Audits Claude subagent configurations for security, completeness, and architectural best practices.
The Meta Audit skill provides a comprehensive diagnostic layer for managing complex Claude Code environments. It systematically analyzes subagent definition files and local settings to ensure that every agent adheres to the principle of least privilege and organizational standards. By identifying overprivileged tools, mismatched model assignments, and inconsistent naming conventions, it allows developers to maintain a secure and high-performing multi-agent ecosystem that scales without configuration drift.
主な機能
01Security analysis of tool assignments to prevent unauthorized write or bash access
02Comprehensive health reporting with prioritized critical and warning alerts
03Automated discovery and validation of agent frontmatter and configurations
04Consistency checks for naming conventions and Claude model selections
05LSP tool coverage verification for language-specific subagents
0610 GitHub stars
ユースケース
01Performing a security review of custom subagents before production deployment
02Enforcing consistent configuration standards across a team of AI developers
03Optimizing agent performance by aligning tool sets with specific agent roles
