image-combinerlisted

# Image Combiner Create one unified image from multiple sources. Treat the capability as a multi-reference image model plus a compact role-aware Prompt, not a temporary combiner endpoint. ## Catalog - Display name: Image Combiner - Category: Layout / composition - Status: Ready - Route label: Multi-reference image composition - Tone: cyan - Short description: Merge multiple images into one coherent scene. ## What this skill does - Assigns a clear role to every source image before generation. - Rebuilds the sources as one scene with consistent scale, perspective, lighting, color, depth, contact, and occlusion. - Preserves only the explicitly requested identities, objects, materials, markings, or scene features. - Rejects pasted-on, split-screen, collage, duplicated, or visibly disconnected results. ## How to use Attach two or more images and say which subject, environment, pose, style, or detail each image contributes. #### Put a subject into a scene ```text Place the product from the first image naturally into the room from the second image. ``` #### Combine people or objects ```text Combine these two subjects into one believable photo while preserving both appearances. ``` #### Mix subject and style references ```text Create one image using the first image as the subject and the second as the visual style. ``` ## Workflow ### 1. Assign source roles Compile only execution-changing information: ```yaml images: 2–5 by default roles: subject | environment | iden