A collection of skills that teach LLM agents how to use specific tools, libraries, and workflows. Each skill is a self-contained folder with documentation, examples, and evaluation tests that can be installed into an LLM agent environment such as Claude Projects or Claude Code.
| Skill | Description |
|---|---|
| langextract-skill | Extract structured information from unstructured text using Google's LangExtract library — supports few-shot extraction, source grounding, and interactive visualization |
Each skill follows a standard layout:
skill-name/
├── README.md # Overview and installation instructions
├── SKILL.md # Technical skill definition (loaded by the agent)
├── evals/ # Evaluation test cases
│ └── evals.json
└── references/ # Supporting documentation and examples
- SKILL.md is the core file that the agent reads. It contains the workflows, parameters, and decision logic the agent needs to use the skill effectively.
- evals/ holds test scenarios for validating that the agent applies the skill correctly.
- references/ provides additional examples, patterns, or domain-specific guidance.
- Browse the available skills above and pick one.
- Open its folder and read its
README.mdfor installation steps and prerequisites. - Follow the installation instructions for your platform (Claude Projects, Claude Code, etc.).
- Start prompting — each skill's README lists the kinds of requests that activate it.
Individual skills carry their own licenses. See the LICENSE file inside each skill folder for details.