# multica-ai/andrej-karpathy-skills

A single CLAUDE.md file to improve Claude Code behavior, derived from Andrej Karpathy's observations on LLM coding pitfalls.

Repository: https://github.com/multica-ai/andrej-karpathy-skills
Canonical: https://ross.abutalabs.com/products/andrej-karpathy-skills
License Family: other
Last push: 2026-04-20T10:05:04+00:00

## Health v2 (maintenance only)
Score: 50/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 78, release rhythm 35, longevity 15
- inputs: {"age_days": 218, "days_push": 135, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 207708, forks 21172 (observed 2026-08-28T04:12:28.442511+00:00)

## What it is
A single CLAUDE.md configuration file that improves Claude Code's coding behavior, based on Andrej Karpathy's observations about LLM coding pitfalls. It encodes four principles—Think Before Coding, Simplicity First, Surgical Changes, and Goal-Driven Execution—as instructions for the AI coding agent.

## Use cases
- make Claude Code stop making silent assumptions and ask clarifying questions
- reduce overengineered and bloated code generated by AI coding agents
- prevent Claude Code from touching unrelated code or comments during edits
- set up behavioral guidelines for AI pair programming in a repository
- get simpler, minimal implementations from LLM coding assistants
- improve code review quality of AI-generated changes

## When to choose
- you use Claude Code and want better default coding behavior without extra tooling
- your AI agent overcomplicates solutions or makes unverified assumptions
- you want a drop-in, single-file configuration to steer an LLM coding agent

## When to avoid
- you use coding agents other than Claude Code that don't read CLAUDE.md
- you need executable software or a tool rather than a prompt/configuration file
- you want guaranteed behavior changes rather than soft behavioral guidance

## Facets
- artifact type: learning-resource
- maturity: active
- function: prompt-engineering, developer-tools, configuration-management
- domain: developer-tools, artificial-intelligence, large-language-models, tutorials
- platform: cli, cross-platform
- tags: claude-code, claudemd, llm-guidelines, coding-agents, prompting, best-practices, andrej-karpathy, ai-assistants

## Member repositories
- multica-ai/andrej-karpathy-skills (main) score 50

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:12:28.442511+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T16:10:29.892422+00:00, confidence not recorded.
  - readme: https://github.com/multica-ai/andrej-karpathy-skills (fetched 2026-08-28T04:12:28.442511+00:00, sha c8e02b0b7517)
- Data as of 2026-08-30T08:39:29.467469+00:00.