documentation-auditor

Auditing documentation drift involves extracting structured claims from hierarchical workspace docs and verifying them against the actual codebase. This process checks file existence, barrel exports, and type signatures using targeted filesystem globs and greps to identify phantom files and incorrect references.

The best way to verify documentation claims is through a reverse-audit that traces claims from hierarchical workspaces back to the repository. It uses existence checks, barrel exports, type greps, and env-var reads to produce a confidence-based discrepancy report with suggested fixes.

No, the audit only suggests specific fixes without editing documentation. It produces a severity-grouped discrepancy report with errors, warnings, and info, leaving the application of fixes to the user for safety.

Yes, the audit separates high and medium-confidence checks from lower-confidence human judgment areas. It extracts only structured, verifiable claims to ensure automated verification focuses on factual codebase alignment rather than subjective documentation quality.

You can detect stale references, missing barrel exports, incorrect type signatures, and phantom files. The audit verifies structured claims against the actual codebase to highlight these synchronization failures following refactors or reorganization.

You need to run a discrepancy report when docs are stored in hierarchical workspaces and must stay synchronized with code, especially after refactors or reorganization. This prevents documentation drift by identifying claims that no longer match the repository.