| name | image-generation |
| description | Use this skill when the user needs text-to-image or image-guided image generation, such as posters, illustrations, style changes, or creative visual drafts. |
Image Generation
Use this skill when the request is to generate a new image from text, or to generate a new image based on an uploaded image plus editing instructions.
Workflow
- Prefer
src.core.skills.image_generation_skill.generate(...) as the canonical runtime entry.
- Let the caller decide whether to upload the returned image to COS, CDN, OSS, or another permanent storage service.
- Keep prompt expansion, style planning, and result selection in the Agent layer instead of hard-coding them into the skill itself.
Notes
- This skill is the canonical image generation entry inside the project.
- The current runtime implementation calls the configured Gemini image generation API.
- The runtime implementation lives in
src/core/skills/builtin/image-generation/runtime.py.