LLM Wiki — a compounding knowledge base.
RAG re-derives knowledge on every question. The LLM Wiki is different: Claude incrementally builds and maintains a persistent wiki from the sources you curate — cross-references already wired, contradictions already flagged. You source and ask; Claude does the summarizing, filing, and bookkeeping.
Three layers
Raw sources (wiki/raw/) — immutable input you curate. Wiki pages (wiki/pages/) — LLM-owned markdown with YAML frontmatter; Claude writes every page. Schema (wiki/schema.md) — tells Claude the focus areas, page types, and tag vocabulary; auto-loaded at session start.
What a page looks like
Every page opens with YAML frontmatter — title, tags, created, updated, and the sources it was built from. The body underneath is plain markdown.
Pages are filed by category, one directory per subsystem or theme, and a directory appears the first time something belongs in it. The index mirrors that structure, so the knowledge base stays browsable by subject rather than by filename.
One rule holds everywhere: code is cited by symbol and file, never by line number. Line numbers drift with the next commit and quietly turn a good page into a wrong one.
Setup
Opt in via /team-management:config — when wiki/ is absent it offers to scaffold wiki/index.md, wiki/log.md, wiki/schema.md, and wiki/raw/README.md. Toggle with wiki.enabled in config.json (default off). The wiki/ directory is unconditionally whitelisted in the DAIC hook, so wiki edits are never blocked regardless of mode.
Three commands
| Command | What it does |
|---|---|
/team-management:wiki-ingest <file> | Read a raw file, discuss takeaways, write a summary page, update related pages, then the index and log. |
/team-management:wiki-tune [section] | Interactively evolve wiki/schema.md via Q&A. |
/team-management:wiki-lint | Health-check: orphan pages, broken links, missing frontmatter, stale or contradictory content. |
Consult, then verify
Reading comes before searching. Claude checks the index and the relevant pages before re-deriving an answer from the codebase — that is the whole point of having written them down.
Every read doubles as a freshness check. When a page informs the work, Claude opens the code it cites and confirms the claim still holds. Code is ground truth. A renamed symbol or a moved file is corrected on the spot and the updated date is bumped. Anything larger or ambiguous is raised with you instead of quietly rewritten.
That is what keeps the knowledge base trustworthy as it grows.
Security: what not to put in wiki/raw/
wiki/raw/ is tracked in git. Never put secrets there — API keys, tokens, PEM/SSH keys, .env contents. git rm does not remove a file from history; if a secret lands there, rotate the credential and run git filter-repo to purge it. There is no automated secret scanning — this warning is the only guard.