Claude Skill 体检器（SKILL.md Doctor）
Agent Skill Quality Inspection and Refactoring Tool
Install & Use
Copy this prompt and send it to your AI assistant (Claude / Cursor / TRAE / Codex / WorkBuddy etc.) to auto-install:
Help me install this AI Skill: Claude Skill 体检器（SKILL.md Doctor）. It is used for: Agent Skill Quality Inspection and Refactoring Tool Full Skill content: https://321skill.com/skills/agent-quality-checker/raw/index.md Read that page and install it.
The prompt includes a link to the full Skill content. You can also view the full content.
Have you ever written a lengthy SKILL.md file but weren't sure if it truly follows Claude/Agent Skill best practices? This tool helps you quickly perform a comprehensive inspection of your Skill. It checks whether the description triggers are precise, if the SKILL.md file is too long, if progressive disclosure (references/ splitting) is used, if scripts are externalized, and if there are security issues like hardcoded absolute paths and secrets. It provides a clear scorecard and prioritized repair recommendations. If you wish, it can also directly help you split a monolithic Skill file into on-demand instruction packages.
Usage is very simple: you just need to provide the content or path of the SKILL.md file you want to check, or simply use trigger phrases like "check my skill," "SKILL.md inspection," or "is this skill compliant?" The tool will automatically start the analysis. It supports multiple trigger methods; almost any request to "evaluate/review/optimize an Agent Skill's quality or structure" will activate it.
This tool is most suitable for developers who are developing or maintaining Claude / general Agent Skills, especially those whose skill files have become bloated and difficult to maintain, or those who want to write a high-quality SKILL.md from scratch. Whether you are an agent developer, a full-stack engineer, or an AI application enthusiast, you can benefit from it.
Note: This inspector only focuses on the structure, portability, and security of the SKILL.md file; it does not evaluate the functional logic correctness of the skill itself. If your Skill has logic bugs, you need to debug them using other methods first. Additionally, the refactoring/splitting suggestions it provides require your manual confirmation before application.
Key Features
While most similar tools only check code quality, this one is specifically designed for Agent Skill SKILL.md files. It covers Agent-specific dimensions such as description trigger quality, progressive disclosure, script externalization, and portability, and can directly perform refactoring and splitting.
Limitations
Only applicable to Agent Skills in the SKILL.md format; not suitable for regular code projects or non-instruction-package-type configurations.
FAQ
How do I use this tool?
Simply tell the AI phrases like "check my skill," "SKILL.md inspection," or similar triggers, and then provide your SKILL.md content or path.
What aspects does it check?
It checks description trigger quality, SKILL.md length, progressive disclosure (references/ splitting), script externalization, portability (no hardcoded absolute paths), security (no hardcoded secrets), and provides a scorecard with repair recommendations.
Installation guide for AI assistants
If your AI coding assistant (Claude Code, Cursor, TRAE etc.) can see this page, send it this message to auto-install:
Visit https://321skill.com/skills/agent-quality-checker/raw/index.md to read the original Skill definition (Markdown format) for Claude Skill 体检器（SKILL.md Doctor）, and install it according to the instructions.
Raw Markdown URL for AI: /skills/agent-quality-checker/raw/index.md