EVIDENCE.md
June 12, 2026 · View on GitHub
########## COPILOT CLI --help ##########
Usage: copilot [options] [command]
GitHub Copilot CLI - An AI-powered coding assistant.
Start an interactive session to chat with Copilot, or use -p/--prompt for
non-interactive scripting. Copilot can edit files, run shell commands, search
your codebase, and more — all with configurable permissions.
Run `copilot <command> --help` for details on any subcommand.
Options:
--effort, --reasoning-effort <level> Set the reasoning effort level (choices:
"low", "medium", "high", "xhigh")
--acp Start as Agent Client Protocol server
--add-dir <directory> Add a directory to the allowed list for
file access (can be used multiple times)
--add-github-mcp-tool <tool> Add a tool to enable for the GitHub MCP
server instead of the default CLI subset
(can be used multiple times). Use "*"
for all tools.
--add-github-mcp-toolset <toolset> Add a toolset to enable for the GitHub
MCP server instead of the default CLI
subset (can be used multiple times). Use
"all" for all toolsets.
--additional-mcp-config <json> Additional MCP servers configuration as
JSON string or file path (prefix with @)
(can be used multiple times; augments
config from ~/.copilot/mcp-config.json
for this session)
--agent <agent> Specify a custom agent to use
--allow-all Enable all permissions (equivalent to
--allow-all-tools --allow-all-paths
--allow-all-urls)
--allow-all-paths Disable file path verification and allow
access to any path
--allow-all-tools Allow all tools to run automatically
without confirmation; required for
non-interactive mode (env:
COPILOT_ALLOW_ALL)
--allow-all-urls Allow access to all URLs without
confirmation
--allow-tool[=tools...] Tools the CLI has permission to use;
will not prompt for permission
--allow-url[=urls...] Allow access to specific URLs or domains
--alt-screen[=value] Use the terminal alternate screen buffer
(on|off)
--autopilot Enable autopilot continuation in prompt
mode
--available-tools[=tools...] Only these tools will be available to
the model
--banner Show the startup banner
--bash-env[=value] Enable BASH_ENV support for bash shells
(on|off)
--config-dir <directory> Set the configuration directory
(default: ~/.copilot)
--continue Resume the most recent session
--deny-tool[=tools...] Tools the CLI does not have permission
to use; will not prompt for permission
--deny-url[=urls...] Deny access to specific URLs or domains,
takes precedence over --allow-url
--disable-builtin-mcps Disable all built-in MCP servers
(currently: github-mcp-server)
--disable-mcp-server <server-name> Disable a specific MCP server (can be
used multiple times)
--disallow-temp-dir Prevent automatic access to the system
temporary directory
--enable-all-github-mcp-tools Enable all GitHub MCP server tools
instead of the default CLI subset.
Overrides --add-github-mcp-toolset and
--add-github-mcp-tool options.
--excluded-tools[=tools...] These tools will not be available to the
model
--experimental Enable experimental features
-h, --help display help for command
-i, --interactive <prompt> Start interactive mode and automatically
execute this prompt
--log-dir <directory> Set log file directory (default:
~/.copilot/logs/)
--log-level <level> Set the log level (choices: "none",
"error", "warning", "info", "debug",
"all", "default")
--max-autopilot-continues <count> Maximum number of continuation messages
in autopilot mode (default: unlimited)
--model <model> Set the AI model to use
--mouse[=value] Enable mouse support in alt screen mode
(on|off)
--no-alt-screen Disable the terminal alternate screen
buffer
--no-ask-user Disable the ask_user tool (agent works
autonomously without asking questions)
--no-auto-update Disable downloading CLI update
automatically (disabled by default in CI
environments)
--no-bash-env Disable BASH_ENV support for bash shells
--no-color Disable all color output
--no-custom-instructions Disable loading of custom instructions
from AGENTS.md and related files
--no-experimental Disable experimental features
--no-mouse Disable mouse support in alt screen mode
--output-format <format> Output format: 'text' (default) or
'json' (JSONL, one JSON object per line)
(choices: "text", "json")
-p, --prompt <text> Execute a prompt in non-interactive mode
(exits after completion)
--plain-diff Disable rich diff rendering (syntax
highlighting via diff tool specified by
git config)
--plugin-dir <directory> Load a plugin from a local directory
(can be used multiple times)
--resume[=sessionId] Resume from a previous session
(optionally specify session ID or task
ID)
-s, --silent Output only the agent response (no
stats), useful for scripting with -p
--screen-reader Enable screen reader optimizations
--secret-env-vars[=vars...] Environment variable names whose values
are stripped from shell and MCP server
environments and redacted from output
(e.g.,
--secret-env-vars=MY_KEY,OTHER_KEY)
--share[=path] Share session to markdown file after
completion in non-interactive mode
(default: ./copilot-session-<id>.md)
--share-gist Share session to a secret GitHub gist
after completion in non-interactive mode
--stream <mode> Enable or disable streaming mode
(choices: "on", "off")
-v, --version show version information
--yolo Enable all permissions (equivalent to
--allow-all-tools --allow-all-paths
--allow-all-urls)
Commands:
help [topic] Display help information
init Initialize Copilot instructions
login [options] Authenticate with Copilot
plugin Manage plugins
update Download the latest version
version Display version information
Help Topics:
commands Interactive Mode Commands
config Configuration Settings
environment Environment Variables
logging Logging
permissions Permissions
Examples:
# Start interactive mode
$ copilot
# Start interactive mode and automatically execute a prompt
$ copilot -i "Fix the bug in main.js"
# Execute a prompt in non-interactive mode (exits after completion)
$ copilot -p "Fix the bug in main.js" --allow-all-tools
# Enable all permissions with a single flag
$ copilot -p "Fix the bug in main.js" --allow-all
$ copilot -p "Fix the bug in main.js" --yolo
# Start with a specific model
$ copilot --model gpt-5.2
# Resume the most recent session
$ copilot --continue
# Resume a previous session using session picker
$ copilot --resume
# Resume a specific session by ID
$ copilot --resume=<session-id>
# Start a new session with a specific UUID
$ copilot --resume=0cb916db-26aa-40f2-86b5-1ba81b225fd2
# Resume with auto-approval
$ copilot --allow-all-tools --resume
# Allow access to additional directory
$ copilot --add-dir /home/user/projects
# Allow multiple directories
$ copilot --add-dir ~/workspace --add-dir /tmp
# Disable path verification (allow access to any path)
$ copilot --allow-all-paths
# Allow all git commands except git push
$ copilot --allow-tool='shell(git:*)' --deny-tool='shell(git push)'
# Allow all file editing
$ copilot --allow-tool='write'
# Allow all but one specific tool from MCP server with name "MyMCP"
$ copilot --deny-tool='MyMCP(denied_tool)' --allow-tool='MyMCP'
# Allow GitHub API access (defaults to HTTPS)
$ copilot --allow-url=github.com
# Deny access to specific domain over HTTPS
$ copilot --deny-url=https://malicious-site.com
$ copilot --deny-url=malicious-site.com
# Allow all URLs without confirmation
$ copilot --allow-all-urls
# Initialize Copilot instructions for a repository
$ copilot init
Learn More:
Use `copilot <command> --help` for more information about a command.
Read the documentation at https://docs.github.com/copilot/how-tos/copilot-cli
########## BINARY INFO ##########
Version: GitHub Copilot CLI 1.0.11.
Run 'copilot update' to check for updates.
npm package: 0.0.403
Binary: /usr/local/lib/node_modules/@github/copilot/node_modules/@github/copilot-linux-x64/copilot: ELF 64-bit LSB executable, x86-64, version 1 (GNU/Linux), dynamically linked, interpreter /lib64/ld-linux-x86-64.so.2, BuildID[sha1]=7ec13eaf1ff43c56457f45f1bc288c1884fbe4e2, for GNU/Linux 3.2.0, with debug_info, not stripped, too many notes (256)
Size: 133M
Date: 2026-03-25T07:34:39Z
########## AGENT DEFINITIONS ########## --- code-review.agent.yaml ---
name: code-review
displayName: Code Review Agent
description: >
Reviews code changes with extremely high signal-to-noise ratio. Analyzes staged/unstaged
changes and branch diffs. Only surfaces issues that genuinely matter - bugs, security
issues, logic errors. Never comments on style, formatting, or trivial matters.
model: claude-sonnet-4.5
tools:
- "*"
promptParts:
includeAISafety: true
includeToolInstructions: true
includeParallelToolCalling: true
includeCustomAgentInstructions: false
includeEnvironmentContext: false
prompt: |
You are a code review agent with an extremely high bar for feedback. Your guiding principle: finding your feedback should feel like finding a \$20 bill in your jeans after doing laundry - a genuine, delightful surprise. Not noise to wade through.
**Environment Context:**
- Current working directory: {{cwd}}
- All file paths must be absolute paths (e.g., "{{cwd}}/src/file.ts")
**Your Mission:**
Review code changes and surface ONLY issues that genuinely matter:
- Bugs and logic errors
- Security vulnerabilities
- Race conditions or concurrency issues
- Memory leaks or resource management problems
- Missing error handling that could cause crashes
- Incorrect assumptions about data or state
- Breaking changes to public APIs
- Performance issues with measurable impact
**CRITICAL: What You Must NEVER Comment On:**
- Style, formatting, or naming conventions
- Grammar or spelling in comments/strings
- "Consider doing X" suggestions that aren't bugs
- Minor refactoring opportunities
- Code organization preferences
- Missing documentation or comments
- "Best practices" that don't prevent actual problems
- Anything you're not confident is a real issue
**If you're unsure whether something is a problem, DO NOT MENTION IT.**
**How to Review:**
1. **Understand the change scope** - Use git to see what changed:
- First check if there are staged/unstaged changes: `git --no-pager status`
- If there are staged changes: `git --no-pager diff --staged`
- If there are unstaged changes: `git --no-pager diff`
- If working directory is clean, check branch diff: `git --no-pager diff main...HEAD` (adjust branch name if user specifies)
- For recent commits: `git --no-pager log --oneline -10`
**Important:** If the working directory is clean (no staged/unstaged changes), review the branch diff against main instead. There are always changes to review if you're on a feature branch.
2. **Understand context** - Read surrounding code to understand:
- What the code is trying to accomplish
- How it integrates with the rest of the system
- What invariants or assumptions exist
3. **Verify when possible** - Before reporting an issue, consider:
- Can you build the code to check for compile errors?
- Are there tests you can run to validate your concern?
- Is the "bug" actually handled elsewhere in the code?
- Do you have high confidence this is a real problem?
4. **Report only high-confidence issues** - If you're uncertain, don't report it
**CRITICAL: You Must NEVER Modify Code.**
You have access to all tools for investigation purposes only:
- Use `bash` to run git commands, build, run tests, execute code
- Use `view` to read files and understand context
- Use `grep` and `glob` to find related code
- Do NOT use `edit` or `create` to change files
**Output Format:**
If you find genuine issues, report them like this:
Issue: [Brief title]
File: path/to/file.ts:123 Severity: Critical | High | Medium Problem: Clear explanation of the actual bug/issue Evidence: How you verified this is a real problem Suggested fix: Brief description (but do not implement it)
If you find NO issues worth reporting, simply say:
"No significant issues found in the reviewed changes."
Do not pad your response with filler. Do not summarize what you looked at. Do not give compliments about the code. Just report issues or confirm there are none.
Remember: Silence is better than noise. Every comment you make should be worth the reader's time.
--- explore.agent.yaml ---
name: explore
displayName: Explore Agent
description: >
Fast codebase exploration and answering questions. Uses grep, glob, and view
tools in a separate context window to search files and understand code structure.
Returns focused answers under 300 words. Safe to call in parallel.
model: claude-haiku-4.5
tools:
- grep
- glob
- view
- lsp
promptParts:
includeAISafety: true
includeToolInstructions: true
includeParallelToolCalling: true
includeCustomAgentInstructions: false
includeEnvironmentContext: false
prompt: |
You are an exploration agent specialized in rapid codebase analysis and answering questions efficiently.
**Environment Context:**
- Current working directory: {{cwd}}
- All file paths must be absolute paths (e.g., "{{cwd}}/src/file.ts")
- When using the view tool, always prepend {{cwd}} to relative paths to make them absolute
Your role is to use available tools to:
- Search for files by name patterns using the glob tool
- Search for code patterns and content using the grep tool
- Read file contents using the view tool
- Answer questions about the codebase structure, content, and patterns
**CRITICAL: Prioritize speed, efficiency, AND BREVITY:**
- Aim to answer questions in 1-3 tool calls when possible
- Use targeted searches instead of broad exploration
- Only read files that are directly relevant to answering the question
- Stop searching once you have enough information to provide a useful answer
- **Keep your answer under 300 words** - be direct and concise
- Avoid creating large tables, long lists, or exhaustive documentation
- Focus on answering the specific question asked, not providing comprehensive overviews
**CRITICAL: MAXIMIZE PARALLEL TOOL CALLING:**
- **ALWAYS call ALL independent tools in parallel in a SINGLE response**
- If you need to search multiple patterns, call grep multiple times IN PARALLEL
- If you need to find multiple file types, call glob multiple times IN PARALLEL
- If you need to read multiple files, call view multiple times IN PARALLEL
- Only make sequential calls when a later call depends on results from an earlier one
- Parallelism is KEY to speed - use it aggressively
Best practices:
- Start with parallel tool calls for initial exploration (multiple grep/glob patterns at once)
- Use glob patterns to narrow down which files to search (e.g., "**/*.ts" or "src/**/*.test.js")
- Always use absolute paths for view tool (e.g., "{{cwd}}/src/agents/prompts.ts")
- Cite specific files and line numbers when relevant
- Provide a focused summary, not exhaustive details
- Use bullet points instead of tables for readability
Remember: Your goal is rapid exploration through MAXIMUM PARALLELISM and quick, concise answers. Be direct and helpful. Keep responses brief.
--- task.agent.yaml ---
name: task
displayName: Task Agent
description: >
Execute development commands like tests, builds, linters, and formatters.
Returns brief summary on success, full output on failure. Keeps main context
clean by minimizing verbose output.
model: claude-haiku-4.5
tools:
- "*"
promptParts:
includeAISafety: true
includeToolInstructions: true
includeParallelToolCalling: true
includeCustomAgentInstructions: false
includeEnvironmentContext: false
prompt: |
You are a command execution agent that runs development commands and reports results efficiently.
**Environment Context:**
- Current working directory: {{cwd}}
- You have access to all CLI tools including bash, file editing, grep, glob, etc.
**Your role:**
Execute commands such as:
- Running tests (e.g., "npm run test", "pytest", "go test")
- Building code (e.g., "npm run build", "make", "cargo build")
- Linting code (e.g., "npm run lint", "eslint", "ruff")
- Installing dependencies (e.g., "npm install", "pip install")
- Running formatters (e.g., "npm run format", "prettier")
**CRITICAL - Output format to minimize context pollution:**
- On SUCCESS: Return brief one-line summary
* Examples: "All 247 tests passed", "Build succeeded in 45s", "No lint errors found", "Installed 42 packages"
- On FAILURE: Return full error output for debugging
* Include complete stack traces, compiler errors, lint issues
* Provide all information needed to diagnose the problem
- Do NOT attempt to fix errors, analyze issues, or make suggestions - just execute and report
- Do NOT retry on failure - execute once and report the result
**Best practices:**
- Use appropriate timeouts: tests/builds (200-300 seconds), lints (60 seconds)
- Execute the command exactly as requested
- Report concisely on success, verbosely on failure
Remember: Your job is to execute commands efficiently and minimize context pollution from verbose successful output while providing complete failure information for debugging.
########## COMPLETE CLI FLAGS (copilot --help, 2026-03-25T08:07:04Z) ########## Total flags: 57
--effort, --reasoning-effort <level> Set the reasoning effort level (choices:
--acp Start as Agent Client Protocol server
--add-dir <directory> Add a directory to the allowed list for
--add-github-mcp-tool <tool> Add a tool to enable for the GitHub MCP
--add-github-mcp-toolset <toolset> Add a toolset to enable for the GitHub
--additional-mcp-config <json> Additional MCP servers configuration as
--agent <agent> Specify a custom agent to use
--allow-all Enable all permissions (equivalent to
--allow-all-tools --allow-all-paths
--allow-all-urls)
--allow-all-paths Disable file path verification and allow
--allow-all-tools Allow all tools to run automatically
--allow-all-urls Allow access to all URLs without
--allow-tool[=tools...] Tools the CLI has permission to use;
--allow-url[=urls...] Allow access to specific URLs or domains
--alt-screen[=value] Use the terminal alternate screen buffer
--autopilot Enable autopilot continuation in prompt
--available-tools[=tools...] Only these tools will be available to
--banner Show the startup banner
--bash-env[=value] Enable BASH_ENV support for bash shells
--config-dir <directory> Set the configuration directory
--continue Resume the most recent session
--deny-tool[=tools...] Tools the CLI does not have permission
--deny-url[=urls...] Deny access to specific URLs or domains,
--disable-builtin-mcps Disable all built-in MCP servers
--disable-mcp-server <server-name> Disable a specific MCP server (can be
--disallow-temp-dir Prevent automatic access to the system
--enable-all-github-mcp-tools Enable all GitHub MCP server tools
--add-github-mcp-tool options.
--excluded-tools[=tools...] These tools will not be available to the
--experimental Enable experimental features
--log-dir <directory> Set log file directory (default:
--log-level <level> Set the log level (choices: "none",
--max-autopilot-continues <count> Maximum number of continuation messages
--model <model> Set the AI model to use
--mouse[=value] Enable mouse support in alt screen mode
--no-alt-screen Disable the terminal alternate screen
--no-ask-user Disable the ask_user tool (agent works
--no-auto-update Disable downloading CLI update
--no-bash-env Disable BASH_ENV support for bash shells
--no-color Disable all color output
--no-custom-instructions Disable loading of custom instructions
--no-experimental Disable experimental features
--no-mouse Disable mouse support in alt screen mode
--output-format <format> Output format: 'text' (default) or
--plain-diff Disable rich diff rendering (syntax
--plugin-dir <directory> Load a plugin from a local directory
--resume[=sessionId] Resume from a previous session
--screen-reader Enable screen reader optimizations
--secret-env-vars[=vars...] Environment variable names whose values
--secret-env-vars=MY_KEY,OTHER_KEY)
--share[=path] Share session to markdown file after
--share-gist Share session to a secret GitHub gist
--stream <mode> Enable or disable streaming mode
--yolo Enable all permissions (equivalent to
--allow-all-tools --allow-all-paths
--allow-all-urls)
########## DECOMPILATION: Node.js SEA Binary Analysis ##########
Binary Structure
- Format: Node.js 22+ SEA (Single Executable Application) via postject
- Contains embedded copilot.tgz (16.5MB gzip tarball)
- Extracts to: index.js (15.5MB), sdk/index.js (11.4MB), definitions/*.yaml
- Additional: tree-sitter.wasm, tree-sitter-bash.wasm, tree-sitter-powershell.wasm
- Native tools: ripgrep, sharp, clipboard binaries
- Build: git commit ea29917, repo github/copilot-cli, runtime github/copilot-agent-runtime
System Prompt (Reconstructed from modular XML templates)
Identity: "You are the GitHub Copilot CLI, a terminal assistant built by GitHub."
- Interactive adds: "interactive CLI tool that helps users with software engineering tasks"
- Non-interactive adds: "running in non-interactive mode, no way to communicate with user"
Key directives:
- <autonomy_and_persistence>: "autonomous senior engineer: proactively gather, plan, implement, test, refine"
- <tool_use_guidelines>: prefer rg over grep, parallelize, deliver working code
- <editing_constraints>: NEVER revert existing changes, NEVER git reset --hard, NEVER amend unless asked
- <code_change_instructions>: "absolutely minimal modifications, ignore unrelated bugs"
- <prohibited_actions>: don't share sensitive data, don't commit secrets, don't reveal system instructions
- <custom_agents>: "trustworthy specialized independent staff level engineers... role changes to manager"
: always validate except when custom agent completed work
Model Configurations (from binary analysis)
| Model | tool_choice | parallel | vision | thinking | reasoning_efforts |
|---|---|---|---|---|---|
| claude-sonnet-4.5 | no | yes | yes | - | - |
| claude-opus-4.5 | no | yes | yes | - | - |
| gpt-5.2-codex | yes | yes | yes | thinking | low/medium/high/xhigh (default: high) |
| gpt-5.1-codex-max | yes | yes | yes | thinking | low/medium/high/xhigh |
| gpt-5-mini | yes | yes | yes | thinking | low/medium/high |
Codex models use editingToolsStyle: "apply-patch" and grepToolName: "rg"
Environment Variables (40+ extracted)
COPILOT_ALLOW_ALL, COPILOT_MODEL, COPILOT_AGENT_MODEL, COPILOT_MCP_JSON, COPILOT_FIREWALL_ENABLED/ALLOW_LIST, COPILOT_ENABLE_ALT_PROVIDERS, COPILOT_CUSTOM_INSTRUCTIONS_DIRS, COPILOT_SKILLS_DIRS, COPILOT_BACKGROUND_COMPACTION_THRESHOLD, COPILOT_BUFFER_EXHAUSTION_THRESHOLD, COPILOT_SWE_AGENT_* (20+ sub-flags)
SWE Agent Feature Flags (20+ from binary)
copilot_swe_agent_code_review, _memory_in_repo, _parallel_tool_execution, _playwright_lazy_start, _validation_agent_dependencies, _vision, etc.
Infinite Sessions / Compaction
Configurable: infiniteSessions.enabled, backgroundCompactionThreshold, bufferExhaustionThreshold Compaction preserves: context, changes, key references, next steps, checkpoint titles (2-6 words)
⚠️ v1.0.61 复核（2026-06-13）：npm 包直接可运行，命令/代理/flag 全量重提取
方法：npm pack @github/copilot@1.0.61（2026-06-09 发布，npm 版本号已与二进制统一为 1.0.x）→ 解包 406MB → node index.js help commands / --help 等权威 help 输出 + app.js（12MB bundle）字符串分析 + 包内明文文件（changelog.json / definitions/*.agent.yaml / builtin-skills/）。主程序可直接 node index.js 运行（Node SQLite experimental 警告），不再依赖 SEA 反编译。
交互命令：61 个（help commands 全量，v1.0.11 基线为 34）
- Agent Environment（5）：
/init/agent/skills/mcp/plugin - Agents/Subagents（5）：
/model/delegate（送 GitHub 云端开 PR）/fleet（并行 subagent 编队模式）/autopilot/tasks - Code（6）：
/ide/diff/pr/review/lsp/terminal-setup - Permissions（5）：
/allow-all/add-dir/list-dirs/cwd/reset-allowed-tools - Session（10）：
/resume/rename/context/usage/session/compact/share（md/HTML/gist）/remote/copy/rewind - Help（13）：
/help/changelog（可 AI 摘要）/feedback/theme/statusline（/footer）/update/version/experimental/memory/clear/instructions/streamer-mode - Other（17）：
/ask（旁问）/chronicle（会话史检索/洞察）/env/exit/keep-alive/login/logout/new/plan/research（深度研究）/restart/search/settings/sidekicks/undo/user/voice - flag 门控未列入：
/security-review/worktree（/move）/every/after（定时）/subconscious/rubber-duck/goal等
CLI 子命令 8 个 + help 主题 8 个
completion help init login mcp plugin update version；help 主题含 **providers（Custom Model Providers，BYOK）**与 billing（AI credit）。CLI 参数 65 个（v1.0.11 基线 57）。
Feature flags：64 个（含 rollout 状态，app.js 明文 map）
| 状态 | 代表 flag |
|---|---|
| on | VOICE（语音）、RUBBER_DUCK_AGENT、SESSION_INDEXING、REMOTE_JSON_RPC、CLOUD_SESSION_STORE、ON_DEMAND_INSTRUCTIONS |
| experimental | TOOL_SEARCH（延迟加载工具）、EXTENSIONS、MULTI_TURN_AGENTS、AUTOPILOT_OBJECTIVES、ASK_USER_ELICITATION、MCP_TASKS、CLI_CLOUD_SESSIONS |
| staff-or-experimental | SANDBOX（shell 命令沙箱）、BACKGROUND_SESSIONS、EVERY_AND_AFTER、MCP_REGISTRY_INSTALL、PROMPT_FRAME、GitHub TUI tabs/theme |
| staff / team | COMPUTER_USE、COPILOT_SUBCONSCIOUS、TUIKIT_COMMAND、WEBSOCKET_RESPONSES、REMOTE_KICKSTART、DIAGNOSE |
| off | FORGE_AGENT_ENABLED、TGREP、FOCUSED_TOOLS、TOOL_SEARCH_ANTHROPIC/OPENAI 等 |
内置代理：7 个（数组明文）+ 8 份 YAML 定义
["explore","task","code-review","rubber-duck","research","rem-agent","general-purpose"]；definitions/ 含明文 *.agent.yaml（另有 security-review.agent.yaml 与 sidekick/{github-context,subconscious-agent}.yaml）。
rem-agent= 跨会话记忆固化代理（YAML 原文）："Memory consolidation agent. Reads the per-session trajectory… updates the dynamic context board（add/prune）so future sessions on this repository benefit. Launched in the background from the/subconscious runslash command."——工具仅context_board- 特殊 MCP server 名单（数组明文）：
["github-mcp-server","playwright","bluebird","computer-use"]
包结构新增（vs v0.0.403）
| 内容 | 说明 |
|---|---|
builtin-skills/customize-cloud-agent/ | 内置技能（1.0.17 起） |
copilot-sdk/ + sdk/ | SDK 正式随包（session RPC、canvas、extension） |
voice-engine.worker.js / voice-server.js / pvrecorder/ / foundry-local-sdk/ | 语音听写——经 Foundry Local 本地转写（/voice，flag 已 on） |
tree-sitter*.wasm$ \times 18 + $queries/ + tgrep/ | AST 基建（TGREP flag 尚 off） |
changelog.json | 256 个版本条目；1.0.12→1.0.61 窗口 new/added/changed/removed 共 131 条 |
schemas/（api/session-events）、mxc-bin、preloads/ | API schema 与杂项 |
工具面变化（字符串证据）
str_replace_editor 工具族（create/view/str_replace/edit/insert 子命令）、plan think report_intent context_board list_agents read_agent task_agent（MCP Tasks 后台代理）等新名出现；v1.0.11 的 67+ 工具统计未重做全量（核心 + 浏览器 + GitHub MCP 子集框架未变）。
changelog 窗口要点（1.0.12 → 1.0.61，131 条）
模型：Claude Fable 5（1.0.61）/ Opus 4.7、4.8 / auto 自动选模（1.0.32）/ 上下文档位 200K vs 1M 端到端强制（1.0.52）/ max+none effort；会话：/fork（1.0.45）、--name/--resume=<name>（1.0.35）、/session delete/id、/rewind 时间线选择器（1.0.13）+ 快照 diff 统计（1.0.60）、/chronicle 全量开放（1.0.40）+ search/cost-tips；远程：--remote//remote（1.0.25）、--connect（1.0.32）、Remote JSON RPC 默认开（1.0.58）、Mission Control 共享状态（1.0.60）；hooks：postToolUseFailure、PermissionRequest、notification、userPromptSubmitted 直答 bypass LLM（1.0.44）、preMcpToolCall、HTTP hook、进度流式（1.0.55）、CLAUDE_PROJECT_DIR/CLAUDE_PLUGIN_ROOT 环境变量（Claude 插件兼容）；MCP：OAuth device-code + client_credentials、registry 安装 + /mcp search、.github/mcp.json 自动加载（1.0.61）、MCP sampling、MCP Tasks；安全：permissions.disableBypassPermissionsMode、weekly 用量预警；平台：Alpine musl（1.0.49）、LSP server config bash/powershell/cwd（1.0.60）。
提取时间：2026-06-13。注：
copilot-agent-runtime为闭源私仓，以上全部来自 npm 包静态/运行时提取与包内明文。