Audits and improves AI agent skills through a systematic framework for quality, compliance, and effectiveness.
The Agent Skill Reviewer provides a comprehensive methodology for evaluating Claude Code skills against established quality standards. It streamlines the auditing process by assessing skills across four key dimensions: correctness, clarity, effectiveness, and maintainability. Designed for developers and maintainers, this skill offers actionable feedback templates, identifies common anti-patterns, and provides pattern-based guidance to troubleshoot underperforming triggers, ensuring your agent skills are production-ready and highly reliable.
Key Features
010 GitHub stars
02Pattern-based improvement guidance to fix common skill implementation issues
03Anti-pattern detection to identify rubber-stamping and style-only focus
04Standardized feedback templates with severity guidelines for PR reviews
05Step-by-step review workflow for structure, frontmatter, and content validation
06Four-dimensional evaluation framework: Correctness, Clarity, Effectiveness, and Maintainability
Use Cases
01Auditing existing AI skills for compliance with the agent-skills standard
02Providing structured, actionable feedback during skill pull request reviews
03Troubleshooting and optimizing skills that fail to trigger or provide proper guidance
