Can I edit existing images with this skill?

Yes, you can pass existing images along with text prompts to perform semantic edits, such as changing the lighting, adding objects, or altering the art style.

Does this skill support grounding with live data?

Yes, it includes advanced support for Google Search grounding, allowing you to generate visuals based on current events or real-time data.

Which AI model does this skill use?

The skill defaults to the gemini-3-pro-image-preview model, providing the highest quality and most flexible image generation capabilities available in the Gemini ecosystem.

How are generated images handled for delivery?

Images are automatically saved to the /tmp/ directory with unique UUID filenames to ensure they are correctly detected and served to the user's interface or Discord integration.

What image formats and resolutions are supported?

Images are generated as JPEGs by default. You can specify resolutions of 1K, 2K, or 4K to balance speed and visual fidelity.

Gemini Image Generation

Name: Gemini Image Generation
Author: chekos

bychekos

0•

Design Tools

Generates and edits high-resolution images using the Gemini API with granular control over resolution, aspect ratios, and multi-turn refinement.

This skill provides direct, low-level access to Google's Gemini Pro image generation models, allowing developers to programmatically create and manipulate visual assets. It offers precise control over technical parameters including 1K-4K resolutions and ten distinct aspect ratios, ranging from cinematic widescreen to vertical formats. Beyond simple generation, it supports advanced features like multi-turn chat refinement for iterative design, semantic image-to-image editing, and Google Search grounding to incorporate real-time data into visual outputs. It is specifically architected to handle unique file naming and directory requirements for seamless delivery within integrated development environments.

Key Features

01Multi-turn refinement capabilities for iterative, chat-based image editing.

02Google Search grounding to generate images based on real-time information.

03Multi-image reference support to combine elements from up to 14 source files.

040 GitHub stars

05Support for 10+ aspect ratios including 1:1, 16:9, 21:9, and 9:16.

06High-resolution image generation supporting 1K, 2K, and 4K quality outputs.

Use Cases

01Iteratively refining visual concepts through natural language feedback and semantic masking.

02Creating high-fidelity product mockups and marketing illustrations for documentation.

03Developing custom UI assets, logos, and stickers directly within a coding workflow.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add chekos/bns-marketplace image-generation

For use in Claude.ai and ChatGPT

Download Skill