Zimage

Z-Image Generation Skill

This skill generates an image from a natural-language prompt using the ModelScope Z-Image API.

When to use

• Generate a new image given a prompt
• Specify size (width × height)
• Control diffusion steps

Workflow

1. Decide if the user wants an image and/or parameters.
2. Collect required inputs:
   • prompt
   • width
   • height
   • steps
   • negative prompt (optional)
3. Run the bundled Node.js script (index.js) with provided inputs.
4. Return the result (image URL or base64 payload) to the user.

Default behavior

• If width or height is omitted, defaults to 1024.
• If steps is omitted, defaults to 30.
• Negative prompt is optional and passed as in API payload.