6eanut

6eanut / llm-wiki

Public

Claude Code skill for building persistent, interlinked knowledge bases from source documents. Knowledge is compiled once and kept current — never re-derived per query. Based on Karpathy's LLM Wiki pattern.

claude-ai claude-code claude-code-skill knowledge-base knowledge-management
13
1
100% credibility
Found Apr 30, 2026 at 13 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Shell
AI Summary

A skill for Claude Code that builds and maintains an interlinked wiki from source documents to provide persistent knowledge for AI queries.

How It Works

1
📖 Discover LLM Wiki

You hear about a clever way to help your AI coding helper remember all your project notes and docs in a growing wiki.

2
🛠️ Install the skill once

You run a quick setup script on your computer to add this wiki tool to your AI assistant's toolkit.

3
📁 Set up in your project

In your work folder, you create a special wiki area where your knowledge will live, ready for the AI to use.

4
📄 Add your documents

Drop your notes, articles, or files into a raw folder, and use a simple chat command to turn them into smart, linked pages.

5
💬 Chat naturally with AI

Start asking questions about your project, and your AI automatically checks the wiki first for the best info.

6
🔍 Get cited answers

Enjoy clear responses pulled from your wiki, complete with links to sources and growing connections between ideas.

✅ Knowledge compounds

Over time, your wiki builds a rich web of info, making your AI sessions smarter and more helpful every day.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 13 to 13 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is llm-wiki?

llm-wiki turns your source documents into a persistent, interlinked knowledge base for Claude Code sessions, inspired by Karpathy's LLM Wiki GitHub gist. Drop markdown or text files into a raw folder, ingest them via slash commands like /wiki-ingest, and Claude automatically checks the wiki's auto-generated index for answers with citations—no re-deriving knowledge per query. Built in Shell as a free Claude code skill, it installs globally with a one-time claude code install script and sets up per-project wikis.

Why is it gaining traction?

It beats typical RAG by compounding knowledge across sessions with bidirectional wikilinks, provenance tracking, and proactive rules via CLAUDE.md, so Claude queries the wiki without prompts. User-facing perks include /wiki-query for index-first retrieval, /wiki-lint for health checks, bilingual support detecting English/Chinese, and session hooks for hot-cache continuity—ideal for claude github integration in long-running projects. As a claude code cli plugin, it leverages Claude's runtime without extra dependencies.

Who should use this?

Claude Code power users managing sprawling project docs, like backend devs ingesting API specs or researchers building concept graphs from papers. Suited for bilingual teams handling English/Chinese codebases, or anyone tired of repeating context in claude github repo sessions. Skip if you're not deep into Claude Code skills or prefer vector DBs.

Verdict

Promising early experiment (13 stars, 1.0% credibility score) with solid claude code docs and uninstall script, but low maturity means expect tweaks—test on non-critical projects. Worth a claude code download for Karpathy llm wiki 2 fans eyeing persistent AI memory.

(187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.