AI Image Generator

•
21
•

Generates and iteratively edits high-resolution images via the Nano Banana 2 MCP server using optimized narrative prompting.

The AI Image Generator skill equips Claude Code with the capability to create, refine, and iteratively edit images via Google Gemini through the Nano Banana 2 MCP integration. It applies prompting best practices—such as structured narrative descriptions instead of comma-separated keyword lists—to achieve high coherence and compositional accuracy. Supporting configurations from quick 1K previews to detailed 4K assets across various aspect ratios, it also incorporates context-window protection by saving output files locally to avoid terminal clutter and token bloat.

主要功能

01Context window preservation via local file disk storage
0221 GitHub stars
03Configurable resolution scaling from 1K up to 4K and custom aspect ratios
04Automated Gemini API key configuration checks and onboarding
05Narrative prompt structuring for enhanced image coherence and detail
06Iterative, multi-turn image editing with seamless continuity

使用场景

01Prototyping visual UI assets, application icons, and branding mockups
02Iteratively modifying existing generated graphics without starting from scratch
03Generating software architecture diagrams, flowcharts, and technical infographics