Karpathy Knowledge Base Pattern

Build and maintain a self-compiling Obsidian markdown knowledge base. The LLM reads raw sources, writes cross-linked wiki articles, files Q&A results back into the corpus, appends every operation to a chronological log.md, and runs lint-and-heal passes.

Each topic lives in its own top-level folder (e.g. ai-harness/) with raw/, wiki/, outputs/, bases/ subtrees plus a topic-level log.md and CLAUDE.md. All topics share a single Obsidian vault at the repo root. Read references/architecture.md for the full rationale and the four-phase pipeline (ingest → compile → query → lint) adapted from Karpathy's three-op core (ingest → query → lint).

The topic's CLAUDE.md (symlinked to AGENTS.md) is the schema document — it tells the LLM the scope, conventions, current articles, and research gaps for that topic. Co-evolve it as the topic matures.

Related skills

This skill orchestrates several companion skills — use them for the sub-tasks of the pipeline:

• firecrawl — scrape URLs and crawl sites to populate <topic>/raw/ during the ingest phase.
• tweetsmash-api — fetch TweetSmash bookmark clusters into <topic>/raw/bookmarks/.
• qmd — build and query per-topic semantic collections over the vault; drives the query phase and source discovery during compile.
• obsidian-markdown — author wiki articles with valid Obsidian Flavored Markdown (wikilinks, callouts, embeds, properties).
• obsidian-bases — create .base files under <topic>/bases/ for dashboard views, filters, and formulas.
• obsidian-cli — interact with the running Obsidian vault from the command line (open notes, search, refresh indexes).