Audit Claude Code agents, skills, and commands for quality and production readiness. Use when evaluating skill quality, checking production readiness scores, or comparing agents against best-practice templates.
日本語の概要は準備中です。原文の説明を表示しています。
Generate a structured handoff document from the current session. Captures scope, relevant files with line numbers, key discoveries, work completed, current status, next steps, and code snippets. Use before ending a session or handing work to another agent.
インストール方法を見るインストールする前に、エージェントに与えられる指示の中身を確認できます。
Generate a structured handoff document from our current conversation and save it to claudedocs/handoffs/handoff_YYYYMMDD_HHMMSS.md (or use the filename provided as $ARGUMENTS[0] if given).
Create the claudedocs/handoffs/ directory if it does not exist.
Include exactly these sections:
# Handoff: [Task Name] - [YYYY-MM-DD HH:MM]
## Task
[One sentence: what is being accomplished]
## Scope
[What needs to be done. Be specific about boundaries: what is in scope and what is explicitly out of scope]
## Files
[All relevant files touched or needed. Format: `path/to/file:line` when a specific line matters]
- `src/auth/middleware.ts:45` - token validation logic
- `tests/auth.spec.ts` - test suite for this change
- `docs/api.md` - needs update after implementation
## Discoveries
[Key findings and insights from the work so far. Things the next agent needs to know that are not obvious from the code]
- [Discovery 1]
- [Discovery 2]
## Work Done
[Completed tasks. Append-only: never remove prior entries. Include commit hashes when available]
- [x] Task A completed (commit: abc1234)
- [x] Task B completed
## Status
[Current state: what is done, what remains, test status, blockers]
## Next Steps
[Actionable checklist of remaining tasks, in order]
1. [ ] Step 1
2. [ ] Step 2
3. [ ] Step 3
## Code
[Relevant code snippets with context. Focus on the parts a resuming agent needs to understand immediately]
\`\`\`typescript
// Key implementation detail
\`\`\`
path/to/file:line format for all file references. Line numbers matter.After saving, confirm the file path.
まだレビューはありません。使ってみた感想をお寄せください。
概要と使いどころ
Audit Claude Code agents, skills, and commands for quality and production readiness. Use when evaluating skill quality, checking production readiness scores, or comparing agents against best-practice templates.
日本語の概要は準備中です。原文の説明を表示しています。
Codebase health audit scoring 7 categories with progression plan
日本語の概要は準備中です。原文の説明を表示しています。
Autonomous improvement loop: scan codebase metrics, scaffold experiment files, run agent-driven iterations until metric improves
日本語の概要は準備中です。原文の説明を表示しています。
Generate bounded independent candidates, score them against a frozen rubric, and verify the selected result with a proof log.
日本語の概要は準備中です。原文の説明を表示しています。
Post-deploy monitoring: watch production after a deploy and alert on regressions
日本語の概要は準備中です。原文の説明を表示しています。
Restore context after /clear by summarizing recent work and project state
日本語の概要は準備中です。原文の説明を表示しています。