菜单
Surface past learnings relevant to the current task before starting work. Searches correction history, recalls past mistakes, and applies prior patterns. Use when starting a task, saying "what do I know about", "previous mistakes", "lessons learned", or "remind me about".
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
SkillOpt-flavored offline training loop for any SKILL.md. Treats accumulated learn-rule corrections as training trajectories, proposes bounded patches via an optimizer LLM, gates each candidate against a held-out validation set built from the user's own past corrections, and ships only candidates that demonstrably improve the score. Inspired by Microsoft SkillOpt's ReflACT pipeline (rollout → reflect → aggregate → select → update → evaluate) adapted to pro-workflow's SQLite store. Use when a skill has accumulated 8+ learn-rule rows and the user wants the skill itself to get better, not just longer.
Auto-configure quality gates, hooks, and settings for a new project. Detects project type and sets up appropriate tooling. Use when onboarding a new codebase.
Analyze permission denial patterns and generate optimized alwaysAllow and alwaysDeny rules. Use when permission prompts are slowing you down or after sessions with many denials.
Complete AI coding workflow system. Orchestration patterns, 18 hook events, 5 agents, cross-agent support, reference guides, and searchable learnings. Works with Claude Code, Cursor, and 32+ agents.
Prevent destructive operations using Claude Code hooks. Three modes — cautious (warn on dangerous commands), lockdown (restrict edits to one directory), and clear (remove restrictions). Uses PreToolUse matchers for Bash, Edit, and Write.
Render a self-contained HTML viewer for a pro-workflow wiki. Pages, sources, claims, seed queue, page-link graph and full-text search all in one file. No external dependencies, no JS framework, S3-uploadable. Use when the user wants to browse a wiki visually, share its current state with someone, audit research progress, or hand off a knowledge base. Inspired by Thariq Shihipar's "Unreasonable Effectiveness of HTML" — favors information density and shareability over markdown-only outputs.
| name | replay-learnings |
| description | Surface past learnings relevant to the current task before starting work. Searches correction history, recalls past mistakes, and applies prior patterns. Use when starting a task, saying "what do I know about", "previous mistakes", "lessons learned", or "remind me about". |
Like muscle memory for your coding sessions. Find and surface relevant learnings before you start working.
Use when starting a new task, saying "what do I know about", "before I start", "replay", or "remind me about".
auth, middleware, refactor).grep -i "auth\|middleware" .claude/LEARNED.md 2>/dev/null
grep -i "auth\|middleware" .claude/learning-log.md 2>/dev/null
grep -A2 "\[LEARN\]" CLAUDE.md | grep -i "auth\|middleware"
REPLAY BRIEFING: <task>
=======================
Past learnings (ranked by relevance):
1. [Testing] Always mock external APIs in auth tests (applied 8x)
Mistake: Called live API in tests, caused flaky failures
2. [Navigation] Auth middleware is in src/middleware/ not src/auth/ (applied 5x)
3. [Quality] Add error boundary around auth state changes (applied 3x)
Session history for similar work:
- 2026-02-01: auth refactor — 23 edits, 2 corrections (8.7% rate)
- 2026-01-28: auth middleware — 15 edits, 4 corrections (26.7% rate)
^ Higher correction rate — review patterns before starting
Suggested approach:
- Mock external APIs (learning #1)
- Check src/middleware/ first for auth code (learning #2)