Image Generation
Use the image_generate tool when the user asks you to create, generate,
render, edit, restyle, or make a deliverable raster image.
The tool comes from the media-tools plugin. If it is not in your tool list,
tell the user to enable it with hybridclaw plugin install media-tools.
Do not use image_generate to inspect or describe an existing image. Use
vision_analyze for image understanding.
Workflow
- Call
image_generate with action: "list" if you need to check whether an
image provider is configured or explain missing auth/model setup.
- For generation, pass a clear
prompt and optional aspectRatio, size,
quality, and count. The native provider layer supports GPT Image 2,
Nano Banana 2, Grok Imagine, and FLUX.2 when those providers are configured.
- For image-to-image edits, pass reference media paths in the
image_generate image or images arguments. Paths may come from
/workspace, /discord-media-cache, /uploaded-media-cache, or safe
current-turn Discord CDN URLs.
- Treat provider
warnings as user-relevant when they change requested
options such as quality, size, aspect ratio, or count.
- Return the generated artifact directly in the final response. The tool owns
provider auth, provider quirks, file persistence, and media delivery paths.
Keep prompts concrete and visual. Include subject, composition, style, colors,
format, and any text that must appear in the image.
1---2name: image-generation3description: Generate or edit raster images with the image_generate tool.4metadata:5 source: https://skillmd.com/skills/hybridaione/image-generation6 hybridclaw:7 category: media8 tags:9 - image10 - generation11 - media12 - raster13user-invocable: true14---1516# Image Generation1718Use the `image_generate` tool when the user asks you to create, generate,19render, edit, restyle, or make a deliverable raster image.2021The tool comes from the `media-tools` plugin. If it is not in your tool list,22tell the user to enable it with `hybridclaw plugin install media-tools`.2324Do not use `image_generate` to inspect or describe an existing image. Use25`vision_analyze` for image understanding.2627## Workflow28291. Call `image_generate` with `action: "list"` if you need to check whether an30 image provider is configured or explain missing auth/model setup.312. For generation, pass a clear `prompt` and optional `aspectRatio`, `size`,32 `quality`, and `count`. The native provider layer supports GPT Image 2,33 Nano Banana 2, Grok Imagine, and FLUX.2 when those providers are configured.343. For image-to-image edits, pass reference media paths in the35 `image_generate` `image` or `images` arguments. Paths may come from36 `/workspace`, `/discord-media-cache`, `/uploaded-media-cache`, or safe37 current-turn Discord CDN URLs.384. Treat provider `warnings` as user-relevant when they change requested39 options such as quality, size, aspect ratio, or count.405. Return the generated artifact directly in the final response. The tool owns41 provider auth, provider quirks, file persistence, and media delivery paths.4243Keep prompts concrete and visual. Include subject, composition, style, colors,44format, and any text that must appear in the image.