🧠 Brain Audit — Knowledge Hygiene & Referential Integrity Auditor

Systematically audits agent memory banks, .agents/context/, .memory/wiki/, and documentation trees. Detects dead markdown links, broken symbol references, orphaned memory notes, leaked credentials, and stale contradictory statements — then closes the loop: severity-routed remediation, re-verification, and routing of unresolved findings to companion skills, reported step by step throughout.

When to Use

Trigger Conditions

Execute this skill when:

1. Pre-Release Documentation Audit: Verifying that all markdown links, table-of-contents anchors, and file references resolve to real files.
2. Memory Bank Hygiene: Cleaning up .memory/wiki/, .agents/context/, or project memory files after extensive multi-agent refactoring.
3. Preventing Agent Hallucinations: Stale or broken documentation causes future agents to hallucinate non-existent files or obsolete APIs.
4. Credential & Secret Sweeps: Auditing knowledge docs to guarantee zero accidentally pasted API keys or tokens.
5. Session or Sprint Closeout: Certifying knowledge hygiene before handing off work, cutting a release, or archiving a session.