/analyze

Audits Claude Code agent definitions (.md files) for triggering reliability, system prompt quality, tool appropriateness, example coverage, and operational safety across 4-7 pillars.

Maps multi-agent system architecture: inventories agents with roles, communication graphs, authority matrix, dependencies, risks, and recommendations.

Optimizes AI agents through performance analysis, user feedback review, failure classification, baseline metrics generation, and prompt engineering enhancements.

Improves specified agent by analyzing recent uses, identifying failure patterns and corrections, updating prompt with examples and constraints, testing, and saving.

Analyzes project code across quality, security, performance, and architecture domains, producing severity-rated findings, actionable recommendations, metrics, and reports.

Analyzes repository code health via complexity metrics, git churn, and test coverage. Generates report with overview, critical issues, warnings, recommendations.