Ai Image Generation

This lets Claude generate and edit images through the RunComfy CLI, which routes to 11+ models including FLUX 2, Nano Banana, GPT Image 2, Seedream, and Qwen. The skill knows when to pick which model based on your actual intent: GPT Image 2 for typography and multilingual text, Seedream 5 for photoreal portraits, Klein 4B for sub-second iteration, Nano Banana for web-grounded references. It ships the documented prompting patterns and exact bash invocations for each endpoint. Useful if you want Claude to handle the model selection logic instead of memorizing which API does what, though you'll still need a RunComfy account and the CLI installed.