What problem does it solve? Repository documentation drifts out of sync with code over time — READMEs contradict changelogs, setup guides reference removed features, and no one knows which file is authoritative. This Skill performs a systematic 18-section audit that surfaces contradictions, dead knowledge, traceability failures, and governance gaps before they mislead developers or AI agents. ## Core Features & Use Cases - Automated Repository Enumeration: A bundled Python script classifies files into priority buckets, detects version string conflicts, verifies internal markdown links, and maps governance artifact presence in one pass. - Ten Finding Categories: Detects Direct Contradictions (DC), Semantic Contradictions (SC), Terminology Drift (GD), Traceability Failures (TF), Documentation Drift (DD), Process Drift (PD), Dead Knowledge (DK), Orphan Knowledge (OK), Coverage Gaps (CG), and Governance Improvements (RGI) with stable, severity-rated IDs. - Five Structured Outputs: Produces a full 18-section audit report plus four standalone findings documents (Direct-Contradictions.md, Semantic-Contradictions.md, Traceability-Failures.md, Proposed-Repository-Governance-Improvements.md) that developers can act on independently. - Use Case: Point the Skill at a codebase whose CLAUDE.md, README, and config files have diverged over months of development; receive a prioritized remediation roadmap with root-cause analysis and measurable done-criteria. ## Quick Start Ask the AI to run a documentation audit on your repository by providing the repo root path or a GitHub URL to clone.
