How does the salmonrk skill save tokens?

It uses template mapping to reference workflows by ID rather than sending full JSON data, and prioritizes local file paths over base64 vision data to keep the context window lean.

Can I add my own custom ComfyUI workflows?

Yes, you can place new API-formatted JSON workflows in the skill's workflow directory and update the mapping script to include them.

Do I need a running ComfyUI instance?

Yes, the skill requires access to a ComfyUI server address (local or remote) configured within the skill's tool parameters.

What are the default output paths for generated images?

Generated assets are stored in the outputs/comfy/ directory relative to your workspace root for easy management.

Does this skill support image-to-image editing?

Yes, it includes built-in support for image editing workflows, such as Qwen-based edits, with automatic image upload handling.

ComfyUI API & Image Generation

Name: ComfyUI API & Image Generation
Author: openclaw

byopenclaw

•

3,390

•

数据科学与机器学习

Integrates ComfyUI into Claude's workflow to automate high-efficiency image generation and editing via optimized API templates.

The salmonrk skill provides a professional bridge between Claude and ComfyUI, enabling seamless image generation and AI-powered editing directly from the developer console. By utilizing a sophisticated token-saving strategy that maps complex workflows to simple template IDs, it allows for high-performance visual asset creation without bloating the LLM's context window. It handles automatic image uploads for editing tasks, provides flexible orientation controls, and ensures efficient asset management, making it an essential tool for developers and designers who need to incorporate AI-generated imagery into their local development environments.

主要功能

01Direct asset delivery via local file openers to bypass base64 data bloat

023,390 GitHub stars

03Flexible orientation controls for portrait and landscape generations

04Extensible workflow architecture for custom API-formatted JSON integrations

05Token-saving template mapping to minimize context window usage

06Automated image-to-image editing with seamless local asset uploads

使用场景

01Generating UI placeholders and concept art directly during the frontend development process

02Automating batch image-to-image editing using advanced vision-model workflows

03Building a centralized image generation pipeline controlled via a remote ComfyUI server

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add openclaw/skills salmonrk

For use in Claude.ai and ChatGPT