What problem does it solve? LLM-maintained wikis accumulate contradictions, stale claims, orphan pages, and broken links over time, and manually auditing a growing wiki space is slow and error-prone. ## Core Features & Use Cases - Seven-category health audit: Detects contradictions, stale claims, orphan pages, concept gaps, broken wiki-links, weak provenance, and index/log drift in a target wiki space. - Severity-triaged findings: Returns findings grouped as critical, medium, or low, each with file path, quoted evidence, suggested fix, and the follow-up operation to run. - Read-only by design: Never modifies wiki content; only appends a run summary to wiki/log.md so future audits can see the run happened. - Use Case: An operation issue with operationType "lint" arrives for a wiki space; the agent walks the space with wiki_search and wiki_read_page, flags three contradictions and two orphan pages, and closes the issue with a triaged findings list. ## Quick Start Run a lint audit on the wiki space referenced by this operation issue and return the findings grouped by severity.
