praneybehl/llm-wiki-plugin
107 stars · Last commit 2026-09-13
Andrej Karpathy's LLM Wiki pattern as a skill & Claude Code plugin — turn accumulated sources into a self-maintaining, scalable markdown knowledge base.
README preview
# LLM Wiki — a second brain for AI agents Turn PDFs, articles, transcripts, and notes into a shared wiki that your AI agents can search, cite, and keep up to date. Add a source once. Ask questions later. Keep the useful answers. Works with Claude Code, Codex, Cursor, Gemini CLI, OpenCode, OpenClaw, Pi, OMP, and Hermes. [Read the documentation](https://praneybehl.github.io/llm-wiki-plugin/). ## What is LLM Wiki? AI agents are good at the task in front of them, but a new session starts with limited context. LLM Wiki gives them a shared memory that can live in one personal wiki across all projects or inside a specific project. When you add a source, the agent turns it into linked Markdown pages. Later, it can find the right section and answer with citations. Useful answers can be saved back into the wiki, so the knowledge grows instead of being rebuilt from scratch. Everything canonical stays in readable Markdown. Default semantic search is local—no hosted vector database or embedding service. ## What's new in v3.2.0 - **Learn from completed work.** `/wiki:learn` captures verified successes/failures and consolidates cited patterns, applicability and counterexamples. - **Turn evidence into tested procedures.** `/wiki:evolve` runs training, evidence consolidation, whole-skill proposals, validation, and independent final testing, retaining rejected attempts. Adapters for nine supported agent hosts provide bounded runs, durable tool traces and cross-agent transfer; see the [execution requirements](https://praneybehl.github.io/llm-wiki-plugin/evolution.html). - **Explicit apply and rollback.** Only a reviewed, passing change can be applied; whole-skill snapshot checks prevent overwriting intervening edits. - **Runnable evaluation examples.** Separate training, validation and final-test tasks cover source-grounded answers and file artifacts, with calibrated judging, paired task-level analysis, and ingestion/graph/hybrid-search checks. Fictional fixtures stay outside the installed skill.