Build and debug ComfyUI image-generation workflows with node graphs.
npx skills add https://github.com/marduk191/qwen3_mcp --skill comfyui-workflow
Please help me install this Agent Skill. Skill: comfyui-workflow Source: https://github.com/marduk191/qwen3_mcp/tree/main/skills/comfyui-workflow Command: npx skills add https://github.com/marduk191/qwen3_mcp --skill comfyui-workflow
Builds and debugs ComfyUI image-generation workflows end-to-end, reducing setup time and integration errors.
Describe the desired ComfyUI workflow and I will generate a ready-to-run node graph.
High-intent search queries and answers about installing and using this skill.
ComfyUI workflows support txt2img, img2img, inpainting, and multi-stage pipelines. They cover conditioning strategies like ControlNet, LoRA, and IPAdapter, along with model loading and node connections for modular assembly.
Debugging ComfyUI pipelines requires validating inputs and outputs to resolve integration errors. The workflow enforces modular assembly validation and provides fallback guidance to handle common misconfigurations.
Yes, ControlNet and LoRA can be used together in a ComfyUI conditioning chain. Workflows support combining multiple conditioning strategies within multi-stage pipelines for image generation.
The best way to structure multi-stage pipelines in ComfyUI is through modular workflow assembly. This enforces input and output validation at each stage, reducing integration errors across the image-generation pipeline.
ComfyUI workflows fail during model loading or node connections due to misconfigurations. Modular assembly validation checks inputs and outputs, while fallback guidance helps resolve common integration errors.
jackspaceGuide ComfyUI workflow creation, node configuration, and optimization.
HuuBarDraft, validate, and iteratively improve AI skills with SKILL.md evaluations.
ThedouglerConstruct, modify, and debug ComfyUI image generation workflows.
hybridindieExplain ComfyUI node connections and JSON structures for image generation workflows.
