REPO PACKAGE
google-gemini gemini-skills
OFFICIAL3skills
0ready to use
LABSCO SUMMARY
gemini-api-dev covers the core generateContent-style calls — text generation, multi-turn chat, multimodal input, image and video generation, streaming, function calling, and structured output — for both the Python and TypeScript SDKs. gemini-live-api-dev is the real-time counterpart: WebSocket audio/video/text streaming, voice activity detection, session management, and migrating older Live models to the current one. gemini-interactions-api documents the newer Interactions API surface — chat, streaming, function calling, Deep Research agents — and gemini-omni-flash-api is narrower still, covering generative video editing and frame-to-video work with one specific model.
All four need a Gemini API key; there's no offline or key-free path through any of them. They fit anyone actively writing code against Gemini's SDKs who wants current best-practice patterns instead of a model's stale memorized idea of the API — they're not a general Gemini tutorial, and assume you already know why you're calling it.
READ THE FULL ANALYSIS
The README's own table lists a fifth skill that's gone. vertex-ai-api-dev is still described there as having moved to google/skills, and that move has already happened — it isn't one of the four we found live in this repository, so the README's fifth row points elsewhere rather than to anything installable here.
The 87%/96% figure in the README is Google's own number, not an independent test. The README cites an internal evaluation showing this skill raised correct-Gemini-API-code generation to 87% with Gemini 3 Flash and 96% with Gemini 3.1 Pro, from a blog post about this same package.
Checked 19 September 2026. The repository carries its own disclaimer that this is not an officially supported Google product, despite being published under the google-gemini org and marked official in our index.
ALSO IN THIS PACKAGE
PLUGINS
1 itemPLUGIN
gemini-skills
A library of skills for the Gemini API, SDK and model interactions.
WHAT'S INSIDE
3 showing · 3 total3 of 3
SETUP
Showing all 3 skills
gemini-api-dev
Writing code that calls Gemini, starting from the question its own docs lead with: which of today's models actually does the job you have.
gemini-live-api-dev
Talk to Gemini out loud and have it talk back, with sound and video flowing both ways while the conversation is still happening.
gemini-omni-flash-api
Makes video: from a written description, from a still image, or by extending and re-editing a clip you already have.
No skill in this package matches that.
HOW TO GET IT
WHOLE PACKAGEInstall the whole package (3 skills)
npx skills add google-gemini/gemini-skillsSINGLE SKILLOr install a single skill
npx skills add google-gemini/gemini-skills --skill <name> --full-depthPick the skill name from the Skills tab — each entry there installs independently.
PLUGIN MARKETPLACEAdd the marketplace, then install the plugin
/plugin marketplace add google-gemini/gemini-skills/plugin install gemini-skills@gemini-skillsTyped inside the agent's own prompt, not in a terminal. The marketplace is called gemini-skills, which is the part after the @.