Agent skill

Karpathy Wiki

by SherwinQ in SherwinQ/karpathy-wiki

A skill your agent uses when building or maintaining a personal knowledge base with LLM assistance.

MITAuto-check passedKnowledge Management

Install Karpathy Wiki

skills CLI
$ npx skills add SherwinQ/karpathy-wiki --skill karpathy-wiki -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install SherwinQ/karpathy-wiki karpathy-wiki --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
karpathy-wiki
GitHub stars
114
Token cost
~967 tokens
SKILL.md length
409 words
Files
25 (incl. scripts, references, assets)
Skills in repo
1
Repo updated
First seen
Licence
MIT

At a glance

A skill your agent uses when building or maintaining a personal knowledge base with LLM assistance.

  • Works in 8 steps: raw/ is immutable — sources are never… → purpose.md anchors direction — read on… → Classify before extract — determine… → …
  • Maintaining a personal knowledge base with LLM assistance
  • SKILL.md covers When to Use, When NOT to Use, Core Principles and Pipeline, plus 5 more sections
  • Ingesting sources

What it does

Karpathy Wiki is an agent skill from SherwinQ/karpathy-wiki. Use when building or maintaining a personal knowledge base with LLM assistance. Triggers on ingesting sources, compiling wiki articles, querying knowledge, running health checks, deep research, or initializing a new wiki vault. Based on Karpathy's LLM Wiki pattern with Obsidian integration.

Its SKILL.md is about 970 tokens, which your agent loads only when the skill is triggered. The skill folder holds 26 other files, including scripts, reference files and assets (for example `CHANGELOG.md`, `README.md` and `assets/concept-index-template.md`).

It sits in Knowledge Management, covering LLM wikis, Deep research and Second brain. It works with Obsidian. The repository describes itself as: 基于 [Andrej Karpathy]提出的 [LLM Wiki 模式]构建的 Agent Skill,通过四阶段流水线将碎片化信息转化为结构化、可检索、持续增长的个人知识库。 The licence is MIT.

When your agent uses it

  • Maintaining a personal knowledge base with LLM assistance
  • Ingesting sources
  • Compiling wiki articles
  • Querying knowledge

Example prompts

  • “/karpathy-wiki”

Workflow steps

8 steps, taken from the first numbered list in SKILL.md.

  1. raw/ is immutable — sources are never edited; re-ingest as new version
  2. purpose.md anchors direction — read on every operation, align output with wiki goals
  3. Classify before extract — determine entity type during Ingest, choose chapter structure during Compile
  4. Analyze before generate — two-step Compile: structured analysis → article generation
  5. Backlink audit is mandatory — bidirectional links are the knowledge graph
  6. Query answers must be archived — good answers are compounding fuel
  7. Token budget first — always read index before full articles (L0→L1→L2→L3)
  8. Human verifies, LLM maintains — LLM writes, user steers

What it can do on your machine

Read from SKILL.md and the folder at commit 1443d53. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Ships 1 file in scripts/, which the agent can run.

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Karpathy Wiki loads about 967 tokens when it runs, and up to ~7.3k if it reads all its reference files. Until then it costs about 76 tokens; SKILL.md has 409 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~76
When it runs · the whole SKILL.md, loaded when a task matches
~967
With references · SKILL.md plus every file in references/, read only if the agent opens them
~7.3k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.

SKILL.md

The full file from SherwinQ/karpathy-wiki at commit 1443d53, republished under its MIT licence (© SherwinQ). 409 words, ~967 tokens.

Download SKILL.mdSave it as .claude/skills/karpathy-wiki/SKILL.md (or your agent's skills folder). This skill also uses 24 other files; get the full folder from GitHub.
name
karpathy-wiki
description
Use when building or maintaining a personal knowledge base with LLM assistance. Triggers on ingesting sources, compiling wiki articles, querying knowledge, running health checks, deep research, or initializing a new wiki vault. Based on Karpathy's LLM Wiki pattern with Obsidian integration.

Karpathy Wiki

LLM-driven personal knowledge base that compounds over time. Raw sources in, structured wiki out.

You never write the wiki. The LLM writes everything. You just steer, and every answer compounds.

When to Use

  • User wants to ingest documents, URLs, or text into a knowledge base
  • User wants to query, compile, or maintain wiki articles
  • User needs deep research to fill knowledge gaps
  • User wants to initialize a new wiki vault

When NOT to Use

  • One-off Q&A without persistent knowledge storage
  • Note-taking without cross-referencing or structure
  • Non-text content management (images, videos alone)

Core Principles

  1. raw/ is immutable — sources are never edited; re-ingest as new version
  2. purpose.md anchors direction — read on every operation, align output with wiki goals
  3. Classify before extract — determine entity type during Ingest, choose chapter structure during Compile
  4. Analyze before generate — two-step Compile: structured analysis → article generation
  5. Backlink audit is mandatory — bidirectional links are the knowledge graph
  6. Query answers must be archived — good answers are compounding fuel
  7. Token budget first — always read index before full articles (L0→L1→L2→L3)
  8. Human verifies, LLM maintains — LLM writes, user steers

Pipeline

StageInputOutput
IngestURLs, files, textraw/ (immutable) + hash cache
Compileraw/ fileswiki/<domain>/ articles by entity type
QueryQuestionsCited answers → outputs/queries/
LintFull wikiHealth report → outputs/reports/

Each stage appends log.md. Lint and Query can trigger Deep Research to fill gaps.

Show full SKILL.md (178 more words)Show less

Token Budget

Level~TokensWhenWhat
L0~200Every sessionSKILL.md frontmatter
L1~1-2KSession startpurpose.md + Concept Index + Dashboard
L2~2-5KSearchResult summaries
L35-20KDeep readFull articles or raw sources

Entity Types

7 types, each with different required sections (see references/entity-types.md):

concept (default) · person · tool · event · comparison · pattern · overview

Commands

CommandAction
init [path] [--template <t>]Initialize vault (templates: general/research/reading/project)
ingest <source>Collect + classify + archive (no wiki generation)
compile [topic]Two-step: analyze → generate wiki article
query <question>Answer from wiki, archive result
promoteElevate quality query answer to wiki article
research <topic>Gap → web search → ingest → compile loop
delete <source>Safe delete with cascade cleanup
lintHealth check + graph analysis

Conventions

  • Obsidian CLI first for page operations (see references/obsidian-cli-cheatsheet.md)
  • Structured diff for updates: Current/Proposed/Reason/Source, user confirms each
  • Sources traceability: frontmatter sources[] links every wiki page to raw/ files
  • Incremental cache: SHA256 hash in .cache/ skips unchanged sources

References

  • references/operations.md — Full procedures for all 9 operations
  • references/frontmatter-schemas.md · compilation-guide.md · entity-types.md · obsidian-cli-cheatsheet.md
  • assets/ — All templates (purpose, wiki article, raw, query, indexes, log)
  • scripts/ — init-wiki.sh (with --template), parse-media.py

© SherwinQ, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 24 other files (scripts, references, assets) in the repository root of SherwinQ/karpathy-wiki.

  • SKILL.md
  • .gitignore
  • CHANGELOG.md
  • LICENSE
  • README.md
  • assets/concept-index-template.md
  • assets/dashboard-template.md
  • assets/log-template.md
  • assets/purpose-project-template.md
  • assets/purpose-reading-template.md
  • assets/purpose-research-template.md
  • assets/purpose-template.md
  • assets/query-output-template.md
  • assets/raw-article-template.md
  • assets/source-index-template.md
  • assets/wiki-article-template.md
  • references/compilation-guide.md
  • references/entity-types.md
  • references/frontmatter-schemas.md
  • … and 6 more

Open the folder on GitHubat commit 1443d53

Compare with similar skills

Karpathy Wiki next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Karpathy Wiki compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Karpathy Wiki this skillSherwinQ/karpathy-wiki114—~967Automated safety check: PassMIT
My LLM WikiMartinLwx/dotfiles140—~2.7kAutomated safety check: PassNone
Obsidian Knowledge BuilderMichael-OvO/obsidian-knowledge-agent205—~917Automated safety check: PassMIT
AI Second Braincharlie947/ai-second-brain189—~3.5kAutomated safety check: PassMIT
LLM Wikiinfranodus/skills119—~8.6kAutomated safety check: NotesNone
Obsidian CLI Read TransportAgriciDaniel/claude-obsidian15k—~762Automated safety check: PassMIT

Similar skills

  • My LLM Wiki

    MartinLwx/dotfiles

    Provides access to the user's personal wiki, including notes, research, project documentation, decisions, and archived knowledge.

    140 GitHub stars~2.7k tokensUpdated 2 days ago
    Knowledge ManagementAuto-check passed
  • Obsidian Knowledge Builder

    Michael-OvO/obsidian-knowledge-agent

    Turns PDFs, slides, syllabi, papers and transcripts into structured, teaching-quality Obsidian notes, matching how much structure is built to how much the material needs.

    205 GitHub stars~917 tokensUpdated 3 mo ago
    Knowledge ManagementAuto-check passed
  • AI Second Brain

    charlie947/ai-second-brain

    Walks the user through Charlie Hills's AI Second Brain + Karpathy Wiki setup from the MarTech AI newsletter.

    189 GitHub stars~3.5k tokensUpdated 5 mo ago
    Knowledge ManagementAuto-check passed
  • LLM Wiki

    infranodus/skills

    Guide users through setting up a personal LLM-maintained wiki — a persistent, compounding knowledge base where the LLM incrementally builds and maintains interlinked markdown pages from raw sources.

    119 GitHub stars~8.6k tokensUpdated 20 days ago
    Knowledge ManagementAuto-check: notes
  • Obsidian CLI Read Transport

    AgriciDaniel/claude-obsidian

    Detects and uses the official Obsidian command-line interface for read-only vault access, falling back to direct file reads, while all mutations stay on a separate transaction core.

    15k GitHub stars~762 tokensUpdated 1 mo ago
    Knowledge ManagementAuto-check passed
  • Obsidian Wiki Ingest

    AgriciDaniel/claude-obsidian

    Files a single source or a bounded batch, pasted text, staged files or an approved URL, into an Obsidian vault as cross-linked notes with provenance, agreeing scope and network access first.

    15k GitHub stars~2k tokensUpdated 1 mo ago
    Knowledge ManagementAuto-check passed

Works with

Questions about Karpathy Wiki

What does Karpathy Wiki do?

A skill your agent uses when building or maintaining a personal knowledge base with LLM assistance. Karpathy Wiki is an agent skill from SherwinQ/karpathy-wiki. Use when building or maintaining a personal knowledge base with LLM assistance.

When should I use Karpathy Wiki?

Karpathy Wiki fits situations like: maintaining a personal knowledge base with LLM assistance; ingesting sources; compiling wiki articles; querying knowledge.

How do I install Karpathy Wiki in Claude Code?

Run `npx skills add SherwinQ/karpathy-wiki --skill karpathy-wiki -a claude-code`. Or copy the skill folder (the SherwinQ/karpathy-wiki repository) into .claude/skills/karpathy-wiki in your project. Claude Code loads it when a task matches its description.

How do I install Karpathy Wiki in Codex?

Run `npx skills add SherwinQ/karpathy-wiki --skill karpathy-wiki -a codex`. Or copy the skill folder (the SherwinQ/karpathy-wiki repository) into .agents/skills/karpathy-wiki in your project. Codex loads it when a task matches its description.

Can I use Karpathy Wiki in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add SherwinQ/karpathy-wiki --skill karpathy-wiki -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/karpathy-wiki, .gemini/skills/karpathy-wiki, .github/skills/karpathy-wiki and .opencode/skills/karpathy-wiki in your project.

What does Karpathy Wiki need to run?

SKILL.md names no scripts, command-line tools or credentials: Karpathy Wiki is instructions for the agent only.

Does Karpathy Wiki access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Karpathy Wiki safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.

What licence does Karpathy Wiki use?

Karpathy Wiki is published under the MIT licence (from the LICENSE file in the skill folder). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Karpathy Wiki use?

About 967 tokens (SKILL.md is roughly 3.9k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 6.4k tokens, read only when the agent opens those files.

What are the alternatives to Karpathy Wiki?

Skills that share tags, products or a category with Karpathy Wiki: My LLM Wiki (MartinLwx/dotfiles, 140 stars), Obsidian Knowledge Builder (Michael-OvO/obsidian-knowledge-agent, 205 stars), AI Second Brain (charlie947/ai-second-brain, 189 stars) and LLM Wiki (infranodus/skills, 119 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Karpathy Wiki?

SherwinQ (a GitHub user) maintains it in SherwinQ/karpathy-wiki, which has 114 GitHub stars. The repository was last updated on April 28, 2026.

Source: SherwinQ/karpathy-wiki on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.