Skill Karpathy — 4 règles pour agents de code (Claude Code, Cursor, Codex)

Ce fichier est la version officielle Agent Skill du repo multica-ai/andrej-karpathy-skills (initialement forrestchang/andrej-karpathy-skills), popularisé via l'article Karpathy's CLAUDE.md Skills File sur AgentPedia.

Andrej Karpathy a listé les défauts récurrents des LLM en coding (tweet du 26 janv. 2026). Forrest Chang les a transformés en 4 principes actionnables pour les agents.

Les 4 principes (résumé)

Comment l'installer

Claude Code (plugin — recommandé)

/plugin marketplace add forrestchang/andrej-karpathy-skills
/plugin install andrej-karpathy-skills@karpathy-skills

Projet existant (CLAUDE.md)

curl https://raw.githubusercontent.com/multica-ai/andrej-karpathy-skills/main/CLAUDE.md >> CLAUDE.md

Cursor / Codex / autres agents

Copie le SKILL.md ci-dessous dans :

• Cursor : .cursor/skills/karpathy-guidelines/SKILL.md ou règle .cursor/rules/karpathy-guidelines.mdc
• Codex : ~/.codex/skills/karpathy-guidelines/SKILL.md
• Claude Code skills : .claude/skills/karpathy-guidelines/SKILL.md

Téléchargement direct : /public/files/karpathy-guidelines/SKILL.md

SKILL.md — copier-coller (MIT, source GitHub)

---
name: karpathy-guidelines
description: Behavioral guidelines to reduce common LLM coding mistakes. Use when writing, reviewing, or refactoring code to avoid overcomplication, make surgical changes, surface assumptions, and define verifiable success criteria.
license: MIT
---

# Karpathy Guidelines

Behavioral guidelines to reduce common LLM coding mistakes, derived from [Andrej Karpathy's observations](https://x.com/karpathy/status/2015883857489522876) on LLM coding pitfalls.

**Tradeoff:** These guidelines bias toward caution over speed. For trivial tasks, use judgment.

## 1. Think Before Coding

**Don't assume. Don't hide confusion. Surface tradeoffs.**

Before implementing:
- State your assumptions explicitly. If uncertain, ask.
- If multiple interpretations exist, present them - don't pick silently.
- If a simpler approach exists, say so. Push back when warranted.
- If something is unclear, stop. Name what's confusing. Ask.

## 2. Simplicity First

**Minimum code that solves the problem. Nothing speculative.**

- No features beyond what was asked.
- No abstractions for single-use code.
- No "flexibility" or "configurability" that wasn't requested.
- No error handling for impossible scenarios.
- If you write 200 lines and it could be 50, rewrite it.

Ask yourself: "Would a senior engineer say this is overcomplicated?" If yes, simplify.

## 3. Surgical Changes

**Touch only what you must. Clean up only your own mess.**

When editing existing code:
- Don't "improve" adjacent code, comments, or formatting.
- Don't refactor things that aren't broken.
- Match existing style, even if you'd do it differently.
- If you notice unrelated dead code, mention it - don't delete it.

When your changes create orphans:
- Remove imports/variables/functions that YOUR changes made unused.
- Don't remove pre-existing dead code unless asked.

The test: Every changed line should trace directly to the user's request.

## 4. Goal-Driven Execution

**Define success criteria. Loop until verified.**

Transform tasks into verifiable goals:
- "Add validation" → "Write tests for invalid inputs, then make them pass"
- "Fix the bug" → "Write a test that reproduces it, then make it pass"
- "Refactor X" → "Ensure tests pass before and after"

For multi-step tasks, state a brief plan:

1. [Step] → verify: [check]
2. [Step] → verify: [check]
3. [Step] → verify: [check]

Strong success criteria let you loop independently. Weak criteria ("make it work") require constant clarification.

Crédits

• Observations : Andrej Karpathy
• Encodage skill : Forrest Chang / multica-ai
• Licence : MIT

Publié sur IA Outils pour la communauté vibe coding francophone — fichier identique au repo GitHub.