AI Skills

A curated collection of compact behavioral skills that change how AI coding agents plan, verify their own work and avoid repeating mistakes, installable across more than 35 agent harnesses.

Agent skills

Visit website github.com

AI Skills is a set of compact, opinionated skills that change how AI coding agents behave: rules that enforce planning discipline, catch claims of being done without verification, and stop an agent from patching symptoms instead of root causes. This repository is a read-only mirror of skills from the Whetstone plugin; edits happen upstream.

Skills install with a single command into more than 35 agent harnesses, including Claude Code, Cursor, Codex, Gemini CLI and GitHub Copilot CLI. On Claude Code, a companion plugin called Whetstone bundles the skills with specialized agents, workflow commands and an MCP server. It is written in Python and MIT licensed.

More in Agent skills

See all 581

Provenance gate
github.com
A deterministic authorization gateway that sits between an AI agent and its tools, tracing where each parameter of a side-effect call came from, and blocking the call or asking a human when the source does not check out.
- Agent skills
- AI evals
Expected Harm Calculus (EHC-4): Asimov's Four Laws for Autonomous AI Agents
github.com
A markdown framework of system rules that restates Isaac Asimov's Four Laws of Robotics as a set of prioritized loss calculations for large-language-model agents to follow, published as a discovery file for agents and crawlers to read.
- Agent skills
Vitruvius
github.com
An engineering research agent that runs a discover, read, synthesize, verify and review loop over technical questions, reading sources directly rather than inferring from titles, and publishing its own scored verifier benchmark, weaknesses included.
- Agent skills
- AI evals

More in Agent skills

Provenance gate

Expected Harm Calculus (EHC-4): Asimov's Four Laws for Autonomous AI Agents

Vitruvius

Cookies