Skip to content
Get packs

Stable Diffusion Image Generation

Lets the agent generate, edit and inpaint images with Stable Diffusion models via Hugging Face Diffusers.

Source Download skill
Repository
Orchestra-Research/AI-Research-SKILLs
Stars
14,195
Licence
MIT
Files
3 files

What it does

Once installed, the agent knows how to write Python code that uses the Hugging Face Diffusers library to turn text prompts into images, transform existing pictures with text guidance, fill in masked areas of a photo, and add control like edges or poses using ControlNet. It also explains settings such as steps, guidance scale, seeds for reproducible results, and negative prompts. The skill is a guide the agent follows; the actual image generation runs on your own machine or cloud GPU through the code it writes.

Great for

  • Designers or hobbyists who want to generate images from text using open Stable Diffusion models and have a suitable GPU
  • Developers building an image generation feature or custom diffusion pipeline in Python
  • Creators who need help with image-to-image edits or inpainting of existing photos

Skip it if

  • People without a GPU or Python setup who want a simple one-click tool; the code expects CUDA hardware and installed libraries
  • Anyone needing guaranteed licensing-safe or commercially cleared images; model licences and training data provenance vary and the skill does not address them
  • Users who expect the agent to call a hosted image API on their behalf; the skill focuses on local Diffusers code

Try asking

  • “Write a Python script that generates a 1024x1024 image of a mountain lake at sunset using SDXL”
  • “Show me how to inpaint the background of photo.jpg with a red car using a mask”
  • “Build an image-to-image pipeline that turns my sketch into a watercolor painting”
  • “Help me reproduce the same image with a fixed seed and swap to a faster scheduler”
Step-by-step guide: how to use Stable Diffusion Image GenerationInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 3 original files, the MIT licence and an attribution note.

    stable-diffusion.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip stable-diffusion.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip stable-diffusion.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Start with the SDXL or SD 1.5 examples and a fixed seed so you can compare changes reliably
  • Use enable_model_cpu_offload() or fp16 if GPU memory is limited, and check the troubleshooting reference for out-of-memory fixes
  • Review any generated or downloaded model code and check each model's licence on Hugging Face before commercial use

What's inside

  • references/advanced-usage.md
  • references/troubleshooting.md
  • SKILL.md

Redistributed unmodified from Orchestra-Research/AI-Research-SKILLs under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.