ComfyUI-OpenClaw Skill

A professional, token-saving agent skill for connecting and controlling ComfyUI via API.

Skill Structure

• Connection: Managed via skills/comfyui/.env (Private, gitignored).
• Config: Shared config.py module — single source of truth for both scripts.
• Workflow Directory: skills/comfyui/workflows/
• Private Workflows: skills/comfyui/workflows/Private/ (gitignored).
• Core Scripts:
  • comfy_client.py: Image Generation/Editing.
  • ltx2_video.py: Private Video Animation.

Tools (CLI)

1. Image Generation/Edit

python3 skills/comfyui/comfy_client.py <template_id> "<prompt>" [input_image_path] [portrait|landscape]

2. Private Video Animation

python3 skills/comfyui/ltx2_video.py <input_image_path> "<movement_prompt>"

Strategy & Security

• Token Efficiency: Refer to templates by ID; use partial file reads for large JSONs.
• Private Security: All sensitive Host IPs/Ports kept in .env (excluded from GitHub).
• Proactive Delivery: Automatically sends results to Telegram and opens on MacBook.
• Character System: Strict LoRA injection for MariClaw/AsukaClaw; auto-disabled for other prompts.