloadRequested.
2. Kept the recovered queued-handoff detail and DownloadItem dispatch intact around that narrower scan.
3. Refreshed the focused regression coverage for the explicit queued scan.

Task A3eh: Clear the retained queued flag when a queued item becomes active [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated CL_Workshop_AdvanceQueue so it now clears the retained queued flag before issuing the queued item's DownloadItem request, matching the queue-pop ownership more closely.
2. Kept the retained downloadRequested tracking intact around that state handoff.
3. Refreshed the focused regression coverage for the queued-flag clear path.

Task A3ei: Restore bootstrap caller ownership of cl_downloadItem seeding [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Added CL_Workshop_SetDownloadRequestCvars and moved the cl_downloadItem seed under the bootstrap caller after successful request- helper returns, matching the retail CL_InitDownloads ownership split.
2. Kept the retained decimal SteamID formatting intact around that move.
3. Refreshed the focused regression coverage for the caller-owned item cvar.

Task A3ej: Restore bootstrap caller ownership of cl_downloadName seeding [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Moved the cl_downloadName seed under CL_Workshop_SetDownloadRequestCvars so the bootstrap caller now owns the Workshop item %i of %i request-label cvar after successful helper returns, matching retail.
2. Kept the retained requestNumber/totalItems labeling intact around that ownership split.
3. Refreshed the focused regression coverage for the caller-owned name cvar.

Task A3ek: Restore bootstrap caller ownership of cl_downloadTime seeding [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Moved the cl_downloadTime seed under CL_Workshop_SetDownloadRequestCvars so the bootstrap caller now owns the request timestamp after successful helper returns, matching the retail CL_InitDownloads call site.
2. Kept the retained cls.realtime source intact around that move.
3. Refreshed the focused regression coverage for the caller-owned time cvar.

Task A3el: Remove item/name/time cvar mutation from the active-item helper [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py, docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Removed the cl_downloadItem, cl_downloadName, and cl_downloadTime mutations from CL_Workshop_SetActiveItem, leaving that helper focused on active-download state and progress-byte refresh rather than bootstrap-cvar ownership.
2. Kept the then-retained UI-facing cl_downloadCount / cl_downloadSize temp-cvar updates intact around that narrower helper contract.
3. Refreshed the focused regression coverage plus the matching abstraction and gap-note documentation for the recovered active-item helper split.

Task A3em: Restore installed-item handling under the client request helper [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated CL_Workshop_RequestDownload so it now owns the installed-item gate, matching the retail SteamWorkshop_RequestDownload helper shape instead of relying on a bootstrap-caller pre-check.
2. Kept the retained queued and active-download branches intact around that helper-owned cache-hit path.
3. Refreshed the focused regression coverage for the recovered installed-item owner split.

Task A3en: Remove the bootstrap caller installed-state pre-check [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Removed the installed-state short-circuit from CL_Workshop_BeginBootstrap so parsed workshop items now flow through the recovered request helper regardless of whether they are already installed.
2. Kept the retained parsed-item count, truncation guard, and SteamID parse surface intact.
3. Refreshed the focused regression coverage for the removed caller pre-check.

Task A3eo: Restore the retail in cache detail under SteamWorkshop_RequestDownload [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Moved the retail Workshop item %llu: in cache. detail under CL_Workshop_RequestDownload, matching the recovered helper that owns that string in retail.
2. Removed that detail from the bootstrap caller owner.
3. Refreshed the focused regression coverage for the recovered cache-hit string owner.

Task A3ep: Restore finalize ownership for request-helper cache hits [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the request helper so installed-item cache hits now call CL_Workshop_FinalizeInstalledItem, matching the retail helper/finalize handoff.
2. Removed that finalize call from the bootstrap caller cache-hit branch.
3. Refreshed the focused regression coverage for the recovered finalize owner.

Task A3eq: Restore queue-gate activation for helper-owned cache hits [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Kept CL_Workshop_RequestDownload setting the retained queue gate active before the installed-item branch, matching the retail helper that marks the queue lane active even when a bootstrap item is already in cache.
2. Left the later queue-complete owner unchanged around that recovered gate activation.
3. Refreshed the focused regression coverage for the helper-owned queue gate.

Task A3er: Keep bootstrap request numbering limited to helper-success returns [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Kept item->requestNumber assignment under the bootstrap caller's successful CL_Workshop_RequestDownload return path, matching the retail CL_InitDownloads caller shape after the installed-item path moved under the helper.
2. Prevented cache-hit items from receiving bootstrap request ordinals.
3. Refreshed the focused regression coverage for the request-number owner.

Task A3es: Route every parsed bootstrap item through the recovered request helper [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Tightened CL_Workshop_BeginBootstrap so every parsed workshop item now flows through CL_Workshop_RequestDownload, matching the retail caller that invokes SteamWorkshop_RequestDownload for each token.
2. Kept the retained truncation and parse guards intact around that caller recovery.
3. Refreshed the focused regression coverage for the per-token request-helper routing.

Task A3et: Restore queue-complete eligibility for all-cached bootstrap passes [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. By moving cache-hit handling under CL_Workshop_RequestDownload, the retained bootstrap lane now marks the queue gate active even when every required item is already installed, matching the retail path that still runs the later queue-complete helper.
2. Kept the retained no-download request-number and cvar behavior intact for those cache-hit items.
3. Refreshed the focused regression coverage for the all-cached queue-complete eligibility lane.

Task A3eu: Update the focused workshop bootstrap regression for helper-owned cache hits [COMPLETED]

Priority: Critical Primary areas: tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the focused client workshop regression so it now asserts the installed-item in cache detail and finalize handoff under CL_Workshop_RequestDownload instead of under the bootstrap caller.
2. Added a direct assertion that the helper keeps the retained queue gate active around that installed-item branch.
3. Kept the existing queued and active-download assertions intact.

Task A3ev: Refresh the abstraction notes for the helper-owned cache-hit split [COMPLETED]

Priority: Critical Primary areas: docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md, IMPLEMENTATION_PLAN.md Parity estimate: before 96% -> after 96%

Completed work:

1. Refreshed the abstraction notes so they now document the recovered SteamWorkshop_RequestDownload cache-hit ownership split and the retained all-cached queue-gate behavior.
2. Recorded the next ten micro-closures in the implementation ledger as A3em through A3ev.
3. Kept the compatibility-boundary framing explicit around those retail-backed helper corrections.

Task A3ew: Restore the plain retail workshop bootstrap announcement string [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated CL_Workshop_BeginBootstrap so it now prints the recovered retail Server requires the following workshop items: %s announcement instead of the newer provider/policy-flavored compatibility detail.
2. Kept the retained bootstrap-unavailable compatibility log intact around that exact-string recovery.
3. Refreshed the focused regression coverage for the restored announcement string.

Task A3ex: Remove the non-retail bootstrap provider/policy announcement locals [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Removed the bootstrap-only workshopProvider / workshopPolicy locals from CL_Workshop_BeginBootstrap after restoring the plain retail announcement string.
2. Kept the later provider/policy-aware fallback diagnostics intact elsewhere in the workshop lane.
3. Refreshed the focused regression coverage for the removed announcement locals.

Task A3ey: Update the bootstrap regression for the restored retail announcement [COMPLETED]

Priority: Critical Primary areas: tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the focused client workshop bootstrap regression so it now asserts the plain retail announcement string and the absence of the removed provider/policy announcement locals.
2. Kept the existing request-helper and queue-owner assertions intact.
3. Reconfirmed the bootstrap regression against the current worktree.

Task A3ez: Restore the retail frame restart-required string [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated CL_Workshop_Frame so the restart-required path now prints the recovered retail Steamworks downloads complete - FS restart is required detail.
2. Kept the retained FS_Restart( clc.checksumFeed ) behavior and downloadsRequested reset intact around that exact-string recovery.
3. Refreshed the focused regression coverage for the restored restart-required string.

Task A3fa: Remove the compatibility-only frame restart lifecycle trace [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Removed the compatibility-only downloads complete; restarting filesystem workshop lifecycle trace from the frame restart path, which retail never printed.
2. Replaced it with the recovered plain restart-required print under the same owner.
3. Refreshed the focused regression coverage for the removed compatibility trace.

Task A3fb: Restore the retail frame completion string before pk3 validation [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Added the recovered Steamworks downloads complete print to the no-restart branch of CL_Workshop_Frame, matching the retail completion helper before it performs the missing-pk3 compare.
2. Kept the retained state-clear and CL_DownloadsComplete handoff ordering unchanged around that exact-string recovery.
3. Refreshed the focused regression coverage for the restored completion string.

Task A3fc: Restore the retail missing-pk3 warning header under the frame helper [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Replaced the non-retail WARNING: You are missing some files referenced by the server: wording in CL_Workshop_Frame with the recovered retail WARNING: Missing pk3s referenced by the server: header.
2. Kept the retained FS_ComparePaks( ..., qfalse ) gate intact around that warning path.
3. Refreshed the focused regression coverage for the restored warning header.

Task A3fd: Restore the retail missing-pk3 consequence line [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Replaced the newer You might not be able to join the game / Go to the setting menu... warning tail in CL_Workshop_Frame with the recovered retail The server will most likely refuse the connection. line.
2. Kept the retained missing-file payload insertion intact around that exact warning tail.
3. Refreshed the focused regression coverage for the restored consequence line.

Task A3fe: Add focused regression coverage for the retail frame strings [COMPLETED]

Priority: Critical Primary areas: tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Added a dedicated CL_Workshop_Frame regression so the restart-required, downloads-complete, and missing-pk3 warning strings now stay pinned to the recovered retail literals.
2. Included negative assertions for the removed compatibility-only restart and missing-file wording.
3. Kept the retained FS_Restart, state-clear, and CL_DownloadsComplete owner assertions intact.

Task A3ff: Refresh the abstraction notes for the restored bootstrap/frame string surface [COMPLETED]

Priority: Critical Primary areas: docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md, IMPLEMENTATION_PLAN.md Parity estimate: before 96% -> after 96%

Completed work:

1. Refreshed the abstraction notes so they now document the restored plain workshop bootstrap announcement plus the recovered retail frame restart, completion, and missing-pk3 warning strings.
2. Recorded the next ten micro-closures in the implementation ledger as A3ew through A3ff.
3. Kept the compatibility-boundary framing explicit around those retail-backed exact-string corrections.

Task A3fg: Restore the retained no-restart workshop completion lane to the retail active-gate teardown [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Rechecked the retail CL_Workshop_Frame HLIL helper and confirmed that the no-restart lane clears only the outer workshop-active gate before tailing into CL_DownloadsComplete().
2. Updated the retained frame helper to drop only cl_steamWorkshopDownloadState.active on the no-restart completion path instead of zeroing the full bootstrap state.
3. Refreshed the focused frame regression so it now pins that exact handoff.

Task A3fh: Remove the compatibility-only bootstrap-state reset from the retained workshop completion frame owner [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Removed the retained CL_Workshop_ClearBootstrapState( qtrue ) call from the workshop no-restart completion lane because the retail helper does not clear the full bootstrap structure there.
2. Kept the recovered retail completion/warning strings and the CL_DownloadsComplete() handoff intact.
3. Added the matching negative assertion to the focused regression.

Task A3fi: Refresh the workshop frame regression for the retail active-gate handoff [COMPLETED]

Priority: Critical Primary areas: tests/test_platform_services.py, IMPLEMENTATION_PLAN.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the focused frame test to assert the retained active-gate clear instead of the older bootstrap-state reset.
2. Kept the surrounding retail restart-required, downloads-complete, and missing-pk3 string checks intact.
3. Recorded the handoff correction in the implementation ledger.

Task A3fj: Bound retained workshop progress ownership to the recovered UI bridge [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Cross-checked the recovered uix86 workshop-progress path and confirmed that it reads cl_downloadItem, calls the native GetItemDownloadInfo import, and reads cl_downloadTime rather than consuming cl_downloadCount / cl_downloadSize.
2. Used that evidence to trim the retained workshop helper ownership back to the native-item download-info bridge instead of generic download cvars.
3. Added focused evidence coverage for that UI bridge sequence.

Task A3fk: Remove the compatibility-only generic progress-cvar write from the retained workshop active-item helper [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Removed the retained CL_Workshop_UpdateProgressCvars() call from CL_Workshop_SetActiveItem.
2. Kept the active-item byte reset and immediate native CL_Workshop_RefreshProgress() query intact.
3. Refreshed the focused helper regression to pin the trimmed cvar surface.

Task A3fl: Remove the compatibility-only generic progress-cvar write from the retained workshop clear-active helper [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Removed the retained CL_Workshop_UpdateProgressCvars() call from CL_Workshop_ClearActiveDownload.
2. Kept the retained active-item and byte-count clear itself intact.
3. Added the matching negative regression assertion for the helper body.

Task A3fm: Refresh the retained workshop helper regression for the trimmed progress-cvar surface [COMPLETED]

Priority: Critical Primary areas: tests/test_platform_services.py, IMPLEMENTATION_PLAN.md Parity estimate: before 96% -> after 96%

Completed work:

1. Extended the workshop helper test to extract CL_Workshop_ClearActiveDownload.
2. Added explicit negative assertions covering both the active-item helper and the clear-active helper so future edits do not reintroduce generic progress-cvar churn on workshop queue handoffs.
3. Recorded that focused regression tightening in the plan ledger.

Task A3fn: Pin the recovered uix86 workshop progress bridge evidence [COMPLETED]

Priority: Critical Primary areas: tests/test_platform_services.py, references/reverse-engineering/ghidra/uix86/source-recreation/ui_reconstruction.c Parity estimate: before 96% -> after 96%

Completed work:

1. Added a focused evidence test covering the recovered workshop progress bridge in ui_reconstruction.c.
2. Pinned the cl_downloadItem read, native GetItemDownloadInfo import call, and cl_downloadTime read as the companion evidence backing the retained client-side ownership trim.
3. Included negative assertions showing that same recovered bridge does not use cl_downloadCount or cl_downloadSize.

Task A3fo: Refresh the abstraction notes for the workshop completion and progress-bridge ownership correction [COMPLETED]

Priority: Critical Primary areas: docs/steam_platform_abstraction.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the abstraction note so it now documents the retail no-restart frame handoff as an active-gate clear rather than a full bootstrap-state reset.
2. Recorded the recovered uix86 workshop progress bridge evidence and the resulting removal of retained generic progress-cvar churn from the workshop active-item and clear-active helpers.
3. Kept the broader online-service compatibility-boundary framing explicit.

Task A3fp: Refresh the source-gap notes and ledger for the workshop completion/progress ownership correction [COMPLETED]

Priority: Critical Primary areas: docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md, IMPLEMENTATION_PLAN.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the RW-G01 source-gap note to match the restored workshop frame teardown and progress-bridge ownership story.
2. Recorded the next ten micro-closures in the implementation ledger as A3fg through A3fp.
3. Kept the evidence wording explicit so the batch is framed as retail-backed ownership correction rather than freehand compatibility instrumentation.

Task A3fq: Restore the retained workshop queue-pop helper name to the recovered retail owner [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Rechecked the retail alias/evidence for sub_469400 and renamed the retained queue-pop helper to CL_Workshop_AdvanceDownloadQueue so it matches the recovered SteamWorkshop_AdvanceDownloadQueue owner.
2. Updated the focused workshop regression to extract/assert that renamed helper directly.
3. Kept the broader workshop bootstrap ownership split intact around the recovered helper family.

Task A3fr: Move the retained active-download clear under the shared queue-pop owner [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Rechecked the retail SteamWorkshop_AdvanceDownloadQueue HLIL and restored the retained active-download clear under the shared queue-pop helper before it scans for queued items.
2. Removed the need for completion/failure callers to open-code that clear.
3. Added focused regression coverage pinning the retained CL_Workshop_ClearActiveDownload() call inside the queue-pop owner.

Task A3fs: Route the retained finalize helper through the shared queue-pop owner [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated CL_Workshop_FinalizeInstalledItem so the active-item completion lane now tails directly into CL_Workshop_AdvanceDownloadQueue().
2. Removed the explicit active-download clear from the finalize helper body because that teardown now belongs to the shared queue-pop owner.
3. Refreshed the focused helper regression so it now asserts the shared completion handoff instead of the old inline clear.

Task A3ft: Route the retained active-download failure helper through the shared queue-pop owner [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated CL_Workshop_FailActiveDownload so it now marks the active item completed and then tails into CL_Workshop_AdvanceDownloadQueue().
2. Kept the helper bounded to retained state management only; the retail failure detail still stays in the callback owner.
3. Added the matching regression assertion for the shared failure handoff.

Task A3fu: Remove the ignored retained EResult parameter from the workshop failure helper [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Dropped the unused result parameter from CL_Workshop_FailActiveDownload because the retained helper never consumed it and the retail shared queue-pop handoff does not take that result code.
2. Kept the actual failure-detail formatting anchored to the DownloadItemResult callback owner where the retail string lives.
3. Updated the focused callback/helper regression to match the trimmed helper signature.

Task A3fv: Remove the callback-inline queue advance from the retained workshop download-result owner [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_main.c, tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Removed the retained CL_Workshop_AdvanceDownloadQueue() call from the CL_Steam_Workshop_OnDownloadItemResult failure lane.
2. Kept the retail failure detail log intact and let the failure helper own the shared queue-pop handoff instead.
3. Added a focused negative assertion so that callback owner no longer grows a second inline queue-advance call.

Task A3fw: Refresh the retained workshop callback regression for the shared failure handoff [COMPLETED]

Priority: Critical Primary areas: tests/test_platform_services.py, IMPLEMENTATION_PLAN.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the workshop callback regression so the download-result owner now asserts CL_Workshop_FailActiveDownload() without the older ignored result argument.
2. Added the matching negative assertion covering the removed callback-inline queue-advance call.
3. Recorded that callback-owner tightening in the implementation ledger.

Task A3fx: Refresh the retained workshop helper regression for the shared queue-pop teardown [COMPLETED]

Priority: Critical Primary areas: tests/test_platform_services.py, IMPLEMENTATION_PLAN.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the focused helper regression to extract CL_Workshop_AdvanceDownloadQueue.
2. Added assertions covering the shared active-download clear in the queue-pop helper plus the new finalize/failure return handoffs.
3. Tightened the downloads-settled assertion so it now references the renamed shared owner as well.

Task A3fy: Refresh the abstraction notes for the retained workshop failure/queue-pop owner split [COMPLETED]

Priority: Critical Primary areas: docs/steam_platform_abstraction.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the abstraction note so it now documents the recovered DownloadItemResult failure handoff into the shared SteamWorkshop_AdvanceDownloadQueue owner.
2. Recorded that the retained queue-pop helper now owns the active-download clear before scanning queued items.
3. Kept the framing explicit that this batch is a bounded retail ownership correction inside the already-bounded online-services lane.

Task A3fz: Refresh the source-gap notes and ledger for the retained workshop failure/queue-pop owner split [COMPLETED]

Priority: Critical Primary areas: docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md, IMPLEMENTATION_PLAN.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the RW-G01 note to match the recovered shared queue-pop ownership for both workshop completion and failure lanes.
2. Recorded the next ten micro-closures in the implementation ledger as A3fq through A3fz.
3. Kept the evidence wording explicit so the batch stays framed as retail helper-owner correction rather than new compatibility instrumentation.

Task A3ga: Rename the retained client UI workshop progress import parameters to the recovered low/high split [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_ui.c, tests/test_client_workshop_bootstrap_parity.py Parity estimate: before 96% -> after 96%

Completed work:

1. Rechecked the companion uix86 workshop-progress reconstruction and confirmed that the native GetItemDownloadInfo import is reached from the parsed cl_downloadItem low/high words.
2. Renamed the retained QL_UI_trap_GetItemDownloadInfo parameters from generic arg1 / arg2 to itemIdLow / itemIdHigh to match that recovered split without changing runtime behavior.
3. Refreshed the focused client workshop import assertion to pin the renamed retained bridge.

Task A3gb: Clarify the retained client workshop progress bridge comment against the recovered uix86 owner [COMPLETED]

Priority: Critical Primary areas: src/code/client/cl_ui.c Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the retained client bridge comment so it now records that the companion uix86 reconstruction parses cl_downloadItem, calls the native import with the low/high words, and then reads cl_downloadTime.
2. Kept the retained bridge behavior unchanged: it still prefers retained client workshop state and falls back to the platform Steamworks helper.
3. Preserved the workshop progress import as a naming/evidence correction rather than a behavior change.

Task A3gc: Rename the retained UI syscall wrapper parameters to the recovered low/high split [COMPLETED]

Priority: Critical Primary areas: src/code/ui/ui_syscalls.c, tests/test_ui_menu_files.py Parity estimate: before 96% -> after 96%

Completed work:

1. Renamed trap_QL_GetItemDownloadInfo to use itemIdLow / itemIdHigh parameter names instead of the older itemHi / itemLo wording.
2. Kept the retained native import call order unchanged while making that order explicit in the wrapper surface.
3. Refreshed the UI menu syscall regression to pin the renamed wrapper signature.

Task A3gd: Rename the retained UI local declaration for the workshop progress import to the recovered low/high split [COMPLETED]

Priority: Critical Primary areas: src/code/ui/ui_local.h Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the retained ui_local.h declaration for trap_QL_GetItemDownloadInfo to use itemIdLow / itemIdHigh.
2. Kept the declaration aligned with the wrapper implementation and retained import cast signature.
3. Preserved the rest of the native UI import surface unchanged.

Task A3ge: Extend the recovered workshop progress evidence test to pin the parsed item-ID split before the import call [COMPLETED]

Priority: Critical Primary areas: tests/test_platform_services.py Parity estimate: before 96% -> after 96%

Completed work:

1. Extended the companion uix86 evidence test so it now asserts the recovered sscanf( "%llu", ... ) parse before the GetItemDownloadInfo import call.
2. Kept the existing assertions covering the native import and cl_downloadTime read intact.
3. Tightened the evidence trail for the retained low/high parameter naming.

Task A3gf: Refresh the retained workshop progress import regression for the renamed client bridge parameters [COMPLETED]

Priority: Critical Primary areas: tests/test_client_workshop_bootstrap_parity.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the focused workshop progress import regression so it now extracts the renamed QL_UI_trap_GetItemDownloadInfo signature.
2. Refreshed the retained bridge assertions to use itemIdLow / itemIdHigh in both the client-state and platform fallback calls.
3. Kept the negative assertions against generic download-count/size ownership intact.

Task A3gg: Refresh the CL-G03 parity gate for the retained workshop progress import naming correction [COMPLETED]

Priority: Critical Primary areas: tests/test_client_full_parity_gate.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the CL-G03 parity gate so it now looks for the renamed itemIdLow / itemIdHigh signature on the retained client workshop progress bridge.
2. Kept the broader workshop publication/bootstrap/mount ownership gate unchanged.
3. Preserved the gate as a source-shape check rather than runtime behavior.

Task A3gh: Refresh the UI menu import-surface regression for the retained workshop progress naming correction [COMPLETED]

Priority: Critical Primary areas: tests/test_ui_menu_files.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the UI menu native-import regression so it now asserts the renamed trap_QL_GetItemDownloadInfo( itemIdLow, itemIdHigh, ... ) wrapper.
2. Kept the rest of the recovered advert/cursor/subscription/measure-text import surface unchanged.
3. Recorded the import-surface naming correction in the focused regression.

Task A3gi: Refresh the abstraction note for the recovered workshop progress import low/high split [COMPLETED]

Priority: Critical Primary areas: docs/steam_platform_abstraction.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the abstraction note so it now explicitly records the recovered GetItemDownloadInfo low/high-word call shape from the parsed cl_downloadItem.
2. Kept the surrounding workshop progress-bridge ownership story intact.
3. Preserved the framing as companion-evidence-backed naming clarification.

Task A3gj: Refresh the source-gap notes and ledger for the retained workshop progress import low/high split [COMPLETED]

Priority: Critical Primary areas: docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md, IMPLEMENTATION_PLAN.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the RW-G01 note so it now explicitly describes the recovered low/high-word GetItemDownloadInfo bridge.
2. Recorded the next ten micro-closures in the implementation ledger as A3ga through A3gj.
3. Kept the batch framed as an evidence-backed interface-naming correction rather than a behavior change.

Task A3gk: Refresh the legacy workshop bootstrap parity suite for the recovered request-helper owner [COMPLETED]

Priority: Critical Primary areas: tests/test_client_workshop_bootstrap_parity.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the legacy workshop parity suite to extract CL_Workshop_RequestDownload instead of the removed pre-audit CL_Workshop_StartDownload helper.
2. Replaced the stale request-helper assertions with the current retail-backed in cache, requesting download, and queueing download owner split.
3. Kept the broader workshop bootstrap coverage anchored to current retained source instead of older compatibility wording.

Task A3gl: Refresh the legacy workshop bootstrap parity suite for the recovered shared queue-pop owner [COMPLETED]

Priority: Critical Primary areas: tests/test_client_workshop_bootstrap_parity.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the legacy workshop parity suite to extract CL_Workshop_AdvanceDownloadQueue.
2. Added the current shared-owner assertions for the active-download clear, queued-item clear, and retained queued download request.
3. Removed the older assumptions that queue progression stayed inline under callback or finalize owners.

Task A3gm: Refresh the legacy workshop bootstrap parity suite for the recovered finalize/failure helper signatures [COMPLETED]

Priority: Critical Primary areas: tests/test_client_workshop_bootstrap_parity.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the suite to match the current qboolean CL_Workshop_FinalizeInstalledItem / CL_Workshop_FailActiveDownload helper shapes.
2. Replaced the stale failure-helper expectation that still passed an ignored EResult parameter.
3. Added the shared queue-pop handoff assertions for both completion and failure helpers.

Task A3gn: Refresh the legacy workshop bootstrap parity suite for the recovered bootstrap exact-string surface [COMPLETED]

Priority: Critical Primary areas: tests/test_client_workshop_bootstrap_parity.py Parity estimate: before 96% -> after 96%

Completed work:

1. Replaced the stale provider/policy-decorated bootstrap string expectations with the restored retail Server requires the following workshop items: %s announcement.
2. Updated the bootstrap-unavailable expectation to the current workshop lifecycle log detail instead of the older raw print wording.
3. Kept the request-number/download-cvar ownership assertions aligned with the current retained bootstrap caller.

Task A3go: Refresh the legacy workshop bootstrap parity suite for the recovered callback bootstrap and failure-owner split [COMPLETED]

Priority: Critical Primary areas: tests/test_client_workshop_bootstrap_parity.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the callback-bootstrap expectation to the current CL_LogWorkshopLifecycle( "callback-bootstrap", detail ) fallback lane.
2. Refreshed the item-installed and download-result callback assertions to match the current null-payload, invalid-app, untracked-item, and active-download guards.
3. Added the current shared failure-handoff expectation so the callback owner no longer expects an inline queue advance.

Task A3gp: Refresh the legacy workshop bootstrap parity suite for the recovered workshop frame string surface [COMPLETED]

Priority: Critical Primary areas: tests/test_client_workshop_bootstrap_parity.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the workshop frame assertions to the restored retail restart-required and downloads-complete print strings.
2. Replaced the older compatibility-only restart log expectation with the current retail output plus missing-pk3 warning assertions.
3. Kept the frame-order and CL_DownloadsComplete() ownership checks intact.

Task A3gq: Refresh the legacy workshop bootstrap parity suite for the current queue-complete/request-cvar ownership split [COMPLETED]

Priority: Critical Primary areas: tests/test_client_workshop_bootstrap_parity.py Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the queue-complete assertion to the current Download completed for all steamworks items detail.
2. Kept the request-number / cl_downloadItem / cl_downloadName / cl_downloadTime ownership assertions aligned with the retained bootstrap caller and helper split.
3. Preserved the workshop progress owner assertions already tightened in the newer focused tests.

Task A3gr: Restore the full legacy workshop bootstrap parity suite to passing status [COMPLETED]

Priority: Critical Primary areas: tests/test_client_workshop_bootstrap_parity.py Parity estimate: before 96% -> after 96%

Completed work:

1. Re-ran the full legacy workshop parity suite after the assertion refresh.
2. Confirmed the older coverage now passes again instead of failing on removed pre-audit helper names.
3. Kept the suite as supplemental validation beside the newer platform-services coverage.

Task A3gs: Refresh the abstraction note for the restored legacy workshop validation alignment [COMPLETED]

Priority: Critical Primary areas: docs/steam_platform_abstraction.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the abstraction note so it now records that focused validation keeps the recovered workshop owner split aligned across both the platform-services and legacy workshop parity suites.
2. Kept the note framed as validation coverage around the already-documented recovered owner split.
3. Avoided re-framing the batch as a runtime behavior change.

Task A3gt: Refresh the source-gap notes and ledger for the restored legacy workshop validation alignment [COMPLETED]

Priority: Critical Primary areas: docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md, IMPLEMENTATION_PLAN.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the RW-G01 note so it now records that the recovered workshop owner split is validated in both focused workshop suites.
2. Recorded the next ten micro-closures in the implementation ledger as A3gk through A3gt.
3. Kept the batch framed as validation-gap closure on top of already-restored retail-backed behavior.

Task A3gu: Refresh the Round 05 workshop wrapper note for the retained client-state progress bridge [COMPLETED]

Priority: Critical Primary areas: docs/reverse-engineering/quakelive_steam_mapping_round_05.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the Round 05 SteamUGC_GetItemDownloadInfo note so it no longer describes the older legacy-download-counter fallback.
2. Recorded that the retained UI import now consults retained client workshop state first and only then falls back to QL_Steamworks_GetItemDownloadInfo, the retained wrapper over the same low/high-word SteamUGC_GetItemDownloadInfo slot.
3. Kept the wrapper promotion itself unchanged.

Task A3gv: Refresh the Round 14 workshop import-call note for the recovered parsed low/high split [COMPLETED]

Priority: Critical Primary areas: docs/reverse-engineering/quakelive_steam_mapping_round_14.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the Round 14 local-fact note so it now records the parsed cl_downloadItem low/high-word handoff before the UI import call.
2. Removed the stale (itemHi, itemLo, ...) wording.
3. Kept the underlying wrapper identification unchanged.

Task A3gw: Refresh the Round 14 workshop import-owner note for the retained wrapper name and fallback story [COMPLETED]

Priority: Critical Primary areas: docs/reverse-engineering/quakelive_steam_mapping_round_14.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the Round 14 note so import 96 is described as QL_UI_trap_GetItemDownloadInfo instead of the older placeholder wrapper name.
2. Recorded the retained-client-state-first progress bridge plus the direct Steam item-info fallback.
3. Kept the wrapper role framed as a UI-owned import seam.

Task A3gx: Refresh the Round 14 alias summary for the recovered workshop import-entry wording [COMPLETED]

Priority: Critical Primary areas: docs/reverse-engineering/quakelive_steam_mapping_round_14.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the alias-summary row for QLUIImport_GetItemDownloadInfo.
2. Recorded that the wrapper is reached from the parsed cl_downloadItem low/high words.
3. Kept the alias candidate itself unchanged.

Task A3gy: Refresh the CL-G03 audit note for the retained workshop request-surface cvars [COMPLETED]

Priority: Critical Primary areas: docs/reverse-engineering/client-full-parity-audit-and-implementation-plan-2026-04-09.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the CL-G03 closure note so it now records the retained cl_downloadItem / cl_downloadName / cl_downloadTime request surface instead of the older byte-counter wording.
2. Kept the surrounding retained bootstrap-owner summary intact.
3. Avoided re-framing the lane as a new runtime change.

Task A3gz: Refresh the CL-G03 audit note for the retained shared queue-pop callback-owner split [COMPLETED]

Priority: Critical Primary areas: docs/reverse-engineering/client-full-parity-audit-and-implementation-plan-2026-04-09.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the CL-G03 closure note so it now records the shared queue-pop owner for installed-item completion and download-result failure.
2. Removed the older phrasing that implied direct callback-owned queue advancement.
3. Kept the polling-fallback note intact.

Task A3ha: Refresh the CL-P3 closure note for the retained workshop request-surface cvars [COMPLETED]

Priority: Critical Primary areas: docs/reverse-engineering/client-full-parity-audit-and-implementation-plan-2026-04-09.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the CL-P3 completed-work note so it now records cl_downloadItem / cl_downloadName / cl_downloadTime instead of the older byte-counter wording.
2. Kept the retained active/queued workshop bootstrap summary intact.
3. Preserved the closure note as historical documentation rather than a new behavior claim.

Task A3hb: Refresh the CL-P3 closure note for the retained shared finalize/failure helper handoff [COMPLETED]

Priority: Critical Primary areas: docs/reverse-engineering/client-full-parity-audit-and-implementation-plan-2026-04-09.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the CL-P3 callback-exactness closure note so it now records the shared finalize/failure helper handoff into SteamWorkshop_AdvanceDownloadQueue.
2. Removed the older immediate callback-owned queue-advancement wording.
3. Kept the listed validation suites intact.

Task A3hc: Add focused validation for the stale reverse-engineering workshop note cleanup and tighten the CL-G03 parity gate [COMPLETED]

Priority: Critical Primary areas: tests/test_client_workshop_bootstrap_parity.py, tests/test_client_full_parity_gate.py Parity estimate: before 96% -> after 96%

Completed work:

1. Added a focused regression covering the refreshed Round 05 / Round 14 / client-parity-plan workshop notes.
2. Tightened the CL-G03 parity gate so it now requires the current request- surface and shared queue-pop wording in the client parity plan.
3. Kept the gate scoped to current workshop ownership/validation shape rather than runtime artifacts.

Task A3hd: Refresh the abstraction notes, source-gap notes, and ledger for stale reverse-engineering workshop note cleanup [COMPLETED]

Priority: Critical Primary areas: docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md, IMPLEMENTATION_PLAN.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the abstraction/source-gap notes so they now explicitly record that the older reverse-engineering workshop notes were brought back into line with the recovered low/high import wording and shared queue-pop owner story.
2. Recorded the next ten micro-closures in the implementation ledger as A3gu through A3hd.
3. Kept the batch framed as stale-note cleanup and validation tightening on top of already-restored retail-backed behavior.

Task A3he: Refresh the client CVar note for the retained workshop request-surface cvars [COMPLETED]

Priority: Critical Primary areas: docs/client_cvars.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the client CVar note so it now records cl_downloadItem, cl_downloadName, and cl_downloadTime as the retained workshop request-surface trio.
2. Anchored that wording to the recovered CL_InitDownloads ownership.
3. Kept the note scoped to documentation rather than new runtime behavior.

Task A3hf: Refresh the client CVar note for the recovered UI progress bridge [COMPLETED]

Priority: Critical Primary areas: docs/client_cvars.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the same note so it now records the parsed cl_downloadItem low/high-word handoff into the native GetItemDownloadInfo probe.
2. Kept cl_downloadTime in the documented retained progress surface.
3. Avoided reintroducing the older legacy-byte-counter wording.

Task A3hg: Refresh the client CVar note for the non-authoritative counter cvars [COMPLETED]

Priority: Critical Primary areas: docs/client_cvars.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the note so it now records cl_downloadCount and cl_downloadSize as UI-facing temp cvars rather than the authoritative workshop progress owner.
2. Kept the non-Steam fallback guidance explicit.
3. Left the underlying CVar registration untouched.

Task A3hh: Refresh the CL-P3 closure note for the retained/direct workshop progress bridge [COMPLETED]

Priority: Critical Primary areas: docs/reverse-engineering/client-full-parity-audit-and-implementation-plan-2026-04-09.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the historical CL-P3 closure note so it now describes the retained-client-state-first workshop progress bridge plus the direct GetItemDownloadInfo fallback keyed by parsed cl_downloadItem words.
2. Removed the older generic "retained client owner" shorthand.
3. Kept the rest of the closure summary intact.

Task A3hi: Add focused validation for the client workshop CVar-note alignment [COMPLETED]

Priority: Critical Primary areas: tests/test_client_workshop_bootstrap_parity.py Parity estimate: before 96% -> after 96%

Completed work:

1. Extended the focused workshop-note regression so it now reads docs/client_cvars.md.
2. Added assertions for the retained request-surface trio, direct GetItemDownloadInfo bridge wording, and the non-authoritative cl_downloadCount / cl_downloadSize note.
3. Kept the test scoped to evidence-note drift rather than runtime behavior.

Task A3hj: Tighten the CL-G03 parity gate for current client workshop CVar wording [COMPLETED]

Priority: Critical Primary areas: tests/test_client_full_parity_gate.py Parity estimate: before 96% -> after 96%

Completed work:

1. Added docs/client_cvars.md as a tracked CL-G03 documentation input.
2. Tightened the gate so it now requires the current retained/direct workshop progress wording in both the client CVar note and the client parity plan.
3. Added a dedicated client_cvar_notes_current detail lane for that check.

Task A3hk: Refresh the abstraction note for the aligned workshop CVar note [COMPLETED]

Priority: Critical Primary areas: docs/steam_platform_abstraction.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the abstraction note so it now calls out the aligned docs/client_cvars.md wording.
2. Recorded that cl_downloadItem, cl_downloadName, and cl_downloadTime remain the retained workshop request/progress bridge.
3. Recorded that cl_downloadCount / cl_downloadSize stay UI-facing temp cvars instead of the authoritative owner.

Task A3hl: Refresh the RW-G01 source-gap note for the aligned workshop CVar note [COMPLETED]

Priority: Critical Primary areas: docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the RW-G01 note so it now calls out the aligned docs/client_cvars.md wording alongside the workshop progress bridge note.
2. Recorded the same retained trio versus UI-temp-counter split there.
3. Kept the note framed as documentation alignment over already-restored behavior.

Task A3hm: Refresh the implementation ledger for the workshop CVar-note cleanup [COMPLETED]

Priority: Critical Primary areas: IMPLEMENTATION_PLAN.md Parity estimate: before 96% -> after 96%

Completed work:

1. Recorded the client workshop CVar-note cleanup as explicit micro-closures instead of leaving it implicit.
2. Kept the ledger framing evidence-backed and documentation-oriented.
3. Preserved the active A3 scope summary unchanged.

Task A3hn: Record the next ten micro-closures for the client workshop CVar-note cleanup batch [COMPLETED]

Priority: Critical Primary areas: IMPLEMENTATION_PLAN.md Parity estimate: before 96% -> after 96%

Completed work:

1. Recorded this batch in the implementation ledger as A3he through A3hn.
2. Kept the batch framed as stale-note cleanup and validation tightening on top of already-restored retail-backed behavior.
3. Avoided re-framing the batch as a runtime code change.

Task A3ho: Refresh the Round 05 workshop wrapper note for the retained direct-helper name [COMPLETED]

Priority: Critical Primary areas: docs/reverse-engineering/quakelive_steam_mapping_round_05.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the Round 05 SteamUGC_GetItemDownloadInfo note so it now names QL_Steamworks_GetItemDownloadInfo directly instead of the older generic "direct Steam item-info probe" wording.
2. Kept the retained-client-state-first UI import story intact.
3. Preserved the wrapper promotion itself unchanged.

Task A3hp: Refresh the Round 14 workshop import-call note for the retained direct-helper name [COMPLETED]

Priority: Critical Primary areas: docs/reverse-engineering/quakelive_steam_mapping_round_14.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the Round 14 local-fact note so it now names QL_Steamworks_GetItemDownloadInfo as the retained fallback helper.
2. Kept the parsed cl_downloadItem low/high-word handoff explicit.
3. Avoided changing the underlying wrapper identification.

Task A3hq: Refresh the Round 14 alias summary for the retained direct-helper bridge wording [COMPLETED]

Priority: Critical Primary areas: docs/reverse-engineering/quakelive_steam_mapping_round_14.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the alias-summary row for QLUIImport_GetItemDownloadInfo.
2. Recorded that the retained cl_ui.c bridge falls back to QL_Steamworks_GetItemDownloadInfo.
3. Kept the alias candidate itself unchanged.

Task A3hr: Refresh the CL-G03 observed-source note for the retained direct-helper bridge [COMPLETED]

Priority: Critical Primary areas: docs/reverse-engineering/client-full-parity-audit-and-implementation-plan-2026-04-09.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the historical CL-G03 observed-source note so it now names QL_Steamworks_GetItemDownloadInfo instead of the older generic direct Steam-item probe wording.
2. Kept the retained-client-state-first bridge summary intact.
3. Left the surrounding workshop mount/bootstrap notes unchanged.

Task A3hs: Refresh the CL-P3 closure note for the retained direct-helper bridge [COMPLETED]

Priority: Critical Primary areas: docs/reverse-engineering/client-full-parity-audit-and-implementation-plan-2026-04-09.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the historical CL-P3 closure note so it now names QL_Steamworks_GetItemDownloadInfo as the retained fallback helper.
2. Kept the retained workshop owner split and legacy-counter exclusion intact.
3. Preserved the closure note as historical documentation rather than a new runtime change.

Task A3ht: Refresh the CL-P3 exit criteria for the recovered low/high workshop progress bridge [COMPLETED]

Priority: Critical Primary areas: docs/reverse-engineering/client-full-parity-audit-and-implementation-plan-2026-04-09.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the workshop-progress exit criterion so it now names the recovered GetItemDownloadInfo low/high-word bridge explicitly.
2. Kept the negative contrast with the older generic legacy counters.
3. Avoided re-framing the lane as a behavior change.

Task A3hu: Add focused validation for the refreshed workshop helper-note wording [COMPLETED]

Priority: Critical Primary areas: tests/test_client_workshop_bootstrap_parity.py Parity estimate: before 96% -> after 96%

Completed work:

1. Extended the focused workshop-note regression to require the refreshed Round 05 / Round 14 helper wording.
2. Added assertions for the current client-parity-plan wording around the retained/direct helper split and the low/high-word exit criterion.
3. Kept the test scoped to evidence-note drift rather than runtime behavior.

Task A3hv: Tighten the CL-G03 parity gate for the refreshed workshop helper-note wording [COMPLETED]

Priority: Critical Primary areas: tests/test_client_full_parity_gate.py Parity estimate: before 96% -> after 96%

Completed work:

1. Tightened the CL-G03 parity gate so it now requires the refreshed retained/direct helper wording in the client parity plan.
2. Kept the existing retained request-surface and shared queue-pop checks intact.
3. Preserved the gate as a documentation/validation guard, not a runtime behavior claim.

Task A3hw: Refresh the abstraction and RW-G01 notes for the retained direct-helper wording alignment [COMPLETED]

Priority: Critical Primary areas: docs/steam_platform_abstraction.md, docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the abstraction and RW-G01 notes so they now record that the older client-parity notes were aligned to the explicit QL_Steamworks_GetItemDownloadInfo fallback naming.
2. Kept the retained-client-state progress bridge and shared queue-pop owner story intact.
3. Left the surrounding workshop/service-boundary notes unchanged.

Task A3hx: Record the next ten micro-closures for the workshop helper-note wording cleanup batch [COMPLETED]

Priority: Critical Primary areas: IMPLEMENTATION_PLAN.md Parity estimate: before 96% -> after 96%

Completed work:

1. Recorded this batch in the implementation ledger as A3ho through A3hx.
2. Kept the batch framed as stale-note cleanup and validation tightening on top of already-restored retail-backed behavior.
3. Avoided re-framing the batch as a runtime code change.

Task A3hy: Refresh the top-level CL-P3 summary for the retained direct-helper name [COMPLETED]

Priority: Critical Primary areas: docs/reverse-engineering/client-full-parity-audit-and-implementation-plan-2026-04-09.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the top-level CL-P3 summary so it now names QL_Steamworks_GetItemDownloadInfo instead of the older generic direct probe wording.
2. Kept the retained-client-state-first bridge summary intact.
3. Preserved the surrounding parity summary unchanged.

Task A3hz: Refresh the top-level CL-P3 summary for the recovered low/high-word bridge wording [COMPLETED]

Priority: Critical Primary areas: docs/reverse-engineering/client-full-parity-audit-and-implementation-plan-2026-04-09.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the same summary so it now records the retail SteamUGC_GetItemDownloadInfo low/high-word slot more explicitly.
2. Kept the parsed cl_downloadItem ownership visible in the wording.
3. Avoided reintroducing generic helper shorthand.

Task A3ia: Refresh the historical A3gu implementation-plan note for the retained direct-helper name [COMPLETED]

Priority: Critical Primary areas: IMPLEMENTATION_PLAN.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the older A3gu completed-work note so it now names QL_Steamworks_GetItemDownloadInfo directly.
2. Kept the retained-client-state-first import story intact.
3. Preserved the historical task framing.

Task A3ib: Refresh the historical A3el implementation-plan note for UI-facing temp-cvar wording [COMPLETED]

Priority: Critical Primary areas: IMPLEMENTATION_PLAN.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the older A3el note so it now describes the then-retained cl_downloadCount / cl_downloadSize updates as UI-facing temp cvars.
2. Kept the note historically accurate for that intermediate step.
3. Avoided implying those counters were ever the authoritative progress owner.

Task A3ic: Add focused validation for the top-level CL-P3 workshop helper wording [COMPLETED]

Priority: Critical Primary areas: tests/test_client_workshop_bootstrap_parity.py Parity estimate: before 96% -> after 96%

Completed work:

1. Extended the focused workshop-note regression to require the refreshed top-level CL-P3 summary wording.
2. Kept the existing Round 05 / Round 14 note checks intact.
3. Left the test scoped to evidence-note drift rather than runtime behavior.

Task A3id: Add focused validation for the historical implementation-plan workshop wording [COMPLETED]

Priority: Critical Primary areas: tests/test_client_workshop_bootstrap_parity.py Parity estimate: before 96% -> after 96%

Completed work:

1. Extended the same regression so it now reads IMPLEMENTATION_PLAN.md.
2. Added assertions for the refreshed A3gu helper wording and A3el temp-cvar wording.
3. Kept the validation surface narrow and evidence-backed.

Task A3ie: Tighten the CL-G03 parity gate for the refreshed implementation-plan workshop wording [COMPLETED]

Priority: Critical Primary areas: tests/test_client_full_parity_gate.py Parity estimate: before 96% -> after 96%

Completed work:

1. Tightened the CL-G03 parity gate so it now requires the refreshed top-level client-plan wording plus the aligned implementation-plan notes.
2. Added a dedicated implementation_plan_notes_current detail lane for that check.
3. Preserved the gate as a documentation/validation guard, not a runtime behavior claim.

Task A3if: Refresh the abstraction note for the aligned top-level summary and implementation-plan wording [COMPLETED]

Priority: Critical Primary areas: docs/steam_platform_abstraction.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the abstraction note so it now calls out the aligned top-level CL-P3 summary and older implementation-plan workshop note wording.
2. Recorded the explicit helper naming and UI-facing temp-cvar framing there.
3. Left the broader workshop boundary story unchanged.

Task A3ig: Refresh the RW-G01 note for the aligned top-level summary and implementation-plan wording [COMPLETED]

Priority: Critical Primary areas: docs/reverse-engineering/source-file-gap-notes/rw-g01-platform-services.md Parity estimate: before 96% -> after 96%

Completed work:

1. Updated the RW-G01 note so it now records the aligned top-level CL-P3 summary and implementation-plan workshop wording.
2. Kept the explicit helper naming and UI-facing temp-cvar framing visible.
3. Preserved the surrounding compatibility-boundary note unchanged.

Task A3ih: Record the next ten micro-closures for the workshop summary-and-ledger wording cleanup batch [COMPLETED]

Priority: Critical Primary areas: IMPLEMENTATION_PLAN.md Parity estimate: before 96% -> after 96%

Completed work:

1. Recorded this batch in the implementation ledger as A3hy through A3ih.
2. Kept the batch framed as stale-note cleanup and validation tightening on top of already-restored retail-backed behavior.
3. Avoided re-framing the batch as a runtime code change.

Task A3ii: Close RW-G01 as a documented bounded divergence across the repo-wide ledgers [COMPLETED]

Priority: Critical Primary areas: AUDIT.md, IMPLEMENTATION_PLAN.md, docs/reverse-engineering/repo-wide-parity-audit-2026-04-21.md, docs/reverse-engineering/source-file-parity-ledger-2026-04-22.md, docs/reverse-engineering/source-file-parity-audit-plan-2026-04-22.md, docs/reverse-engineering/source-file-gap-notes/, tools/reverse-engineering/generate_source_file_audit.py Parity estimate: before 96% -> after 98%

Completed work:

1. Reclassified the RW-G01 online-services lane as an intentional documented divergence instead of active repo-wide parity debt, while keeping the strict-retail Windows target unchanged at 100%.
2. Updated the source-file audit generator and regenerated the plan, ledger, and per-file note set so the seven RW-G01 owners now land under documented-divergence notes while RW-G02 remains the only active file-level compatibility gap family.
3. Refreshed the top-level repo-wide ledgers so the checked-in whole-repo estimate now reads 98%, with RW-G02 and RW-G04 left as the active remaining repo-wide gap families.

Task A4j: Add a bounded silent Linux sound sink and close the OSS shutdown leak [COMPLETED]

Priority: Medium Primary areas: src/code/unix/linux_snd.c, tests/test_non_windows_portability.py, docs/reverse-engineering/source-file-gap-notes/rw-g02-linux-snd.md, docs/platform/toolchain-matrix.md, docs/build/linux-glibc-32bit.md Parity estimate: before 98% -> after 98%

Completed work:

1. Added an explicit silent Linux DMA sink selected through snddevice null, none, or silent, giving headless/client smoke probes a non-OSS sound path without pulling in ALSA, PulseAudio, SDL, or other external library code.
2. Reconstructed the Linux sound lifecycle around the retained OSS backend so audio_fd starts at -1, failed init paths route through shared cleanup, and SNDDMA_Shutdown() now unmaps the mmap DMA buffer and closes the sound descriptor instead of leaving restart/error cleanup implicit.
3. Kept the result deliberately bounded: the default Linux sound backend is still the legacy /dev/dsp OSS path, the silent sink is not an audible backend, and RW-G02 remains open until the broader Linux client renderer, audio, input, and validation boundary is settled.
4. Expanded the focused non-Windows portability suite and refreshed the platform/gap notes so the new silent sink, DMA-position simulation, and OSS shutdown cleanup stay source-pinned as compatibility work rather than strict-retail Windows parity.

Task A4k: Bound the retained Linux joystick device and restart lifecycle [COMPLETED]

Priority: Medium Primary areas: src/code/unix/linux_joystick.c, src/code/unix/linux_glimp.c, src/code/unix/linux_local.h, tests/test_non_windows_portability.py, docs/reverse-engineering/source-file-gap-notes/rw-g02-linux-joystick.md, docs/reverse-engineering/source-file-gap-notes/rw-g02-linux-glimp.md, docs/platform/toolchain-matrix.md Parity estimate: before 98% -> after 98%

Completed work:

1. Reworked the retained Linux joystick lane so startup now scans modern /dev/input/js0 through /dev/input/js3 before the historical /dev/js0 through /dev/js3 nodes, while still avoiding SDL or any other external input library.
2. Added explicit joystick shutdown/restart handling: IN_ShutdownJoystick() releases queued joystick key state, clears tracked axes/buttons, closes joy_fd, and resets ui_joyavail; linux_glimp.c now calls that path on input shutdown and when latched in_joystick changes trigger a restart.
3. Bounded event translation so Linux button events cannot exceed the K_JOY1 through K_JOY32 range and axis events cannot exceed the retained eight-axis / sixteen-direction-key table.
4. Kept RW-G02 open because this is still retained Linux joystick compatibility, not a validated modern Linux client input stack, and pinned that distinction in the focused portability suite and gap notes.

Task A4l: Bound the retained Linux mouse/input shutdown lifecycle [COMPLETED]

Priority: Medium Primary areas: src/code/unix/linux_glimp.c, tests/test_non_windows_portability.py, docs/reverse-engineering/source-file-gap-notes/rw-g02-linux-glimp.md, docs/reverse-engineering/function-parity-gap-audit-2026-04-24.md, docs/platform/toolchain-matrix.md, docs/build/linux-glibc-32bit.md Parity estimate: before 98% -> after 98%

Completed work:

1. Reconstructed the Linux input shutdown path around the retained X11 mouse grab lifecycle so IN_Shutdown() now calls IN_DeactivateMouse() before clearing mouse_avail, avoiding stale grab/activity state across in_restart and normal input teardown.
2. Kept joystick teardown in the same host-owned shutdown path and now clears mouse_active explicitly after the deactivate call, so a later IN_Init() can reactivate the mouse according to the current cvar state instead of inheriting a stale active flag.
3. Added focused source assertions and refreshed the RW-G02 notes so this remains bounded Linux-host compatibility work rather than a claim that the legacy X11/GLX/DGA client path is modern or retail-equivalent.

Task A4m: Bound the retained Unix native-module loader roots and outputs [COMPLETED]

Priority: Medium Primary areas: src/code/unix/unix_main.c, tests/test_non_windows_portability.py, docs/reverse-engineering/source-file-gap-notes/rw-g02-unix-main.md, docs/reverse-engineering/function-parity-gap-audit-2026-04-24.md, docs/platform/toolchain-matrix.md, docs/build/linux-glibc-32bit.md Parity estimate: before 98% -> after 98%

Completed work:

1. Reworked the retained Unix Sys_LoadDll() root probe into an explicit bounded search over cwd, fs_homepath, fs_basepath, and fs_cdpath, so archived/native module roots can be reached through the same configured data-root lane used elsewhere in the Unix host.
2. Reset the legacy entryPoint output before probing, matching the safer failure contract used by the recovered Win32 loader shape and preventing a failed Unix load from leaving stale caller state behind.
3. Added focused source assertions and refreshed the RW-G02 notes so this remains Unix host compatibility work, not a claim that the non-Windows native-module runtime is retail-equivalent.

Task A4n: Bound the Unix event-loop packet payload copy [COMPLETED]

Priority: Medium Primary areas: src/code/unix/unix_main.c, tests/test_non_windows_portability.py, docs/reverse-engineering/source-file-gap-notes/rw-g02-unix-main.md, docs/reverse-engineering/function-parity-gap-audit-2026-04-24.md, docs/platform/toolchain-matrix.md, docs/build/linux-glibc-32bit.md Parity estimate: before 98% -> after 98%

Completed work:

1. Reconstructed the retained Unix Sys_GetEvent() packet queue copy so SE_PACKET events preserve only unread bytes after netmsg.readcount, matching the recovered Win32 event-loop packet path and its SOCKS-aware comment.
2. Kept the change deliberately bounded: the current Unix UDP receiver still leaves readcount at zero, but the event loop now honors that field if a future Unix packet transport consumes header bytes before queuing.
3. Added focused source assertions and refreshed the RW-G02 notes so this is recorded as Unix host event-loop compatibility, not proof of full non-Windows runtime parity.

Task A4o: Bound Unix native-module candidate validation [COMPLETED]

Priority: Medium Primary areas: src/code/unix/unix_main.c, tests/test_non_windows_portability.py, docs/reverse-engineering/source-file-gap-notes/rw-g02-unix-main.md, docs/reverse-engineering/function-parity-gap-audit-2026-04-24.md, docs/platform/toolchain-matrix.md, docs/build/linux-glibc-32bit.md Parity estimate: before 98% -> after 98%

Completed work:

1. Reworked the retained Unix Sys_LoadDll() probing loop so each opened native-module candidate is validated before returning a handle: it must expose dllEntry and then satisfy either the Quake Live export-table path, a vmMain export, or the legacy dllEntry()-returned entry point path.
2. Closed incompatible dlopen() handles, reset failed-load outputs, and continued to the next configured root instead of treating the first opened but incompatible shared object as the final loader outcome.
3. Added focused source assertions and refreshed the RW-G02 docs so this stays recorded as bounded Unix host compatibility work aligned with the recovered Win32 loader shape, not proof of full non-Windows runtime parity.

Task A4p: Bound the null sound host with an explicit silent DMA sink [COMPLETED]

Priority: Medium Primary areas: src/code/null/null_snddma.c, tests/test_non_windows_portability.py, docs/reverse-engineering/source-file-gap-notes/rw-g02-null-snddma.md, docs/reverse-engineering/function-parity-gap-audit-2026-04-24.md, docs/platform/toolchain-matrix.md Parity estimate: before 98% -> after 98%

Completed work:

1. Replaced the null sound host's outright SNDDMA_Init() failure with a local silent DMA sink that seeds the shared dma shape, owns a deterministic 22,050 Hz 16-bit stereo buffer, and advances its DMA cursor from Sys_Milliseconds().
2. Added explicit buffer/state cleanup through SNDDMA_Shutdown(), SNDDMA_BeginPainting(), and S_ClearSoundBuffer() while preserving the compatibility-only no-op behavior for submit, activation, sound registration, local sound, and voice samples.
3. Added focused source assertions and refreshed the RW-G02 docs so this remains a bounded null-host compatibility sink, not an audible backend or a broader runtime parity claim.

Task A4q: Make the null renderer host refuse fake GL initialization [COMPLETED]

Priority: Medium Primary areas: src/code/null/null_glimp.c, tests/test_non_windows_portability.py, docs/reverse-engineering/source-file-gap-notes/rw-g02-null-glimp.md, docs/reverse-engineering/function-parity-gap-audit-2026-04-24.md, docs/platform/toolchain-matrix.md Parity estimate: before 98% -> after 98%

Completed work:

1. Replaced the null QGL_Init() success stub with an explicit qfalse result, so the compatibility host no longer claims that a GL backend was loaded when no renderer library or function pointers exist.
2. Made GLimp_Init() fail with a clear fatal message for the null renderer host and routed GLimp_Shutdown() through QGL_Shutdown(), which now clears retained extension pointers and logging state.
3. Added focused source assertions and refreshed the RW-G02 docs so this stays recorded as a bounded renderer-host compatibility boundary, not a real graphics context or swap path.

Task A4r: Route the null key-event pump through explicit no-device input state [COMPLETED]

Priority: Medium Primary areas: src/code/null/null_input.c, tests/test_non_windows_portability.py, docs/reverse-engineering/source-file-gap-notes/rw-g02-null-input.md, docs/reverse-engineering/function-parity-gap-audit-2026-04-24.md, docs/platform/toolchain-matrix.md Parity estimate: before 98% -> after 98%

Completed work:

1. Extended the null input compatibility state so it clears the retained in_joystick->modified latch while continuing to pin ui_joyavail to 0.
2. Routed Sys_SendKeyEvents() through that no-device refresh once null input is initialized, making the key-event pump an explicit disabled-input maintenance path instead of an unstructured empty placeholder.
3. Added focused source assertions and refreshed the RW-G02 docs so this stays recorded as a null-host input boundary, not a real input backend or event source.

Task A4s: Bound the Linux GLX shutdown and end-frame lifecycle [COMPLETED]

Priority: Medium Primary areas: src/code/unix/linux_glimp.c, tests/test_non_windows_portability.py, docs/reverse-engineering/source-file-gap-notes/rw-g02-linux-glimp.md, docs/reverse-engineering/function-parity-gap-audit-2026-04-24.md, docs/platform/toolchain-matrix.md Parity estimate: before 98% -> after 98%

Completed work:

1. Reworked GLimp_Shutdown() so partial Linux GLX init state is cleaned up instead of returning early when ctx is missing: the path now deactivates mouse state, detaches any current context, destroys retained context/window state only when present, restores VidMode/gamma state, closes the QGL log, clears the GLX globals, and releases QGL before clearing renderer state.
2. Guarded GLimp_EndFrame() against missing display/window/swap state so a failed partial init or post-shutdown call cannot fall through into stale GLX function-pointer state.
3. Added focused source assertions and refreshed the RW-G02 docs so this stays recorded as Linux renderer-host lifecycle containment, not a modern GLX or SDL-style Linux client parity claim.

Active tasks

Task A4: Modernise or explicitly contain the non-Windows portability lanes [COMPLETED 2026-06-05]

Priority: High Primary areas: src/code/unix/, src/code/null/, docs/platform/toolchain-matrix.md, docs/build/linux-glibc-32bit.md, docs/reverse-engineering/non-windows-portability-boundary-2026-06-05.md Parity estimate: before 98% -> after 99%

Scope:

1. Replace the remaining Unix Sys_* helper placeholders that still matter after the restored low-memory, Linux/glibc function compare/checksum, q3monkeyid marker-probe, bounded gprof-control paths, and bounded clipboard retrieval path, bounded data-root probe, and Linux silent sound sink / OSS shutdown cleanup plus bounded Linux joystick descriptor/restart handling plus Linux mouse-grab shutdown cleanup plus Linux GLX partial-init shutdown/end-frame containment and the bounded Unix native-module root probe plus candidate validation, Unix packet event copy, the null renderer GL-init refusal, the explicit null silent DMA sink, and the null no-device key pump, or keep them clearly classified as compatibility-only. The null host/client/audio shims now match the current contract surface closely enough that the remaining work is primarily about the real Unix host boundary rather than stale null host/client/audio/input entry points.
2. Decide whether Linux client/renderer/audio/input support is a real target or whether the current server-only path should remain the bounded endpoint.
3. Refresh the toolchain/runtime guidance and validation coverage once the intended portability boundary is settled.

Completed work:

1. Re-baselined the non-Windows portability lanes around the current target: hosted POSIX source builds and dedicated/server-module validation are the active support endpoints.
2. Chose explicit containment for the legacy Linux client/runtime stack: the retained X11/GLX/DGA renderer/input path, OSS-first audio host, silent sink, and null host/client/audio/input/renderer shims are compatibility-only carries, not active modern Linux client parity targets.
3. Added docs/reverse-engineering/non-windows-portability-boundary-2026-06-05.md and refreshed the RW-G02 gap notes, repo-wide audit, function-gap summary, toolchain matrix, Linux glibc preset note, and focused portability tests so this boundary is enforced as a documented closure.
4. Future Linux client/runtime work now needs a successor task that explicitly adopts a modern renderer/audio/input dependency stack and reproducible validation.

Task A6: Refresh build/runtime evidence for the repo-wide ledger [COMPLETED 2026-06-05]

Priority: Medium Primary areas: tracked runtime probe bundles, native build helpers, platform docs Parity estimate: before 99% -> after 99%

Scope:

1. Refresh the archived runtime evidence bundles on a current toolchain instead of relying solely on the April 2026 tracked artifacts, and promote the stable latest aliases only when reruns remain sufficient.
2. Re-run native build validation where the required Windows toolchain is available and fold the results back into the repo-wide audit.
3. Capture the remaining glibc and continuous/self-hosted publication follow-ups called out in docs/platform/toolchain-matrix.md, including the remaining retail-module runtime freshness tail.

Completed work:

1. Added docs/reverse-engineering/runtime-build-evidence-refresh-2026-06-05.json and docs/reverse-engineering/runtime-build-evidence-refresh-2026-06-05.md as the A6 evidence ledger.
2. Confirmed the retail-module latest alias now points at artifacts/module_validation/logs/retail_module_runtime_evidence_20260602.json, with no warnings, no missing markers, retail ui/qagame/cgame loads, Server: bloodrun, Going from CS_PRIMED to CS_ACTIVE, screenshots, and vm-trace create/free/call traffic.
3. Verified Visual Studio 2022 v143 availability and ran the modern native validation lane with optional codecs disabled. MSBuild produced a clean Release|x86 build with 0 warnings and 0 errors; the wrapper timed out after the successful build while continuing post-build checks, so tools/ci/assert-dll-exports.ps1 was rerun directly.
4. Tightened tools/ci/assert-dll-exports.ps1 so it prefers the newest Release artifact before inspecting exports. The direct export audit now validates the fresh build/win32/Release/modules/qagamex86/qagamex86.dll and build/win32/Release/modules/cgamex86/cgamex86.dll outputs rather than older bin/baseq3 copies.
5. Recorded that strict retail VC10 staged-runtime evidence is still blocked on this checkout because tools/ci/audit-retail-toolchain.ps1 -Strict expects PlatformToolset v100, while the checked-in project files currently use v141. This is now a documented validation boundary, not an uninspected A6 freshness gap.

Earlier 2026-06-05 preflight, now superseded by the A6 refresh ledger:

1. docs/reverse-engineering/windows-native-validation-preflight-2026-06-05.md now records the local native Windows validation preflight. The wrapper-level repo-root inference issue was fixed in tools/ci/audit-retail-toolchain.ps1 and tools/ci/validate-windows-native.ps1, so documented -File invocations now reach the project metadata checks.
2. powershell -NoProfile -ExecutionPolicy Bypass -File .\tools\ci\audit-retail-toolchain.ps1 -Strict stops before build because the strict retail audit expects PlatformToolset=v100 while src/code/quakelive_steam.vcxproj currently reports v141.
3. powershell -NoProfile -ExecutionPolicy Bypass -File .\tools\ci\validate-windows-native.ps1 -RuntimeProfile retail verifies the current v141 defaults for qagamex86, cgamex86, uix86, quakelive_steam, and awesomium_process, then hits the same strict v100 audit blocker. The later A6 refresh above adds modern-host build evidence and current retail-module runtime evidence while keeping this strict VC10 blocker documented.

Task 23: Ownerdraw/stat payload completion and validation [COMPLETED]

Priority: High Primary areas: src/code/game/, src/code/cgame/, runtime validation harnesses Parity estimate: before 99% -> after 100%

Completed work:

1. Revalidated the ownerdraw debug payload against the current runtime capture path and confirmed that the live pickupAvg slab is emitted as fixed-point seconds rather than the older integer-only fixture assumption.
2. Updated the ownerdraw runtime harness in tests/test_ownerdraw_stats_logging.py so the debug-ownerdraw assertion path now accepts and normalizes float pickupAvg CSV payloads while keeping the integer stat families strict.
3. Re-ran the focused ownerdraw/stat validation surface on the current worktree; no concrete unsupported ownerdraw or PLAYER_STATS payload field remains in the active repo-level gap register.

Task 24: Gameplay validation sweep [COMPLETED]

Priority: High Primary areas: physics, Race, gametype-specific rules Parity estimate: before 99% -> after 100%

Completed work:

1. Expanded the gametype lifecycle harness so it now builds through the shared compiler helper on Windows as well as POSIX, and it directly validates duel warmup/configstring sequencing, Race lifecycle routing, and CA/RR round-warmup init dispatch.
2. Tightened the ready-up and match-state regression guards around the retail gametype truth tables, including the both-teams-present gate and the Attack & Defend inclusion in the round-controller team-count publisher set.
3. Reconfirmed the focused Race and shared pmove regression lanes on the current worktree; the gameplay validation sweep is now covered by dedicated fixtures rather than left as an open catch-all validation reminder.

Task 25: UI import contract retail remap guard [COMPLETED]

Priority: High Primary areas: tests/test_ui_menu_files.py, src/code/ui/ui_public.h, src/code/ui/ui_syscalls.c Parity estimate: before 98% -> after 100% for the focused uiImport_t legacy-to-retail-native import contract.

Completed work:

1. Rechecked the committed uix86.dll native import evidence and preserved the current runtime split: uiImport_t remains the legacy source/QVM syscall ABI, while uiQlImport_t carries the recovered retail native slot numbers.
2. Added a parser-based parity guard that pins every uiImport_t value, verifies the complete UI_MapNativeImport handoff into recovered retail native slots, and keeps direct-only retail native slots out of the legacy syscall surface.
3. Re-ran the full UI static regression lane so future drift in either the enum or remap table fails before it can affect retail-module loading.

Working priority order

1. Keep the online-service compatibility boundary explicitly documented if a future open replacement path is introduced.

Reference audits for closed surfaces

• docs/reverse-engineering/non-windows-portability-boundary-2026-06-05.md
• docs/reverse-engineering/runtime-build-evidence-refresh-2026-06-05.md
• docs/reverse-engineering/repo-wide-parity-audit-2026-04-21.md
• docs/reverse-engineering/source-file-parity-ledger-2026-04-22.md
• docs/reverse-engineering/source-file-parity-audit-plan-2026-04-22.md
• docs/reverse-engineering/historical-audit-index-2026-04-22.md
• docs/reverse-engineering/engine-full-parity-audit-and-implementation-plan-2026-04-10.md
• docs/reverse-engineering/game-module-parity-audit-and-implementation-plan-2026-04-10.md
• docs/reverse-engineering/ui-full-parity-audit-and-implementation-plan-2026-04-05.md
• docs/reverse-engineering/qcommon-full-parity-audit-and-implementation-plan-2026-04-10.md
• docs/reverse-engineering/qshared-retail-helper-parity-audit-2026-04-17.md
• docs/reverse-engineering/renderer-full-parity-audit-and-implementation-plan-2026-04-09.md
• docs/reverse-engineering/server-full-parity-audit-and-implementation-plan-2026-04-10.md
• docs/reverse-engineering/engine-host-support-full-parity-audit-and-implementation-plan-2026-04-10.md
• docs/reverse-engineering/platform-specific-engine-parity-audit-and-implementation-plan-2026-04-16.md
• docs/reverse-engineering/engine-netcode-parity-audit-2026-04-16.md
• docs/reverse-engineering/awesomium-browser-host-parity-audit-and-implementation-plan-2026-04-16.md

Historical closure anchors kept for parity-gate compatibility

The entries below are intentionally terse. They preserve the exact historical task headings and parity-estimate lines that checked-in parity gates still read from this file, without keeping pages of duplicated completed-task prose in the active plan.

Task 34: Cgame Attack and Defend round-scoreboard owner parity closure [COMPLETED]

Task 31: Strict retail game-module parity closure [COMPLETED]

Task 73: Renderer internal helper-family ownership closure tranche [COMPLETED]

Parity estimate: before 93% -> after 96% (RG-P5 complete; RG-G06 closed)

Task 75: Renderer strict retail-font-stack re-audit and closure-plan refresh [COMPLETED]

Task 87: Client CL-P6 parity gate and runtime-evidence closure [COMPLETED]

Parity estimate: before 99% -> after 100%

Task 92: Qcommon QC-P2 parity gate and Windows-friendly harness closure [COMPLETED]

Task 94: Qcommon QC-P3 retail homepath closure [COMPLETED]

Task 97: Qcommon QC-P4 collision leaf ownership closure [COMPLETED]

Task 100: Server SV-P7 parity gate and dedicated runtime-evidence closure [COMPLETED]

Parity estimate: before 97% -> after 100%

Task 102: Qcommon QC-P5 fallback VM closure [COMPLETED]

Task 103: Remaining engine host/support EH-P2 Win32 Unicode clipboard closure [COMPLETED]

Task 104: Strict retail game-module final ledger and runtime-evidence reconciliation [COMPLETED]

Task 104: Qcommon QC-P6 runtime-evidence and ledger closure [COMPLETED]

Parity estimate: before 98% -> after 100%

Task 110: Qshared shared-helper exactness and validation-lane closure [COMPLETED]

Parity estimate: before 99% -> after 100%

Task 105: Remaining engine host/support EH-P3 raw-input host closure [COMPLETED]

Task 111: Client/input m_cpi and general mouse parity closure [COMPLETED]

Parity estimate: before 96% -> after 100% (m_cpi math was already retail-mapped; this closes the remaining Win32 mouse host/key-number gaps)

Task 112: Win32 game-cursor mouse ownership closure [COMPLETED]

Parity estimate: before 99% -> after 100% (browser/UI/cgame absolute-coordinate lanes now preserve the retail cursor-owner split; UI/cgame game cursors suppress the OS cursor while browser cursor overrides remain browser-owned)

Task 106: Remaining engine host/support EH-P6 parity gate and evidence closure [COMPLETED]

Parity estimate: before 89% -> after 92%

Task 107: Remaining engine host/support EH-P4 botlib internal proof closure [COMPLETED]

Parity estimate: before 92% -> after 95%

Task 109: Remaining engine host/support EH-P1 boundary formalisation closure [COMPLETED]

Parity estimate: before 100% -> after 100% (EH-P1 complete; strict-retail scope classification formalised)

Task 113: Clientinfo compact identity/color transport closure [COMPLETED]

Parity estimate: before 94% -> after 98% for the scoped CS_PLAYERS clientinfo color, SteamID, and country transport lane

Task 114: Clientinfo identity-tail and native sidecar reconstruction [COMPLETED]

Parity estimate: before 82% -> after 90% for the scoped cgame clientinfo identity/speaking/avatar tail. This pass confirmed the retail 0x738 clientinfo stride, documented the 0x710..0x730 privilege, queue, speaking, identity, and avatar tail, and reconstructed source wiring for cn, xcn, rp, p, cached clean names, avatar handles, and in-record speaking state. This pass left the binary-exact cgame animation/clientinfo layout split open: retail uses 37 animation records of size 0x18, which is closed by the follow-up Task 115 animation-record pass below.

Task 115: Clientinfo cgame animation record and retail stride closure [COMPLETED]

Parity estimate: before 90% -> after 94% for the scoped cgame clientinfo layout/animation lane. This pass confirmed the retail animation cache at clientinfo offset 0x318 with 37 records of size 0x18, custom sounds at 0x690, and the retail outer clientinfo stride of 0x738. Source cgame now uses a local cgAnimation_t record without the shared GPL flipflop field, cgame animation parsing/playback consumes that type, and clientInfo_t carries an explicit temporary pad so the x86 source stride matches retail. The shared game/UI animation_t remains untouched. The remaining gap is the internal member-order repack to move the animation, sound, identity, avatar, and voice fields to their retail byte offsets naturally, which is closed by Task 116.

Task 116: Clientinfo internal anchor-offset repack [COMPLETED]

Parity estimate: before 94% -> after 96% for the scoped cgame clientinfo byte-layout lane. This pass moved the recovered native sidecar fields behind sounds[MAX_CUSTOM_SOUNDS] and aligned the strongest retail anchors in the x86 source layout: fixed animation metadata at 0x2E0, model handles at 0x2FC, animations at 0x318, custom sounds at 0x690, privilege/spectator queue at 0x710..0x718, speaking state at 0x71C..0x720, Steam identity at 0x728..0x72C, avatar cache at 0x730, and final size 0x738. The remaining gap is the early 0x008..0x2DF mixed retail/source-compatibility region, including model/head-model scratch strings, color override cache, country flag cache, and scoreboard/team-overlay mirrors.

Task 117: Clientinfo early-slot retail offset map [COMPLETED]

Parity estimate: before 96% -> after 97% for the scoped cgame clientinfo byte-layout lane. This pass promoted the early 0x000..0x2FC retail map into named CG_CLIENTINFO_RETAIL_OFFSET_* constants and documented the evidence chain from CG_NewClientInfo, CG_CopyClientIdentity, placement/profile text, country flag lookup, and head/model scale consumers. High-confidence early anchors now include display name 0x008, clean/native identity name 0x048, country 0x0C8, team 0x108, bot skill 0x10C..0x110, colors 0x114..0x128, handicap 0x130, wins/losses 0x134..0x138, fixed animation metadata 0x2E0..0x2F8, and model handles 0x2FC..0x314. The source struct is not forcibly repacked around this early map yet because the current tree keeps source-compatibility color override, country flag, and scoreboard/team-overlay fields in that region; moving them should be a deliberate sidecar migration once the remaining medium-confidence slots are closed.

Task 118: Clientinfo model-string band and pre-animation scalar reconstruction [COMPLETED]

Parity estimate: before 97% -> after 98% for the scoped cgame clientinfo byte-layout lane. This pass promoted the six 64-byte retail model string slots from a medium-confidence band to named high-confidence offsets: body model/skin at 0x154..0x1D3, icon body model/skin at 0x1D4..0x253, and head model/skin at 0x254..0x2D3. The adjacent pre-animation fields are now pinned as deferred flag 0x2D4, model bounding-box scale 0x2D8, and new-animation flag 0x2DC. Source reconstruction now separates clientInfo_t.modelScale from the animation.cfg headOffset vector, moves newAnims onto the retail 0x2DC lane, and keeps focused regression coverage over the Ghidra/HLIL constructor, load, copy, sound, icon, and tag-flag evidence.

Task 119: Clientinfo score, compact tinfo, and effect-timer lane closure [COMPLETED]

Parity estimate: before 98% -> after 99% for the scoped cgame clientinfo byte-layout lane. This pass promoted the live score mirror at 0x12C, team-task scalar at 0x13C, medkit timer at 0x144, and invulnerability start/stop timers at 0x148..0x14C using HLIL/Ghidra producer and consumer evidence. It also documents the remaining 0x150 breath-puff timer candidate as medium confidence and reconstructs CG_ParseTeamInfo to accept the retail compact tinfo <count> <client...> shape while preserving legacy six-field source-server rows for teammate location, health, armor, weapon, and powerup compatibility.

Task 120: Clientinfo animation metadata and render-handle slot closure [COMPLETED]

Parity estimate: before 99% -> after 99.3% for the scoped cgame clientinfo byte-layout lane. This pass split the broad 0x2E0..0x314 animation metadata/model-handle bands into individually named retail offsets: fixedlegs, fixedtorso, headOffset[3], footsteps, gender, legs/torso/ head model and skin handles, and modelIcon. The evidence now ties animation cfg parsing, model/skin/icon registration, player/model-preview rendering, and the retail deferred-asset copier together. It also preserves the retail detail that CG_CopyClientInfoModel copies from 0x2E8 onward and intentionally does not propagate fixedlegs or fixedtorso during deferred asset reuse.

Task 121: Steam callback ABI launch/runtime reconstruction [COMPLETED]

Parity estimate: before 74% -> after 93% for the scoped Steam callback ABI lane. This pass aligned the source CCallbackBase-compatible vtable order to the Steam runtime dispatcher shape (Run, RunCallResult, GetCallbackSizeBytes), preserved the retail split where SteamUGCQueryCompleted_t is a call-result rather than a normal registered callback, and recorded the evidence in docs/reverse-engineering/quakelive_steam_mapping_round_429.md.

Task 122: Steam initialization app-id ownership cleanup [COMPLETED]

Parity estimate: before 82% -> after 91% for the scoped Steam initialization/app-id lane. This pass removed the init-time diagnostic QL_Steamworks_GetAppID() side effect so app-id reads remain owned by the retail-mapped wrapper callers, especially the stats_clear gate, and recorded the evidence in docs/reverse-engineering/quakelive_steam_mapping_round_430.md.

Task 123: SteamClient_Init com_build launch guard reconstruction [COMPLETED]

Parity estimate: before 63% -> after 90% for the scoped Steam com_build launch-guard lane. This pass added the source-side build-script gate at the top of SteamClient_Init, before platform-service refresh, callback/lobby bootstrap, voice command registration, stats-clear, and rich-presence writes, and recorded the evidence in docs/reverse-engineering/quakelive_steam_mapping_round_431.md.

Task 124: SteamClient_Frame retail pump-order reconstruction [COMPLETED]

Parity estimate: before 86% -> after 94% for the scoped SteamClient_Frame runtime-pump lane. This pass moved the source-only callback recovery shim behind the retail-equivalent frame pump so initialized frames run SteamAPI_RunCallbacks, outgoing voice send, channel-0 stats-report packet drain, and incoming voice packet processing in the order shown by the retail HLIL, and recorded the evidence in docs/reverse-engineering/quakelive_steam_mapping_round_432.md.

Task 125: Steam auth-ticket common error cleanup reconstruction [COMPLETED]

Parity estimate: before 78% -> after 92% for the scoped Steam auth-ticket error-cleanup lane. This pass added the retail-observed SteamClient_CancelAuthTicket() call to Com_Error after publishing com_errorMessage and before the error-code branches diverge, keeping retained ticket cleanup consistent across disconnect, drop, need-CD, and fatal paths. The evidence is recorded in docs/reverse-engineering/quakelive_steam_mapping_round_433.md.

Task 126: Pre-filesystem Steam com_build guard reconstruction [COMPLETED]

Parity estimate: before 70% -> after 91% for the scoped pre-filesystem Steam startup guard lane. This pass added the retail-observed com_build skip to Com_InitSteamClientForFilesystem() before its platform-service refresh, so build-script launches do not probe Steam in the source-only early filesystem SteamID path. The evidence is recorded in docs/reverse-engineering/quakelive_steam_mapping_round_434.md.

Task 127: Lazy Steam service refresh guard reconstruction [COMPLETED]

Parity estimate: before 67% -> after 93% for the scoped lazy Steam refresh lane. This pass added SteamClient_ShouldRefreshPlatformServices() and applied it before source-only lazy QL_RefreshPlatformServices() calls in the SteamID, auth-ticket, subscription, UGC download-info, IP-country, callback-recovery, and frame-pump paths, keeping com_build launches aligned with retail retained initialized-flag semantics. The evidence is recorded in docs/reverse-engineering/quakelive_steam_mapping_round_435.md.

Task 128: Steam frame-pump initialized-state owner reconstruction [COMPLETED]

Parity estimate: before 78% -> after 96% for the scoped Steam frame-pump owner lane. This pass removed source-only QL_RefreshPlatformServices() and SteamClient_SetInitializedState() calls from SteamClient_Frame() and kept callback recovery from re-latching Steam state, matching the retail data_e30218 != 0 gate before SteamAPI_RunCallbacks(), voice send, stats report packet drain, and voice packet drain. The evidence is recorded in docs/reverse-engineering/quakelive_steam_mapping_round_436.md.

Task 129: Steam identity bootstrap retained-state reconstruction [COMPLETED]

Parity estimate: before 74% -> after 95% for the scoped client identity bootstrap owner lane. This pass removed direct QL_Steamworks_Init() calls from SteamClient_SyncPersonaNameCvar() and CL_Steam_SeedCountryCvar(), gated both helpers on the retained SteamClient_IsInitialized() state after the retail com_build skip, and kept the retail name=anon persona fallback when Steam identity is unavailable. The evidence is recorded in docs/reverse-engineering/quakelive_steam_mapping_round_437.md.

Task 130: Steam wrapper retained-state owner reconstruction [COMPLETED]

Parity estimate: before 72% -> after 94% for the scoped small Steam wrapper owner lane. This pass added SteamClient_InitForFilesystem() as the explicit source-only early filesystem Steam owner, moved the pre-filesystem platform refresh into that client-owned helper, and removed wrapper-level QL_RefreshPlatformServices() / SteamClient_SetInitializedState() calls from the SteamID, auth-ticket, subscription, UGC download-info, and country wrapper thunks. The evidence is recorded in docs/reverse-engineering/quakelive_steam_mapping_round_438.md.

Task 131: Steam client shutdown resource-callback owner reconstruction [COMPLETED]

Parity estimate: before 78% -> after 95% for the scoped Steam client shutdown/resource callback owner lane. This pass rechecked the retail SteamAPI_Shutdown thunk at 0x00460540, the SteamDataSource_Shutdown callback unregister path at 0x00464440, and the source platform avatar callback owner, then moved CL_ShutdownSteamResources() into the public SteamAPI_Shutdown() wrapper ahead of CL_Steam_ShutdownCallbacks() and QL_Steamworks_Shutdown(). The evidence is recorded in docs/reverse-engineering/quakelive_steam_mapping_round_440.md.

Task 132: Steam game-start rich-presence initialized-gate reconstruction [COMPLETED]

Parity estimate: before 82% -> after 96% for the scoped game.start / rich-presence runtime lane. This pass rechecked retail sub_4F38F0, found the demo playback early return before "game.start" publication, and matched the lanIp and "Playing a match" rich-presence writes to the retained SteamClient_IsInitialized() gate. The source now suppresses the packed browser event during demo playback and only writes adjacent Steam presence state when the retained Steam client initialization flag is live. The evidence is recorded in docs/reverse-engineering/quakelive_steam_mapping_round_442.md.

Task 133: SteamClient_Init post-success side-effect boundary reconstruction [COMPLETED]

Parity estimate: before 84% -> after 97% for the scoped SteamClient_Init post-success bootstrap lane. This pass rechecked retail sub_461500, confirmed the failed SteamAPI_Init() return before callback, lobby, voice, stats, and main-menu rich-presence side effects, and moved the source fallback branch to return immediately when the retained SteamClient_IsInitialized() state is false. The evidence is recorded in docs/reverse-engineering/quakelive_steam_mapping_round_443.md.

Task 134: SteamClient_Init terminal success diagnostic reconstruction [COMPLETED]

Parity estimate: before 90% -> after 99% for the scoped SteamClient_Init terminal success diagnostic lane. This pass rechecked retail sub_461500 at 0x004615CA, confirmed the "Steam API initialized.\n" log after the initial "At the main menu" rich-presence write, and added the matching client-owner success marker after CL_Steam_SetMainMenuRichPresence() in the source bootstrap path. The evidence is recorded in docs/reverse-engineering/quakelive_steam_mapping_round_445.md.

Task 135: Steam connect_lobby command and lobby bootstrap return reconstruction [COMPLETED]

Parity estimate: before 76% -> after 99% for the scoped connect_lobby command handler lane, and before 88% -> after 97% for the scoped SteamLobby_Init cvar/command/return boundary. This pass rechecked retail sub_464AA0 and the SteamLobby_Init registration at 0x00465840, removed source-only argc and provider guards from CL_Steam_ConnectLobby_f(), stopped using the initial lobby callback registration result as a main Steam bootstrap gate, and pinned the lobby cvar flags / command registration against retail HLIL. The evidence is recorded in docs/reverse-engineering/quakelive_steam_mapping_round_447.md.

Task 136: Steam lobby chat-message entry-type gate reconstruction [COMPLETED]

Parity estimate: before 82% -> after 99% for the scoped lobby chat-message publish gate. This pass rechecked retail sub_4645A0, confirmed that lobby.%s.chat is published only when Steam's lobby chat entry type is 1, added the matching source gate in CL_Steam_Lobby_OnLobbyChatMessage(), and pinned the HLIL branch / publish instruction in the Steam platform tests. The evidence is recorded in docs/reverse-engineering/quakelive_steam_mapping_round_448.md.

Task 137: Steam browser lobby method initialized-gate reconstruction [COMPLETED]

Parity estimate: before 78% -> after 97% for the scoped browser-facing Steam lobby method initialized-state gate. This pass rechecked retail sub_4649B0, sub_4649E0, sub_465630, sub_464AC0, sub_464B10, and sub_464BB0, confirmed each route begins at the retained sub_460510 Steam-client initialized gate before SteamMatchmaking/SteamFriends calls, and added the matching source gates to the lobby/social browser wrappers. The evidence is recorded in docs/reverse-engineering/quakelive_steam_mapping_round_450.md.

Task 138: Steam browser non-lobby method initialized-boundary reconstruction [COMPLETED]

Parity estimate: before 74% -> after 94% for the scoped browser-facing non-lobby Steam method initialized boundary. This pass rechecked retail sub_431E50 qz_instance dispatch cases for OpenSteamOverlayURL, RequestUserStats, ActivateGameOverlayToUser, and GetAllUGC, plus mapped sub_460DC0 (SteamWorkshop_GetAllUGC), then added the retained source initialization gate before the source wrappers reach SteamFriends, SteamUserStats, or SteamUGC. The evidence is recorded in docs/reverse-engineering/quakelive_steam_mapping_round_451.md.

Task 139: Steam browser native server-list retained-state reconstruction [COMPLETED]

Parity estimate: before 80% -> after 95% for the scoped native SteamMatchmakingServers list/detail availability boundary. This pass rechecked retail sub_462E80, sub_462EB0, sub_463090, and sub_4630B0, confirmed the retained JSBrowser owner directly uses SteamMatchmakingServers refresh, cancel, gamedir=baseq3 filtered-list, and detail-request paths, and added the retained SteamClient_IsInitialized() gate to the source native-list availability guard before native list/detail requests can start. The evidence is recorded in docs/reverse-engineering/quakelive_steam_mapping_round_452.md.

Task 140: Steam browser favorite-server retained-state reconstruction [COMPLETED]

Parity estimate: before 82% -> after 96% for the scoped browser SetFavoriteServer retained-state boundary. This pass rechecked retail sub_431E50 qz_instance dispatch case 0x13, confirmed the browser dispatcher uses SteamUtils and SteamMatchmaking directly for add/remove favorite-game slots after the incoming argument conversion, and added the retained SteamClient_IsInitialized() gate before the source web-host favorite wrapper can reach QL_Steamworks_SetFavoriteServerForApp. The evidence is recorded in docs/reverse-engineering/quakelive_steam_mapping_round_453.md.

Task 141: Steam GameServer frame retained-state reconstruction [COMPLETED]

Parity estimate: before 84% -> after 97% for the scoped Steam GameServer packet/frame retained-state boundary. This pass rechecked retail sub_465D50 and sub_466850, confirmed both native packet ingress and the server callback / P2P maintenance loop are gated by the retained data_e30358 GameServer initialized flag before reaching SteamGameServer, SteamGameServerNetworking, or SteamGameServer_RunCallbacks, and added the matching source owner guards in SV_SteamServerHandleIncomingPacket and SV_SteamServerNetworkingFrame. The evidence is recorded in docs/reverse-engineering/quakelive_steam_mapping_round_454.md.

Task 142: Steam GameServer key-value retained-state reconstruction [COMPLETED]

Parity estimate: before 86% -> after 98% for the scoped Steam GameServer serverinfo key-value publication boundary. This pass rechecked retail sub_465A60, confirmed the info-string walker checks the retained data_e30358 GameServer initialized flag before splitting serverinfo pairs or calling SteamGameServer vtable slot 0x50, and added the matching source guard to QL_Steamworks_ServerSetKeyValuesFromInfoString. The evidence is recorded in docs/reverse-engineering/quakelive_steam_mapping_round_455.md.

Task 143: Steam GameServer heartbeat retained-state reconstruction [COMPLETED]

Parity estimate: before 88% -> after 98% for the scoped Steam GameServer heartbeat publication boundary. This pass rechecked retail sub_465DB0, confirmed it checks the retained data_e30358 GameServer initialized flag before calling SteamGameServer vtable slot 0x9C, and added the matching source guard to QL_Steamworks_ServerEnableHeartbeats before it reaches QL_Steamworks_GetGameServer. The evidence is recorded in docs/reverse-engineering/quakelive_steam_mapping_round_456.md.

Task 144: Steam GameServer unauthenticated client identity reconstruction [COMPLETED]

Parity estimate: before 45% -> after 96% for the scoped unauthenticated server-client identity wrapper. This pass rechecked retail sub_465DF0 and its SV_BotAllocateClient call site, corrected the alias away from public server identity publication, reconstructed the retained data_e30358-gated SteamGameServer vtable slot 0x64 wrapper as QL_Steamworks_ServerCreateUnauthenticatedUserConnection, and wired bot/local server-owned client allocation through that source bridge. The evidence is recorded in docs/reverse-engineering/quakelive_steam_mapping_round_458.md.

Task 145: Steam GameServer connected serverinfo replay reconstruction [COMPLETED]

Parity estimate: before 84% -> after 97% for the scoped connected-success Steam GameServer publication sequence. This pass rechecked retail sub_466800, confirmed the connect callback marks the Steam server connected flag, publishes identity, pushes a full state refresh, runs the dedicated/server callback continuation, and finally replays the serverinfo key/value batch through sub_465A60. The source callback now mirrors that reconnect path by republishing CVAR_SERVERINFO with QL_Steamworks_ServerSetKeyValuesFromInfoString after the identity/state/stats refresh. The evidence is recorded in docs/reverse-engineering/quakelive_steam_mapping_round_462.md.

Task 146: Steam GameServer negative callback log reconstruction [COMPLETED]

Parity estimate: before 72% -> after 98% for the scoped connect-failure/disconnect callback lane. This pass rechecked retail sub_465C10 and sub_465C30, confirmed both callbacks clear the retained Steam server connected flag and emit fixed log messages without reading the Steam callback payload/result fields, and updated the source callbacks to ignore the adapter payload while preserving the wrapper signatures. The evidence is recorded in docs/reverse-engineering/quakelive_steam_mapping_round_464.md.