菜单
Activate when a subagent completes work and needs continuation check. Activate when a task finishes to determine next steps or when detecting work patterns in user messages. Governs automatic work continuation and queue management.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Activate when starting new work to check for established patterns. Activate when ensuring consistency with team standards or when promoting successful memory patterns. Searches and applies best practices before implementation.
Activate when user wants to save knowledge, search past decisions, or manage persistent memories. Handles architecture patterns, implementation logic, issues/fixes, and past implementations. Uses local SQLite + FTS5 + vector embeddings for fast hybrid search. Supports write, search, update, archive, and list operations.
Activate when user explicitly requests the development workflow process, asks about workflow phases, or says "start work", "begin development", "follow the process". Activate when creating PRs or deploying to production. NOT for simple questions or minor fixes. Executes AUTONOMOUSLY - only pauses when human decision is genuinely required.
Activate when reviewing code, before committing, after committing, or before merging a PR. Activate when user asks to review, audit, check for security issues, or find regressions. Analyzes code for logic errors, regressions, edge cases, security issues, and test gaps. Fixes findings AUTOMATICALLY. Required at process skill quality gates.
Activate when user asks to commit, push changes, create a PR, open a pull request, or submit changes for review. Activate when process skill reaches commit or PR phase. Provides commit message formatting and PR structure. PRs default to dev branch, not main. Works with git-privacy skill.
Activate when asked to auto-review and merge a PR. Runs a closed-loop workflow: subagent Stage 3 review -> fix findings -> re-review -> post ICC-REVIEW receipt -> merge (optional via workflow.auto_merge).
| name | autonomy |
| description | Activate when a subagent completes work and needs continuation check. Activate when a task finishes to determine next steps or when detecting work patterns in user messages. Governs automatic work continuation and queue management. |
Invoke automatically after subagent completion or when deciding next actions.
| Trigger | Action |
|---|---|
| Subagent returns completed work | Check .agent/queue/ for next item |
| Task finishes successfully | Update status, pick next pending item |
| Work pattern detected in user message | Add to work queue if L2/L3 |
| Multiple tasks identified | Queue all, parallelize if L3 |
.agent/queue/After work completes:
1. Mark current item completed in .agent/queue/
2. Check: Are there pending items in queue?
3. Check: Did the work reveal new tasks?
4. If yes → Add to queue, execute next pending item
5. If no more work → Report completion to user
Triggers queue addition:
Direct response (no queue):
Uses .agent/queue/ for cross-platform work tracking:
See work-queue skill for queue management details.
Level stored in autonomy.level (L1/L2/L3)