librarian

To detect drift between architecture documentation and your codebase, you can run parallel read-only scouts that audit architecture docs, catalogs, memory files, and dead references against repository state. This generates a consolidated report highlighting inconsistencies and guiding corrective actions.

Code drift detection is the process of comparing architecture docs and memory mappings against the actual repository state to find inconsistencies. It works by dispatching parallel scouts to verify catalogs and dead references, returning structured findings that expose where documentation no longer matches code.

You can fix memory mapping inconsistencies and dead references by analyzing the codebase to identify drift, then applying synthesized reports to auto-fix the discrepancies. This process audits memory files and dead references in parallel to propose targeted corrective actions and consolidate memory.

Yes, you can audit skill catalogs and memory files without modifying your repository because the verification process uses four read-only scouts. These scouts analyze architecture docs and dead references in parallel, ensuring the audit is non-destructive while returning structured findings.

You need to run a workspace health check for architecture docs and memory mappings when you suspect documentation drift has occurred. This on-demand check surveys your codebase to verify that catalogs and memory files accurately reflect the current repository state, returning a structured drift report.