skills.rest
ShinyGua
ShinyGuaCOMMUNITY
Explore all skills by this author →
GitHub

replacing-backgrounds

Compiles a Visual Intent Object into a ComfyUI background-replacement prompt JSON using inpainting.

Updated Apr 8, 2026
Search Tags:#image editing#background replacement#sdxl#mask#comfyui#inpainting#prompt json
One-click install
npx skills add https://github.com/ShinyGua/ComfyUI-Agent --skill replacing-backgrounds-shinygua
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: replacing-backgrounds
Source: https://github.com/ShinyGua/ComfyUI-Agent/tree/main/.claude/skills/replacing-backgrounds
Command: npx skills add https://github.com/ShinyGua/ComfyUI-Agent --skill replacing-backgrounds-shinygua

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve? Swapping the backdrop of an image while preserving the subject requires manually wiring a ComfyUI inpaint workflow with the correct mask, prompts, and sampler settings. This Skill automates that by compiling a Visual Intent Object (VIO) into a ready-to-run ComfyUI prompt JSON for background replacement. ## Core Features & Use Cases - Template-driven compilation: Patches a hand-authored SDXL inpaint template with the new background prompt, negative prompt, source image, and background mask. - Parameter overrides: Supports steps, cfg, sampler, scheduler, denoise, seed, and checkpoint overrides from the VIO, with negative seeds resolved to a random value. - Goal validation: Rejects VIOs whose goal is not replace_background, returning a structured JSON error. - Use Case: A planner emits a step with goal=replace_background and a background mask; this Skill produces the exact ComfyUI prompt JSON to regenerate the backdrop at high denoise while keeping the subject intact. ## Quick Start Provide a Visual Intent Object with goal replace_background, a background description prompt, the source image, and a background mask PNG to generate the ComfyUI prompt JSON.

Frequently Asked Questions about replacing-backgrounds

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I replace the background of an image in ComfyUI?▼

Use an inpaint workflow where the mask covers the background region and denoise is set high (0.95) to force strong regeneration. This Skill compiles a Visual Intent Object containing the new background prompt, source image, and mask into the ComfyUI prompt JSON automatically.

What inputs does background replacement with inpainting require?▼

It requires a source image, a mask PNG covering the background region to regenerate, and a text prompt describing the new backdrop. If no mask is available, a matting step must run first to produce one.

Can I override sampler settings like steps, cfg, and denoise?▼

Yes, the params field of the Visual Intent Object accepts steps, cfg, sampler_name, scheduler, and denoise overrides. A negative seed value is replaced with a time-based random seed, and model.checkpoint overrides the default SDXL checkpoint.

Why does the compile script fail with a goal error?▼

The script validates that the VIO goal equals replace_background and exits with a structured JSON error for any other goal such as inpaint. Ensure the planner step emits goal=replace_background before invoking compilation.

What are the limitations of mask-based background replacement?▼

The approach depends entirely on the quality of the caller-supplied background mask; inaccurate masks cause subject leakage or leftover backdrop. It also regenerates only the masked region, so lighting mismatches between subject and new background may still occur.

Related Skills

Explore More Design & Creative Skills→
prime-skillsprime-skills
31

image-inpainting

Remove or replace unwanted image regions using masks or descriptions.

Official
Advanced
dawsonblockdawsonblock

cli-anything-comfyui

Manage ComfyUI workflow queues and download generated images via REST API.

Community
Intermediate
williamsforealwilliamsforeal
1

json-to-comfy

Convert image-description JSON into ComfyUI prompts and workflow settings.

Community
Intermediate
prime-skillsprime-skills
31

image-edit

Route image editing requests to the appropriate RunComfy model and execute them.

Official
Advanced