review-skill

To audit a Claude Code skill definition, you validate its SKILL.md file against a tiered binary checklist covering Critical, Important, and Polish levels. This automated review enforces Anthropic best practices, frontmatter compliance, directory structure, and script invocation.

Skill validation checks Agent Skills specification compliance by reviewing frontmatter, directory structure, script invocation, and resource referencing. It identifies issues across a tiered checklist and enforces community guidelines for skill quality.

Yes, you can automatically fix non-compliant SKILL.md files. The automated fixing feature corrects identified issues in skill definitions and bundled resources after validating them against Anthropic's best practices and the tiered checklist.

You validate skill frontmatter and directory structure by running an automated code audit against a tiered binary checklist. This process checks resource referencing and script invocation to ensure reliability and discoverability before publication.

The limitations of automated skill review involve its dependency on a binary checklist system. While it validates frontmatter, directory structure, and script invocation, it focuses on specification compliance rather than subjective quality or functional logic testing.

Yes, the review-skill audit works with bundled resources and scripts. It validates script invocation and resource referencing within the directory structure, and the automated fixing feature corrects identified issues across these bundled components.