原始内容
Gemini Claude Skills
Claude Code skills for integrating Google's Gemini 3 Pro models into your workflow. For more context, see this blog post.
Skills Included
1. Nano Banana Pro (nano-banana-pro)
Generate images using Google's advanced Nano Banana Pro model (gemini-3-pro-image-preview).
Capabilities:
- Accurate infographics with real data (uses Google Search grounding)
- Text rendering in images
- Cartographic visualizations and maps
- Detailed instruction following
- Multiple aspect ratios:
1:1,2:3,3:2,3:4,4:3,4:5,5:4,9:16,16:9,21:9 - Resolutions:
1K,2K,4K
2. Gemini Consultant (gemini-consultant)
Get a second opinion from Gemini 3 Pro (gemini-3-pro-preview) with Google Search grounding and vision.
Capabilities:
- Real-time web information via Google Search grounding
- Image analysis (single or multiple images)
- Configurable thinking depth (
lowfor speed,highfor complex reasoning) - Context passing for detailed questions
Prerequisites
You need a Google AI API key with access to Gemini 3 Pro models.
export GEMINI_API_KEY="your-api-key-here"
Get your API key from Google AI Studio.
Installation
# In Claude Code, add the marketplace
/plugin marketplace add stared/gemini-claude-skills
# Install the plugin
/plugin install gemini-skills@gemini-claude-skills
Usage Examples
Once installed, Claude will automatically use these skills when appropriate. You can also explicitly request them:
Image Generation
"Generate an infographic showing the water cycle"
"Create a 16:9 landscape image of a futuristic city at 4K resolution"
AI Consultation
"Ask Gemini what the latest React 19 features are"
"Get a second opinion on this code architecture from Gemini"
"Have Gemini analyze this screenshot" (attach image)
Dependencies
The scripts use uv with inline script dependencies - no manual package installation required. Dependencies are declared in each script:
# /// script
# dependencies = ["google-genai"]
# ///
Author
Piotr Migdał (@pmigdal)
License
MIT