dsh-memory-setup
August 18, 2026 ยท View on GitHub
Solve the AI goldfish brain ๐ โ a local, auditable personal memory layer for DeepSeek Harness. Remembers your preferences, project conventions, workflows, and error lessons, and injects them back into every session.
่งฃๅณ AI ็"้้ฑผ่"๏ผๆฌๅฐใๅฏๅฎก่ฎก็ไธชไบบ่ฎฐๅฟๅฑโโๅๅฅฝใ้กน็ฎ็บฆๅฎใๅทฅไฝๆนๅผใ็บ ้ๆ่ฎญ๏ผไผ่ฏ้ด่ชๅจ็ปงๆฟใ
Install
dsh plugin --profile <profile> add dsh-memory-setup
Tools
| Tool | What it does |
|---|---|
memory_setup | One-time onboarding: language, code style, tools, conventions, workflows |
memory_status | Read current memory + changelog (also auto-injected guidance at boot) |
memory_update | Update one memory path (e.g. preferences.codeStyle) with a changelog entry |
memory_project | Auto-extract project conventions from workspace files (README / package.json / configs), preview or apply |
memory_lesson | Record an error lesson (error โ fix โ evidence) so the same mistake is not repeated |
memory_review | v0.2 โ formalize an incident into a lesson with root cause; similar lessons are auto-merged (dedupe + hit counter) |
memory_export | v0.2 โ export the full memory + changelog to a Markdown file for review/backup |
knowledge_add | v0.3 โ add a knowledge entry (title/content/tags/source); similar titles auto-merge |
knowledge_search | v0.3 โ keyword retrieval (title ร3 / tags ร2 / content ร1 scoring) |
knowledge_list / knowledge_remove | v0.3 โ browse / delete knowledge entries |
memory_diff | v0.4 โ diff current memory against memory.json.bak, optionally written to memory-diff.md |
memory_review_session | v0.5 โ bulk incident review: submit many failures at once, dedupe per item |
memory_snapshot / memory_list_snapshots / memory_restore | v0.6 โ snapshot the memory (keeps N), list, and restore with auto-backup of the current state |
memory_troubleshoot | v0.6 โ given an error, search past lessons + knowledge base for a known fix |
memory_stats | v0.7 โ aggregate stats across memory, knowledge base and snapshots |
memory_promote | v0.7 โ promote recurring lessons (hits โฅ threshold) into standing conventions; auto-runs on save |
memory_import / memory_merge | v0.8 โ import memory from JSON (auto-migrate) / merge two memories (newer or both) |
kb_export / kb_import | v0.8 โ knowledge base JSON round-trip |
memory_focus | v0.9 โ relevance-based injection: only memory matching a topic is injected |
memory_tier | v1.0 โ hot/warm/cold tiers (hot is injected, cold is archived) |
memory_audit | v1.0 โ sha256 integrity check + changelog audit report |
memory_import_claude | v1.1 โ import conventions from CLAUDE.md |
memory_export_all / memory_import_all | v1.2 โ full backup bundle (memory + KB + snapshots) |
memory_annotate | v1.2 โ owner/purpose annotations on entries |
knowledge_embed | v0.5 โ backfill embeddings for KB entries (needs embeddingEndpoint); enables semantic search |
Storage & auditability
- Location:
<workspace>/.dsh-memory-setup/memory.jsonโ plain JSON, easy to read/back up - Every mutation appends to
changelog(when / what / why) โ memory is auditable by design - Lessons carry an optional
evidencefield (file/command/observation) โ no evidence, no lesson - Local-first: nothing leaves your machine
Config (optional)
| Field | Default | Description |
|---|---|---|
| memoryDir | .dsh-memory-setup | memory dir relative to the session workspace |
| injectOnBoot | true | inject live memory into the system prompt (dynamic context, refreshed on save) |
| maxMemoryChars | 6000 | cap for rendered memory text |
| lessonTtlDays | 90 | lessons expire after this many days (0 disables) |
| changelogCap | 100 | max changelog entries kept |
| backupOnSave | true | write memory.json.bak before every save |
| reviewReminder | true | append self-review reminder to guidance |
| embeddingEndpoint | (empty) | OpenAI-compatible embeddings endpoint (enables semantic KB search) |
| embeddingKey | (empty) | Bearer key for the embeddings endpoint |
| embeddingModel | text-embedding-3-small | embeddings model name |
| snapshotKeep | 10 | max memory snapshots kept |
| troubleshootReminder | true | append troubleshoot/snapshot reminder to guidance |
Roadmap
- v0.2 โ
: incident review with dedupe (
memory_review), lesson/convention expiry + changelog cap (auto-pruned on save),memory.json.bakbackup on every save, Markdown export (memory_export) - v0.3 โ
: personal knowledge base (
knowledge_*, keyword retrieval, title-merge dedupe); dynamic memory injection via a livesystemPrompt.context()section โ refreshed at boot (from the workspace path) and after every memory save (throttled 30s), with static guidance as fallback - v0.4 โ
: BM25 retrieval for the knowledge base (title ร3 / tags ร2 / content ร1, IDF-scaled โ no embeddings, no deps), memory diff export (
memory_diffvs backup), self-review reminder in the injected guidance - v0.5 โ
: optional embeddings provider (OpenAI-compatible endpoint;
knowledge_embedbackfill + cosine retrieval, BM25 fallback), bulk incident review (memory_review_session), own-tool fs failure tracking surfaced into the injected context - v0.6 โ
: memory snapshots & restore (
memory_snapshot/memory_list_snapshots/memory_restore, capped, index-file based), fault troubleshooting (memory_troubleshootโ lessons + knowledge lookup), troubleshoot reminder in guidance - v0.7 โ
: KB included in snapshots (snapshot/restore both memory + knowledge), lesson auto-promotion (recurring lessons with hits โฅ threshold become standing conventions, auto-run on save โ the memory literally learns from repeated mistakes), memory stats (
memory_stats) - v0.8 โ
: schema v2 migration (auto on load), memory import/merge (
memory_import/memory_merge, newer/both conflicts), knowledge base JSON round-trip (kb_export/kb_import) - v0.9 โ
: lesson health evaluation (failing/resolved/active โ auto on save, โ ๏ธ markers in render), relevance-focused injection (
memory_focus) - v1.0 โ
: tiers (hot/warm/cold), integrity audit (
memory_audit, sha256), privacy redaction in exports (sensitiveKeys) - v1.1 โ
: optimistic locking (revision-based CAS โ multi-session/multi-agent safe), CLAUDE.md import (
memory_import_claude) - v1.2 โ
: full backup bundle (
memory_export_all/memory_import_allโ memory + KB + snapshots), annotations (memory_annotate, owner/purpose) โ team-ready - v1.3+: lesson auto-detection (pending a tool-call event API), web stats view
Security
Memory plugins are the highest-trust plugin type โ see SECURITY.md for the audit posture.