karpathy-llm-wiki
Karpathy LLM Wiki organizes your knowledge into two layers: immutable source files and compiled wiki articles that grow together. Sources feed into raw directories, then compile into topic-organized wiki pages with automatic fact-checking against originals. The wiki becomes a persistent, compounding artifact you query and maintain over time.
Karpathy LLM Wiki helps you build a personal knowledge base by organizing source material and compiling it into structured wiki articles.
AI-generated summary based on this skill's SKILL.md
Decision gist · record as of 2026-07-23
Karpathy LLM Wiki helps you build a personal knowledge base by organizing source material and compiling it into structured wiki articles. Karpathy LLM Wiki organizes your knowledge into two layers: immutable source files and compiled wiki articles that grow together. Sources feed into raw directories, then compile into topic-organized wiki pages with automatic fact-checking against originals. The wiki becomes a persistent, compounding artifact you query and maintain over time.
Use it when
- Karpathy-llm-wiki maintains two distinct layers: raw directories store immutable source files you ingest.
- Karpathy-llm-wiki performs fact verification and grounding invariant enforcement when you ingest sources into the wiki.
Verify before relying
Read SKILL.md below before installing (16 files). Open directory: indexed for reading, not audited.
Install
Astro-Han/karpathy-llm-wiki/karpathy-llm-wiki · repository language: Python
Open directory. Skills are indexed for reading, not audited. Review a skill's body before installing it.
Frequently asked questions
AI-generated answers based on this skill's SKILL.md and metadata
How do I create an LLM wiki with karpathy-llm-wiki?
Karpathy-llm-wiki organizes your knowledge into two layers: immutable source files and compiled wiki articles. You ingest sources into raw directories, which then compile into topic-organized wiki pages with automatic fact-checking. The system uses LLM capabilities to verify facts against originals and enforce grounding invariants, creating a persistent, compounding knowledge artifact you build and query over time.
What is the structure of karpathy-llm-wiki's raw and wiki directories?
Karpathy-llm-wiki maintains two distinct layers: raw directories store immutable source files you ingest, while wiki directories contain compiled articles organized by topic. Sources feed into the raw layer, then compile into structured wiki pages. This separation ensures your original sources remain unchanged while allowing the wiki to grow and evolve as a permanent knowledge artifact.
How does karpathy-llm-wiki verify facts when ingesting sources?
Karpathy-llm-wiki performs fact verification and grounding invariant enforcement when you ingest sources into the wiki. The system checks facts against original sources to maintain integrity across articles. This verification process ensures that compiled wiki pages remain grounded in your source material, preventing drift and maintaining evidence integrity as your knowledge base grows.
Can I query my personal wiki and archive the results?
Yes. Karpathy-llm-wiki lets you query your personal wiki to retrieve and synthesize knowledge on specific topics. You can then archive synthesized query results as permanent wiki pages with proper citations. This allows you to build on previous research and create new articles from query synthesis, making your wiki a continuously compounding resource.
How does karpathy-llm-wiki maintain quality and consistency?
Karpathy-llm-wiki includes linting capabilities to check wiki quality, consistency, and evidence integrity across articles. The system verifies that compiled pages remain grounded in source material and enforces structural consistency. Regular linting helps catch drift and maintain the reliability of your knowledge base as it grows over time.
What license does karpathy-llm-wiki use?
Karpathy-llm-wiki is released under the MIT license, allowing you to freely use, modify, and distribute the system for personal knowledge management and other purposes.
SKILL.md
Rendered from the published skill. Quoted content, verbatim.
Karpathy LLM Wiki
Build and maintain a personal knowledge base using LLMs. You manage two directories: raw/ (immutable source material) and wiki/ (compiled knowledge articles). Sources go into raw/, you compile them into wiki articles, and the wiki compounds over time.
Core ideas from Karpathy: - "The LLM writes and maintains the wiki; the human reads and asks questions." - "The wiki is a persistent, compounding artifact."
Architecture
Three layers, all under the user's project root:
raw/ — Immutable source material. You read, never modify. Organized by topic subdirectories (e.g., raw/machine-learning/).
wiki/ — Compiled knowledge articles. You have full ownership. Organized by topic subdirectories, one level only: wiki/<topic>/<article>.md. Contains two special files:
- wiki/index.md — Global index. One row per article, grouped by topic, with link + summary +
(truncated - see the full file via the links below)
File tree — 15 files
.gitignore
LICENSE
README.md
SKILL.md
assets/karpathy-tweet.png
examples/2026-03-19-claude-code-statusline-landscape.md
examples/README.md
examples/ai-coding-tools-index.md
examples/claude-code-statusline-landscape.md
examples/log-sample.md
references/archive-template.md
references/article-template.md
references/index-template.md
references/raw-template.md
scripts/check_evidence.py
Let your AI agent find skills like this
Example. Real query, live index.
You found this page by searching. An agent finds it by wishing: SkillFed indexes 56,283 agent skills by what they can do, searchable in plain language.
wish › “Build and maintain a personal LLM-powered knowledge base with structured raw and compiled wiki layers”
Give your agent the search over MCP, or paste the wish link into any chat. No install? Search from any chat →
Related skills
This skill guides you through establishing a wiki knowledge base on NanoClaw following Karpathy's LLM Wiki architecture. It walks you through choosing a group, designing your wiki's structure and conventions, creating the necessary directories and schema layer, and configuring how the agent ingests and maintains sources. Each step is idempotent—safe to re-run without clobbering existing work.
LLM Wiki lets you build and maintain a self-contained personal knowledge base as plain markdown, organized into immutable source material and compiled articles you own. Sources land in a `raw/` directory, you compile them into `wiki/` articles, and the system keeps everything local—readable on GitHub and openable as an Obsidian vault. Query your knowledge, ingest new sources, lint for health, audit claims against their sources, or critique reasoning, all without external servers or embeddings.
Wiki Manager organizes AI-generated knowledge bases as interconnected markdown articles compiled from ingested sources. It handles hub resolution, topic sub-wikis, incremental compilation, and semantic queries without external tools—all through Claude Code.
llm-wiki lets you construct and maintain a compounding knowledge base as organized markdown files, with the agent handling summarization, cross-referencing, and consistency checks while you direct curation. Unlike RAG systems that rediscover knowledge per query, your wiki compiles once and stays current, with contradictions flagged and synthesis reflecting all ingested material.
Wiki is a knowledge base manager that helps you build interconnected markdown articles from ingested sources. Initialize topic-scoped wikis, compile raw documents incrementally, query with Codex, and maintain structured indexes—all while keeping your sources immutable and synthesis honest.
Transform loose content into a living knowledge base by ingesting articles, documents, or notes into organized wiki pages. The skill automatically extracts concepts, products, and patterns, creates interconnected pages with cross-references, and maintains a searchable index—all while preventing duplicates and keeping your knowledge network coherent.
More skills llm-wiki (MIT) · Llm Wiki (unlicensed) · llm-wiki (MIT)