karpathy-guidelines

Name: karpathy-guidelines
Availability: InStock
Author: tiancaiamao

Review code changes for hidden assumptions, unnecessary complexity, and missing verification.

5|Updated Feb 8, 2026

Search Tags:#code review #refactoring #verification #coding #assumptions #simplicity

Ready to run

npx skills add https://github.com/tiancaiamao/ai --skill karpathy-guidelines-tiancaiamao

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: karpathy-guidelines
Source: https://github.com/tiancaiamao/ai/tree/main/skills/karpathy-guidelines
Command: npx skills add https://github.com/tiancaiamao/ai --skill karpathy-guidelines-tiancaiamao

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps you avoid common AI coding mistakes by encouraging cautious planning, minimal edits, and explicit verification before making changes.

Core Features & Use Cases

Surface assumptions early: State uncertainties and tradeoffs instead of silently choosing an interpretation.
Keep changes surgical: Limit edits to the requested scope and avoid speculative refactors or added complexity.
Define success criteria: Turn vague coding tasks into testable steps for implementation, review, and refactoring.
Use cases: Helpful for code generation, bug fixes, refactors, and reviews where overengineering or hidden assumptions could cause regressions.

Quick Start

Use the karpathy-guidelines skill to review this code change for hidden assumptions, unnecessary complexity, and missing verification steps.

Frequently Asked Questions about karpathy-guidelines

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

FAQPage Schema

How do I prevent AI coding mistakes during code generation and refactoring?▼

To prevent AI coding mistakes during code generation and refactoring, you must surface assumptions early, keep changes surgical to the requested scope, and define explicit verifiable success criteria before implementation.

Why does AI introduce unnecessary complexity and speculative refactors into my code?▼

AI introduces unnecessary complexity and speculative refactors when it silently chooses interpretations without explicit assumptions. Requiring minimal surgical edits limits changes to the requested scope and avoids added risk during software changes.

What is the best way to review code changes for hidden assumptions?▼

The best way to review code changes for hidden assumptions is to check for explicit uncertainties, verify minimal surgical edits, and ensure testable success criteria were defined before the software changes were implemented.

Can I use surgical code edits for bug fixes without causing regressions?▼

Yes, you can use surgical code edits for bug fixes without causing regressions by limiting edits to the requested scope, avoiding speculative refactors, and establishing verifiable success criteria to complete the task safely.

When do I need explicit assumptions for software refactoring?▼

You need explicit assumptions for software refactoring whenever overengineering or hidden interpretations could cause regressions. Stating uncertainties and tradeoffs early prevents AI from silently choosing an interpretation during code changes.

How do I define verifiable success criteria for vague coding tasks?▼

To define verifiable success criteria for vague coding tasks, turn the implementation, review, and refactoring requirements into testable steps that explicitly confirm the requested scope was met without added complexity.