saa-agent
by mirabarukaso
Image & Video GenerationBrief
Enables AI agents to generate images using the Character Select Stand Alone App (SAA) image generation backend.
Description
What This Skill Does
A command-line tool that lets AI agents send image generation requests to a running Character Select Stand Alone App (SAA) backend over WebSocket. Supports both ComfyUI and WebUI backends with parameters for model selection, prompts, dimensions, and sampling settings.
Lets AI agents drive image generation programmatically against an existing local SAA instance without any manual UI interaction.
When to Use It
- Generating anime character art from a text prompt via a local SAA server
- Creating side-by-side two-character compositions using regional prompting
- Producing reproducible images by locking a specific seed value
- Saving base64-encoded image output for programmatic downstream handling
- Diagnosing failed generations with verbose logging to capture error details
Example Workflow
User asks: Generate an anime girl with long blue hair
- 1
Confirm SAA backend is running and SAAC feature is enabled
- 2
Obtain WebSocket address from user
- 3
Execute saa-agent.py with --positive describing the character and --negative listing unwanted qualities
- 4
Check exit code and report success or busy-state error to user
- 5
Provide generation summary including prompts, seed, and output file path
Image saved to generated_image.png with a summary of prompts and seed
Requirements
- Character Select Stand Alone App (SAA) backend installed and running at version 2.4.0 or above
- SAAC (SAA Client) feature enabled in the SAA settings
- WebSocket address of the running SAA instance
Installation
clawhub install mirabarukaso/saa-agentRun the CLI command, or paste the prompt into Claude Code, Cursor, Codex or whichever coding agent you use.