What problem does it solve? Maintaining a large LLM-operated knowledge wiki leads to accumulating quality issues: leaked sensitive data, contradictory pages, orphaned documents, broken links, and stale status metadata. This Skill runs a structured audit workflow over the entire wiki and produces a dated report of every issue found. ## Core Features & Use Cases - Automated mechanical checks: Runs wiki/script/lint.py and wiki/script/check-status.py to detect sensitive patterns (emails, API keys, passwords), frontmatter errors, link integrity violations, orphan pages, and status-chain inconsistencies. - Judgment-based review: Detects contradictions between pages, verifies resolved decisions are propagated downstream, identifies data gaps, and flags stale content that scripts cannot catch. - Governance enforcement: Checks provenance (cited sources match frontmatter), dependency direction (wiki must not link to implementation docs), and never auto-fixes sensitive data — it reports file and line numbers for user confirmation. - Use Case: After a week of wiki edits, ask the agent to lint the wiki; it runs the scripts, reviews the remaining judgment items, writes wiki/synthesis/lint-YYYY-MM-DD.md, and appends a summary entry to log.md. ## Quick Start Run the wiki lint to check the current state of the wiki and generate a report of any issues found.