nano-banana
Generate Images with Google Gemini
Also available from: inference-sh-9,inferen-sh,toolshell,tul-sh
Users need an easy way to create AI images using Google Gemini models directly from their terminal. This skill provides commands and examples for generating images with Gemini 3 Pro and Gemini 2.5 Flash via the inference.sh CLI.
Download the skill ZIP
Upload in Claude
Go to Settings → Capabilities → Skills → Upload skill
Toggle on and start using
Test it
Using "nano-banana". Generate a futuristic cityscape
Expected outcome:
The skill provides the command: infsh app run google/gemini-3-pro-image-preview --input '{"prompt": "a futuristic cityscape at sunset with flying cars"}'
Using "nano-banana". Create 4 product photos
Expected outcome:
Shows how to use num_images parameter to generate multiple variations in one command
Using "nano-banana". Edit my image
Expected outcome:
Demonstrates providing an images array with the prompt to modify an existing image
Security Audit
SafeAll static findings are false positives. This skill provides documentation with example bash commands for the inference.sh CLI tool. External commands are example commands users run themselves in their terminal, not code executed by the skill. Network URLs are legitimate documentation links. The skill is safe for publication.
Risk Factors
⚙️ External commands (31)
🌐 Network access (6)
Quality Score
What You Can Build
Create AI Illustrations
Generate custom illustrations for projects, presentations, or creative work using text prompts
Edit Photos with AI
Modify existing images by describing changes you want the AI to make
Quick Prototyping
Rapidly generate concept art and visual mockups during product development
Try These Prompts
Show me how to generate a simple image using the Gemini 2.5 Flash model with the prompt 'a red apple on a wooden table'
Generate 4 logo concepts for a tech startup using Gemini 3 Pro
Create a landscape image with aspect ratio 16:9 showing a mountain at sunset
Show me how to edit an image by adding a rainbow to a landscape photo
Best Practices
- Be specific in your prompts - include details about style, lighting, composition, and mood
- Use Gemini 2.5 Flash for faster results when prototyping or iterating on ideas
- Use Gemini 3 Pro for final outputs requiring the highest quality
Avoid
- Do not use vague prompts like 'make something cool' - be specific about your vision
- Do not skip the login step - you must authenticate with inference.sh first
- Do not exceed the maximum of 14 input images when editing