Claude Code Skills

Open-source skills for Claude Code.

Each skill lives in its own directory and is centered around a SKILL.md entry point plus any bundled scripts it needs.

Available Skills

Skill	What it does	Requirements
gemini-image-gen	Generate images with the official Gemini API and save organized output folders with prompt.md, generated image files, and optional request.json.	GEMINI_API_KEY or GOOGLE_API_KEY

Install

npx skills add aouos/skills

Install a single skill with Claude Code:

claude skill add --from https://github.com/aouos/skills/tree/main/gemini-image-gen

For gemini-image-gen, set GEMINI_API_KEY or GOOGLE_API_KEY in your environment before use.

Get a key from Google AI Studio.

How These Skills Work

• Claude Code uses a skill's description to decide when to trigger it.
• SKILL.md contains the workflow and usage guidance the agent follows.
• Bundled scripts are usually non-interactive CLI tools; the interactive experience comes from the agent workflow, not from the script prompting on its own.

Project Structure

skills/
├── <skill-name>/
│   ├── SKILL.md
│   └── scripts/
└── ...

Contributing

PRs are welcome. When adding a new skill:

1. Create a directory named after the skill.
2. Add a concise SKILL.md with accurate name and description frontmatter.
3. Keep scripts small and dependency-light; prefer standard library tooling when possible.
4. Put durable workflow guidance in SKILL.md and avoid overloading it with fast-changing product details.
5. Verify the skill can be installed and that its main command path works end to end.

License

MIT