What problem does it solve? Project documentation drifts out of sync with the code it describes: claims become wrong, facts get duplicated across files, and stale history accumulates. A wrong doc is worse than a missing one because it gets trusted, and manual audits rarely verify every claim against the actual codebase. ## Core Features & Use Cases - Claim Verification: Mechanically checks every documented claim (endpoints, env vars, paths, behaviors, absence claims) against the real code and produces an evidence table classifying each as OK or WRONG. - Doc Classification and Reduction: Classifies every doc as WRONG, DUPLICATE, INVENTORY, HISTORY, SPECULATIVE, NAVIGATION, or LOAD-BEARING, then deletes, merges, or rewrites with a required user approval gate. - Secret and Staleness Scanning: Greps docs for leaked credentials and self-admitted staleness markers before any restructuring. - Regrowth Prevention: Installs or verifies a lean-docs rule and wires it into code review, docs-update steps, and contributor guides so the doc set stays small. - Use Case: After a major release, run a scoped audit on the docs a PR touched to confirm every changed claim matches the code, delete migration narratives that git already carries, and report before/after sizes. ## Quick Start Ask the assistant to run ar-optimize-docs to audit the docs in this repository and verify every claim against the code.
