The Claude Skill Reviewer is a specialized validation suite designed for developers building within the Claude Code ecosystem. It utilizes a three-agent orchestration pattern to concurrently analyze metadata persistence, code mechanics (security and hooks), and architectural integrity such as two-tier separation. By providing actionable, structured reports categorized by severity, this skill ensures that custom agents and skills meet the latest v0.5 guidelines, facilitating higher quality standards and seamless deployment for AI-driven workflows.
主要功能
01Parallel orchestration of metadata, implementation, and architecture review agents
02Architectural compliance checks for two-tier skill/agent design patterns
03Security-focused code reviews targeting JSON fencing and hook implementation
041 GitHub stars
05Automated YAML frontmatter and registry.yaml consistency validation
06Actionable reporting with grouped errors, warnings, and optimization tips
使用场景
01Debugging agent tool-use hooks and dependency declarations during development
02Ensuring legacy agent packages comply with modern Plugin Storage Conventions
03Auditing new Claude Code skills before production deployment