blueprint-audit

Audit orchestration blueprints by reading all blueprint files and cross-referencing agent instructions, workflows, templates, and documentation to detect contradictions, stale references, and misalignments. The audit produces a structured report with prioritized remediation steps.

Find missing rationale and documentation drift by scanning blueprint directories to identify directives lacking explanation, ambiguous gating language, and exception clauses. The audit checks for conflicting guidance between CLAUDE files, README docs, and agent definitions.

You can run repository static tests using pytest -m static to surface mechanical regressions like missing files or broken frontmatter. The audit integrates these static test results with cross-file consistency checks for comprehensive validation.

Fix stale references and role mismatches by running a full blueprint audit before release. This generates a prioritized remediation plan listing exact file edits needed to resolve stale paths, role mismatches, and contradictions across the blueprint structure.

You need .claude orchestration configurations in your repository to audit blueprint directories. The audit specifically targets blueprint_* directories within repositories containing these orchestration configurations to validate agent instructions, workflows, and templates.

Blueprint audit reports show contradictions when lead instructions, agent files, workflows, and templates contain conflicting guidance or mismatched agent lists. The audit cross-references these files to identify exact inconsistencies and documentation misalignments requiring fixes.