| name | image-generation |
| description | Generate images via OpenAI (DALL-E / gpt-image-1) or Stable Diffusion (AUTOMATIC1111-compatible) endpoints. |
| version | 2026.2.21 |
| publisher | Local Publisher |
| handler_language | typescript |
| handler_file | handler.ts |
| inputs_schema | {"type":"object","properties":{"prompt":{"type":"string"},"provider":{"type":"string","enum":["openai","stable_diffusion"],"default":"openai"},"size":{"type":"string","enum":["256x256","512x512","1024x1024","1024x768","768x1024"],"default":"1024x1024"},"n":{"type":"integer","minimum":1,"maximum":4,"default":1},"style":{"type":"string","default":""}},"required":["prompt"],"additionalProperties":false} |
| outputs_schema | {"type":"object","properties":{"provider":{"type":"string"},"model":{"type":"string"},"images":{"type":"array","items":{"type":"object","properties":{"data_url":{"type":"string"},"seed":{"type":"number"},"width":{"type":"number"},"height":{"type":"number"}},"required":["data_url"]}}},"required":["images"]} |
Image Generation Skill
Generate images from text prompts using either OpenAI's image generation APIs or a Stable Diffusion endpoint (AUTOMATIC1111-compatible).
Configuration
- OpenAI:
OPENAI_API_KEY (required)OPENAI_IMAGE_MODEL (optional, default gpt-image-1)
- Stable Diffusion (AUTOMATIC1111):
STABLE_DIFFUSION_URL (optional, default http://127.0.0.1:7860)STABLE_DIFFUSION_API_KEY (optional, for reverse proxies)
Approval Policy
Image generation produces new media artifacts and can incur external API costs. Treat this as write scope and require user approval when cost controls or restricted content policies apply.
Egress / Network Policy
- Allowlist:
api.openai.com- Host for
STABLE_DIFFUSION_URL
- Denylist:
- Any other outbound domains
Scope Mapping
image.generate: write (creates media artifacts)image.generate.safe: write (same scope, but enforce safe prompt policies if configured)
Notes
This skill returns data URLs by default for easy rendering in clients. Persist outputs to object storage if needed.
