# multica\-ai/andrej\-karpathy\-skills

multica\-ai/andrej\-karpathy\-skills: Claude Code Guidelines

- HackerByte page: <https://hackerbyte.net/en/projects/multica-ai-andrej-karpathy-skills-claude-code-guidelines-1142983825>
- GitHub: <https://github.com/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.

## AI-generated summary

This project provides a single CLAUDE.md file containing guidelines to improve Claude Code's behavior, inspired by Andrej Karpathy's observations on common LLM coding pitfalls. It outlines four core principles: Think Before Coding, Simplicity First, Surgical Changes, and Goal\-Driven Execution, to address issues like making wrong assumptions, overcomplicating code, and touching unrelated parts of a codebase.

## Key points

- Provides four principles: Think Before Coding, Simplicity First, Surgical Changes, and Goal\-Driven Execution.
- Aims to prevent LLMs from making incorrect assumptions, overcomplicating solutions, and making unintended code modifications.
- Can be installed as a Claude Code plugin or by appending the CLAUDE.md file to a project.
- Includes integration for Cursor IDE via a committed project rule.
- Emphasizes defining clear success criteria for LLMs to drive execution and verification.
