mirror of
https://github.com/catalyst-cooperative/agent-skills.git
synced 2026-09-19 07:35:36 +08:00
main
Update pixi lockfile for week of 2026-09-13
Catalyst Cooperative Agent Skills
This repository contains experimental agent skills related to PUDL (the Public Utility Data Liberation Project).
datapackageprovides agents context on how to work with the Frictionless Datapackage metadata standardpudlis for exploring and working with PUDL open energy data and metadata
Installing
Most LLM agents support skills, but depending on the agent there are many different ways that skills are installed and activated.
One fairly generic, agent-agnostic method of installing them is npx skills.
If you have npm installed you should be able to do:
npx skills add catalyst-cooperative/agent-skills -s datapackage
npx skills add catalyst-cooperative/agent-skills -s pudl
Agent Skills Resources
For Users
- Using Agent Skills in VS Code
- Extend Claude with Agent Skills
- npx skills (CLI for installing skills)
- Agent Skill Installation CLI
- ⚠️ The Agent Skills Directory ⚠️
For Skills Authors
- The Agent Skills Standard
- Skill Authoring Best Practices
- Equipping Agents for the Real World With Agent Skills
- Claude Developer Guide Agent Skills Overview
- Create a Claude Plugin Marketplace
- Evals for AI Agents
- Testing Agent Skills Systematically with Evals
Agentic (Data) Engineering
- Agentic Engineering Patterns (Simon Willison)
- Zero Degree-of-Freedom LLM Coding using Executable Oracles (John Regehr)
- Dagster University AI Driven Data Engineering (Dagster)
- Best practices for LLM Dagster Development (Dagster)
- What Is Code Review For (Glyph)
- Your job is to deliver code you have proven to work (Simon Willison)
Agentic (Meta)Data Exploration
- Coding agents for data analysis (Simon Willison)
Other related skills
Agent skills defined outside of this repo that we either used in creating the Catalyst Cooperative agent skills, or that we delegate to within the skill.
Description
pudl: Explore and understand PUDL energy data: discover which tables exist, look up column meanings and usage warnings, and load Parquet files from S3 or a local…; datapackage: Explore and query any dataset annotated with a Frictionless Data Package descriptor (datapackage.json). Use this skill whenever a user wants to discover what…
Languages
Python
99.4%
Dockerfile
0.6%