audit-specs

To automate code specification audits, you run semantic analysis scripts that compare specs against the actual codebase to detect drift and discrepancies. This process validates library-wide consistency and outputs structured drift reports with severity findings.

Semantic specification review is the process of verifying that code specifications align with the actual code in a library. By running automated audits to identify inconsistencies, it prevents drift and ensures documentation matches the implementation.

Yes, you can scope specification verification to recent code changes by providing a DIFF_CONTEXT parameter. This contextual verification restricts the semantic audit to only claims affected by the specific changes, optimizing the review process.

Yes, you need Python scripting and analysis libraries to run codebase verification. The skill requires access to both spec and code repositories to execute the semantic audits and generate structured drift reports with severity findings.

The best way to audit code specs across an entire library is to run a default parallelized audit. You can also target specific stems for scoped checks, generating drift reports that categorize inconsistencies into high, medium, and low severity issues.