What problem does it solve? Writing effective prompts for Qwen-Image and Qwen-Image-Edit in ComfyUI is error-prone: vague instructions produce poor edits, unquoted text renders incorrectly, and wrong CFG/step settings waste generation time. This Skill codifies the official prompting best practices so every generation or edit request follows proven structure. ## Core Features & Use Cases - Edit Taxonomy Guidance: Classifies requests into the six official edit types (add/delete/replace, attribute change, style transfer, text editing, tone adjustment, multi-image) and rewrites vague prompts into precise instructions. - Parameter Recipes: Provides CFG, step count, resolution, seed, and negative prompt presets for scenarios from quick composition tests to final print output. - ComfyUI Workflow Wiring: Documents how to patch positive_prompt slots in workflows like qwen_multiref_3img.json and validate them before queueing. - Use Case: When asked to "make it like Ghibli", the Skill rewrites it to "Keep the girl's pose and clothing unchanged, transfer the image to Studio Ghibli anime style, soft pastel colors, hand-drawn texture, cinematic composition" with CFG 4.5 and 50 steps. ## Quick Start Use the prompt-best-qwen skill to write an optimized positive and negative prompt for editing this image with Qwen-Image-Edit in ComfyUI.
