Skill Reviewer FAQs

Question 1

Can the Skill Reviewer automatically fix issues found during an audit?

Accepted Answer

Yes, the skill is designed to auto-fix unambiguous issues such as YAML syntax errors, broken links, and formatting inconsistencies. It only prompts the user for high-level architectural decisions or confirmation on breaking changes to maintain a smooth development workflow.

Question 2

When should I use the Skill Reviewer skill?

Accepted Answer

You should invoke this skill when preparing a skill for marketplace submission, after major package updates, if a skill hasn't been verified in over 3 months, or when you notice contradictory examples or broken links in your documentation.

Question 3

How does this skill improve Claude's performance and token usage?

Accepted Answer

It implements a progressive disclosure architecture review, ensuring that core metadata is lightweight (~100 tokens) while detailed resources are only loaded on demand. It also flags over-explained concepts to keep the context window focused on solving tasks rather than reading redundant theory.

Question 4

What technical standards does the Skill Reviewer enforce?

Accepted Answer

It enforces strict rules including 64-character lowercase/hyphenated names, 1024-character description limits, and SKILL.md body lengths under 500 lines. It also checks for anti-patterns like Windows-style paths, missing dependencies, and inconsistent terminology.

Question 5

What is the primary function of the Skill Reviewer?

Accepted Answer

The Skill Reviewer performs a comprehensive 14-phase systematic audit of Claude Code skills. It validates technical accuracy, enforces YAML standards for marketplace readiness, checks for API version drift using WebFetch, and ensures documentation follows progressive disclosure best practices to optimize token efficiency.

Skill Reviewer

Skill Reviewer

주요 기능

사용 사례

주요 기능

사용 사례