multica-ai/andrej-karpathy-skills

This page introduces an external open-source repository. It is not an HDATF product.

What it does

A single CLAUDE.md file, based on Andrej Karpathy's observations on LLM coding pitfalls, that gives Claude Code four principles: think before coding, simplicity first, surgical changes and goal-driven execution.

How it helps ATF

It could be a plain reference for instructions given to coding executors through Harness, especially stating assumptions and setting verifiable success criteria, which is close to how Harness checks results with evidence.

License

No license No license file found. Without a license, reuse of the code is not permitted by default.

More in this category

• openai/skills
  A now-deprecated catalog of Agent Skills for Codex. Skills are folders of instructions, scripts and resources that agents can discover and use for specific tasks, grouped here as system, curated and experimental.
• VoltAgent/awesome-openclaw-skills
  A curated list of OpenClaw skills. It gathers more than 5,400 skills filtered and sorted into categories from the official OpenClaw Skills Registry.
• anthropics/knowledge-work-plugins
  Open-source plugins for Claude Cowork, also usable in Claude Code, that bundle skills, connectors, slash commands and sub-agents for job functions such as productivity, sales and customer support. Teams can customize them with their own tools and terminology.
• blader/humanizer
  An agent skill written in Markdown that rewrites AI-sounding text so it reads like a person wrote it, without changing its meaning. It can follow a writing sample to match the user's voice.
• Jeffallan/claude-skills
  A Claude Code plugin with 67 skills for full-stack development across 12 categories, which activate based on the request and load matching reference files. It also has workflow commands that manage epics with Jira and Confluence.

Only repositories in the ranked Trendshift lists are included, and the lists are used only to find candidates. We do not copy their ranks. Descriptions, licenses and star counts come from each GitHub repository. The notes are our own reading. We have not tested these projects, and a place on a trending list does not prove quality.