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。
Template for commands that implement a structured recipe: validate preconditions, then execute numbered steps. Fork this and replace the placeholder content. The 'Context Validation Checkpoints' section is the key pattern, forcing Claude to verify preconditions before starting.
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。
This is a template. Replace all
[placeholder]content with your specifics.
Before executing any step, verify all of these are true. If any checkpoint fails, stop and explain why to the user.
[Clear instructions for this step. Be specific about what to read, what to check, what to write.]
Validation: [How Claude verifies this step completed correctly before moving to Step 2]
[Instructions]
Validation: [How to verify]
[Instructions]
Validation: [How to verify]
Summarize what was done:
| Situation | Response |
|---|---|
| [Error condition 1] | [What to do: retry / stop / ask user] |
| [Error condition 2] | [What to do] |
| Precondition checkpoint fails | Stop. Explain which checkpoint failed and what the user needs to fix before re-running |
If $ARGUMENTS[0] was provided: [how to use it]
If no argument was provided: [default behavior]
The "Context Validation Checkpoints" section is the key pattern from this template. It forces Claude to verify preconditions explicitly rather than discovering failures mid-execution. Inspired by recurring patterns in Packmind commands (Apache 2.0).
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。