workflow-audit

To audit workflow health, analyze evolution artifacts like error-registry.json, model-performance.json, and session-postmortems to measure effectiveness, detect staleness, and identify optimization opportunities. The audit generates a structured health report with actionable recommendations and traceable evidence.

Workflow system auditing is a periodic self-audit assessing rule effectiveness, model performance, error patterns, and configuration staleness. Run it monthly or on demand to determine whether your workflow is improving, stagnating, or accumulating dead weight across evolution artifacts.

Identify recurring error patterns by analyzing error-registry.json and session-postmortems/ directories. The audit enforces guardrails by flagging recurring errors and citing data-backed evidence without making automated changes without user approval.

Yes, you can review model performance data from model-performance.json safely. The audit must cite data-backed evidence and avoids making automated changes without explicit user approval, ensuring you control all configuration modifications.

The best way to assess workflow stagnation is analyzing evolution data across CLAUDE.md, MEMORY.md, and workflow-changelog.md. This detects stale or conflicting rules and produces a structured health report with optimization opportunities for engineering teams.

No, the workflow audit requires no external dependencies or components. It operates directly on existing evolution artifacts within the ~/.claude/ directory structure to generate health reports and flag low-performance areas.