init
Bootstrap a new mnemo knowledge base with taxonomy directory structure and a starter SCHEMA.md. Use when starting a new wiki, setting up a second brain, initializing a personal knowledge base, or when the user says "set up my wiki", "create a knowledge base", "initialize mnemo", or "start my second brain". Run once per project before the first ingest. After init, optionally wires the wiki into CLAUDE.md for future session memory.
Repository SourceNeeds Review
ingest
Process raw files in raw/ into synthesized, categorized wiki pages via LLM analysis. Extracts entities and concepts, creates bidirectional wikilinks, and enforces source citations. Use when the user says "ingest this", "add this to my wiki", "process my notes", "compile this paper into the knowledge base", "add to second brain", or drops new files into raw/ and asks to process them.
Repository SourceNeeds Review
mine
Extract knowledge from the current conversation session worth persisting to the mnemo knowledge base. Scans for decisions, new entities, insights, and conclusions. Invoke explicitly (/mnemo:mine) or when the user expresses intent to save something ("remember this", "note that", "important") or when the agent detects high-value signals ("we decided", "in conclusion", "key insight") — in any language.
Repository SourceNeeds Review
query
Search the mnemo wiki knowledge base. Uses qmd (hybrid BM25 + vector semantic search) if configured at init, otherwise falls back to BM25 ranked retrieval. Supports tag:, since:, category:, backlinks:, and top-linked modifiers. Use when the user asks "what does my wiki say about X", "search my notes for Y", "find pages about Z", "query the knowledge base", "what do I know about X", or "look up X in my second brain". Falls back to global knowledge base if local returns no results.
Repository SourceNeeds Review