skill-review

To review an agent skill for quality, evaluate its purpose, approach, implementation, frontmatter, and progressive disclosure to ensure clarity, correctness, and efficiency while avoiding unnecessary token consumption.

Agent skill quality assurance is the process of evaluating skills for completeness and correctness. It is needed to prevent wasted AI tokens, ensure clear intent, and improve overall agent performance and discoverability.

To check if a skill has unnecessary scripts or references, evaluate the necessity of its components against its core purpose, ensuring it remains concise, effective, and avoids unnecessary token consumption.

When evaluating agent skill frontmatter and documentation, check for adherence to best practices, clear intent, first-principles checks, and proper progressive disclosure for optimal discoverability and usability.

Yes, you can use a code review process to validate a newly developed agent skill before deployment by reviewing its SKILL.md and associated scripts to ensure they are clear, secure, and efficient.

Automated skill review focuses on structural and documentation checks like frontmatter and progressive disclosure, but may not fully validate complex runtime execution logic or external API integration security without deeper contextual analysis.