andrej-karpathy-skills
Codex-first conversion of Karpathy-inspired coding agent guidelines.
Safety scan
A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.
Badge for your README
Markdown
[](/listings/duolahypercho--andrej-karpathy-skills)HTML
<a href="/listings/duolahypercho--andrej-karpathy-skills"><img src="/api/listings/duolahypercho--andrej-karpathy-skills/badge.svg" alt="findsafeskills safety rating" /></a>Install
Install instructions from the README
As written by the repo's author. findsafeskills has not verified these steps.
Use this repository as a local Codex plugin source.
1. Add the repository to your Codex plugin sources. 2. Codex reads `.codex-plugin/plugin.json`. 3. Codex loads `skills/andrej-karpathy-skill/SKILL.md`.
Plugin name:
andrej-karpathy-skills
Skill name:
andrej-karpathy-skill
What was scanned
The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.
README.md · README · 88 lines
View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)
- # Andrej Karpathy Skill for Codex
- A Codex-native packaging of Andrej Karpathy-inspired coding-agent guidelines.
- - `AGENTS.md` is the repository-level operating context.
- - `.codex-plugin/plugin.json` is the plugin manifest.
- - `skills/andrej-karpathy-skill/SKILL.md` is the reusable skill.
- - `instruction.md` is the copy-paste version for Codex Custom Instructions.
- - `EXAMPLES.md` shows Codex-style examples.
- English | [Simplified Chinese](./README.zh.md)
- ## Why This Exists
- Coding agents often fail in predictable ways:
- - they assume too much,
- - they overbuild simple requests,
- - they edit unrelated code,
- - they call work done without a clear success check.
- This package turns the original Karpathy-style guidelines into a Codex-ready skill, repo instruction file, and Custom Instructions block.
- ## The Four Guidelines
- | Guideline | Codex Behavior |
- |-----------|----------------|
- | Think before coding | Surface assumptions, confusion, alternatives, and tradeoffs. |
- | Keep it simple | Solve the current request without speculative features or abstractions. |
- | Make surgical changes | Touch only what the task requires and preserve surrounding code. |
- | Define and verify the goal | Turn the request into a checkable outcome before calling it done. |
- ## Install In Codex
- Use this repository as a local Codex plugin source.
- 1. Add the repository to your Codex plugin sources.
- 2. Codex reads `.codex-plugin/plugin.json`.
- 3. Codex loads `skills/andrej-karpathy-skill/SKILL.md`.
- Plugin name:
- ```text
- andrej-karpathy-skills
- ```
- Skill name:
- ```text
- andrej-karpathy-skill
- ```
- ## Use Without Installing
- If someone does not want to install the plugin or add an `AGENTS.md`, they can use the standalone Custom Instructions version:
- 1. Open `instruction.md`.
- 2. Copy the block under "Paste the block below".
- 3. Paste it into Codex Settings -> Custom Instructions.
- That version is self-contained, so Codex can apply the same behavior without reading any repo files.
- ## Repository Layout
- ```text
- .codex-plugin/plugin.json
- AGENTS.md
- EXAMPLES.md
- instruction.md
- LICENSE
- README.md
- README.zh.md
- skills/andrej-karpathy-skill/SKILL.md
- ```
- ## Attribution
- Inspired by Andrej Karpathy's public comments on LLM coding behavior and by the original community repo that packaged those ideas as reusable agent guidance.
- Original repo credit:
- ```text
- multica-ai/andrej-karpathy-skills
- ```
- ## License
- MIT
Ratings and comments
Sign in to rate this skill.
Sign inFree to use. If it’s helped, pass it on.