karpathy-guidelines

Apply behavioral guidelines to minimize LLM coding mistakes during review and refactoring.

Updated Apr 11, 2026
One-click install
npx skills add https://github.com/wukangcheng2944/claude-skills --skill karpathy-guidelines-wukangcheng2944
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: karpathy-guidelines
Source: https://github.com/wukangcheng2944/claude-skills/tree/main/karpathy-guidelines
Command: npx skills add https://github.com/wukangcheng2944/claude-skills --skill karpathy-guidelines-wukangcheng2944

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Behavioral guidelines to reduce common LLM coding mistakes by emphasizing explicit thinking, cautious scope, and verifiable outcomes during development, review, and refactoring.

Core Features & Use Cases

  • Think Before Coding: surface assumptions, state uncertainties, and compare alternatives before acting.
  • Simplicity First: prefer minimal, deterministic changes over speculative rewrites.
  • Surgical Changes: touch only what is necessary and keep surrounding code intact.
  • Goal-Driven Execution: define verifiable success criteria and iterate to ensure outcomes meet them.
  • Use Case: code reviews, tutoring sessions, and incremental refactors to keep projects maintainable under evolving requirements.

Quick Start

Review a piece of code using these guidelines and produce a minimal, verifiable set of changes that improve clarity and safety.

Frequently Asked Questions about karpathy-guidelines

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I prevent LLM coding mistakes during refactoring?

To prevent LLM coding mistakes during refactoring, apply behavioral guidelines that enforce explicit thinking, surgical edits, and verifiable success criteria to ensure safe, minimal changes.

What are the best guidelines for LLM code review?

Best guidelines for LLM code review emphasize surfacing hidden assumptions, preferring minimal deterministic changes over speculative rewrites, and defining goal-driven success criteria to verify outcomes.

Why do LLMs make overcomplicated code changes and how can I fix it?

LLMs make overcomplicated code changes due to hidden assumptions and speculative rewrites; you fix it by enforcing simplicity first, touching only necessary code, and requiring explicit thinking before acting.

Can I use these coding guidelines for incremental refactors?

Yes, you can use these coding guidelines for incremental refactors to keep projects maintainable under evolving requirements by producing minimal, verifiable sets of changes that improve clarity and safety.

Does applying surgical edit guidelines work with existing development workflows?

Applying surgical edit guidelines works with existing development workflows by integrating into coding, review, and tutoring sessions without requiring dependencies, keeping surrounding code intact during minimal changes.

When should I not use an LLM for code refactoring?

You should be cautious using an LLM for code refactoring when a task lacks verifiable success criteria or explicit assumptions, as the guidelines require deterministic outcomes to ensure safe, minimal changes.