Skip to content
Back to search

llm-wiki-skill

Karpathy-style LLM knowledge base Agent Skill for OpenClaw/Codex. Experimental — will iterate over time.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/lewislulu--llm-wiki-skill/badge.svg)](/listings/lewislulu--llm-wiki-skill)

HTML

<a href="/listings/lewislulu--llm-wiki-skill"><img src="/api/listings/lewislulu--llm-wiki-skill/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

# Copy the skill into your agent's skills directory
cp -r llm-wiki/ ~/.claude/skills/llm-wiki/
# or for Codex
cp -r llm-wiki/ ~/.codex/skills/llm-wiki/

Then reference it in your agent config, or simply paste `llm-wiki/SKILL.md` into your agent context.

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

README.md · README · 120 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. # llm-wiki
  2. **An OpenClaw / Codex Agent Skill for building Karpathy-style LLM knowledge bases.**
  3. > Experimental skill — will iterate over time.
  4. > Please send your feedbacks in github issues.
  5. Inspired by [Andrej Karpathy's llm-wiki Gist](https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f) and the community's work building on it.
  6. ## What this is
  7. Instead of RAG (re-retrieving raw docs on every query), this pattern has the LLM **compile** raw sources into a persistent, cross-linked Markdown wiki. Every `compile`, `ingest`, `query`, `lint`, and `audit` pass makes the wiki richer. Knowledge compounds over time.
  8. - You own: sourcing raw material, asking good questions, steering direction, filing feedback on things the AI got wrong.
  9. - LLM owns: all writing, cross-referencing, filing, bookkeeping, and acting on your feedback.
  10. The skill comes with two companion tools in this repo:
  11. - **`plugins/obsidian-audit/`** — an Obsidian plugin: select text in any page, leave a comment with severity, the comment is written into `audit/` as an anchored markdown file.
  12. - **`web/`** — a local Node.js preview server: renders the wiki with mermaid, KaTeX, and wikilinks, lets you select + file feedback from the browser, and shows open audits per page.
  13. Both tools share a single TypeScript library (`audit-shared/`) so audit files written from Obsidian and the web viewer are byte-identical in shape.
  14. ## Install
  15. ```bash
  16. # Copy the skill into your agent's skills directory
  17. cp -r llm-wiki/ ~/.claude/skills/llm-wiki/
  18. # or for Codex
  19. cp -r llm-wiki/ ~/.codex/skills/llm-wiki/
  20. ```
  21. Then reference it in your agent config, or simply paste `llm-wiki/SKILL.md` into your agent context.
  22. ## Quick start
  23. ```bash
  24. # 1. Scaffold a new wiki
  25. python3 llm-wiki/scripts/scaffold.py ~/my-wiki "My Research Topic"
  26. # 2. Add a source
  27. cp my-article.md ~/my-wiki/raw/articles/
  28. # 3. Tell your agent: "ingest raw/articles/my-article.md"
  29. # 4. Ask questions: "what does the wiki say about X?"
  30. # 5. Run lint periodically
  31. python3 llm-wiki/scripts/lint_wiki.py ~/my-wiki
  32. # 6. File a comment from the web viewer or Obsidian plugin, then process it
  33. python3 llm-wiki/scripts/audit_review.py ~/my-wiki --open
  34. # then tell the agent: "audit: process the open comments"
  35. ```
  36. ## Repo contents
  37. ```
  38. llm-wiki-skill/
  39. ├── llm-wiki/ ← The skill
  40. │ ├── SKILL.md ← Main skill file (read by agent)
  41. │ ├── references/
  42. │ │ ├── schema-guide.md ← CLAUDE.md schema template
  43. │ │ ├── article-guide.md ← Article writing (divide & conquer, mermaid, KaTeX)
  44. │ │ ├── log-guide.md ← log/ folder convention
  45. │ │ ├── audit-guide.md ← audit file format + processing workflow
  46. │ │ └── tooling-tips.md ← Obsidian, qmd, plugin + web
  47. │ └── scripts/
  48. │ ├── scaffold.py ← Bootstrap new wiki directory
  49. │ ├── lint_wiki.py ← 7-pass health check (links, audit, log shape)
  50. │ └── audit_review.py ← Group open/resolved audits by target
  51. ├── audit-shared/ ← Shared TypeScript library
  52. │ └── src/{schema,anchor,id,serialize,index}.ts
  53. ├── plugins/obsidian-audit/ ← Obsidian plugin — file audit from vault
  54. └── web/ ← Local Node.js preview + feedback server
  55. ├── server/ ← Express + markdown-it + KaTeX + wikilinks
  56. └── client/ ← Vanilla-TS SPA with mermaid + selection popover
  57. ```
  58. ## Running the web viewer
  59. ```bash
  60. # one-time setup (builds audit-shared, installs deps, bundles client)
  61. cd audit-shared && npm install && npm run build && cd ..
  62. cd web && npm install && npm run build && cd ..
  63. # start the server against a wiki
  64. cd web
  65. npm start -- --wiki "/path/to/your/wiki-root" --port 4175
  66. # open http://127.0.0.1:4175
  67. ```
  68. ## Building the Obsidian plugin
  69. ```bash
  70. cd audit-shared && npm install && npm run build && cd ..
  71. cd plugins/obsidian-audit
  72. npm install
  73. npm run build
  74. npm run link -- "/path/to/your/Obsidian vault"
  75. # Enable 'LLM Wiki Audit' in Obsidian → Settings → Community plugins.
  76. ```
  77. ## Use cases
  78. - **Research deep-dive** — reading papers/articles on a topic over weeks
  79. - **Personal wiki** — Farzapedia-style: journal entries compiled into personal encyclopedia
  80. - **Team knowledge base** — fed by Slack threads, meeting notes, docs
  81. - **Reading companion** — building a rich companion wiki as you read a book
  82. ## Related work
  83. - [Karpathy's original Gist](https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f)
  84. - [pedronauck/skills karpathy-kb](https://github.com/pedronauck/skills/tree/main/skills/karpathy-kb) — full Obsidian vault integration
  85. - [Astro-Han/karpathy-llm-wiki](https://github.com/Astro-Han/karpathy-llm-wiki) — example implementation
  86. - [qmd](https://github.com/tobi/qmd) — semantic search for Markdown wikis
  87. ## License
  88. MIT

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.