What problem does it solve? After a development session, project docs (CLAUDE.md, README.md, docs/) and agent memory drift out of sync with the actual code, causing future AI sessions and human teammates to act on stale or conflicting information. ## Core Features & Use Cases - Three-layer knowledge reconciliation: Audits agent memory, project-root markdown (CLAUDE.md/AGENTS.md), and docs/ as separate audiences with distinct responsibilities. - Anti-bloat enforcement: Applies size limits and a "rules not changelog" policy to keep CLAUDE.md lean, migrating historical narrative to docs or deleting it. - Change-impact matrix: Maps code changes (new APIs, env vars, database tables, cross-project changes) to the exact documents that must be updated, including downstream integration guides. - Use Case: After shipping a new API flow across two projects, trigger the skill to enumerate every doc, update the integration guide, architecture doc, runbook, and handoff notes on both sides, prune stale memory entries, and produce a change summary. ## Quick Start Ask the agent to sync up the docs and memory after this development session so everything matches the current code.