kfchou

kfchou / wiki-skills

Public

LLM-maintained personal wiki skills for Claude Code — implements Karpathy's LLM Wiki pattern

43
11
100% credibility
Found Apr 06, 2026 at 27 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
AI Summary

A plugin for an AI assistant that builds and maintains a structured, growing wiki of knowledge by adding sources and saving answers to questions.

How It Works

1
🔍 Discover Wiki Skills

You come across a helpful plugin that lets your AI build a growing collection of knowledge on any topic you care about.

2
➕ Add to your AI

In your AI's plugin store, you easily add this tool to unlock its special knowledge-building powers.

3
📂 Create your wiki

You tell your AI to start a new knowledge collection for something like research notes or personal projects, and it organizes everything neatly from scratch.

4
📄 Add sources

You share articles, files, or web pages, and your AI extracts the main ideas, links them up, and adds them to your collection.

5
❓ Ask and expand

You ask questions about your collection, get clear answers with references, and save the best ones as new entries that make it even smarter.

6
🩺 Check and improve

Your AI scans for gaps, broken connections, or updates needed, then suggests simple ways to keep your knowledge fresh and complete.

🎉 Treasure your wiki

You now have a living, interconnected knowledge base that grows richer with every source and question, always there to help you learn and remember.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 27 to 43 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 wiki-skills?

wiki-skills is a Claude Code plugin that implements Karpathy's LLM Wiki pattern, creating an LLM-maintained personal wiki as a persistent knowledge base. Instead of re-deriving answers from raw docs via RAG every time, it builds structured markdown pages that grow richer with sources and queries—using skills like wiki-init to bootstrap, wiki-ingest for adding papers or URLs, wiki-query for Q&A, wiki-lint for audits, and wiki-update for revisions. Developers get a compounding wiki for any domain, installed via Claude's marketplace.

Why is it gaining traction?

It stands out by turning Claude into a wiki curator with proactive behaviors: ingest previews key takeaways before writing, queries always cite the wiki and offer to save answers, and linting flags contradictions or gaps with fix suggestions. The backlink audits and diff previews during updates make maintenance feel effortless compared to manual wikis or basic RAG tools. Early adopters hook on the append-only log and synthesis overview for tracking knowledge evolution without losing context.

Who should use this?

Researchers ingesting papers or transcripts on a topic, backend devs documenting APIs and data flows in a codebase, or analysts tracking competitive products over time. It's ideal for anyone building personal wiki skills around self-improvement notes, book summaries, or evolving project docs where knowledge compounds session-to-session.

Verdict

Try it if you're deep into Claude Code and want Karpathy's pattern without building from scratch—solid docs and workflow make the 15 stars and 1.0% credibility score forgivable for an early plugin. Skip for production until more adoption proves reliability.

(187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.