Files
ComfyUI_Skills_OpenClaw/docs/llms.txt
科林_Cyril Pilgrim 8712aafb22 docs(core): move LLM summary files to docs directory and sync links
Motivation:
To maintain a cleaner root directory structure and follow the project's documentation organization standards.

Implementation:
- Relocated llms.txt and llms-full.txt to the docs/ folder.
- Updated file paths in README.md, README.zh.md, and PROJECT_DISCOVERY_CHECKLIST.md to point to the new location.
2026-03-08 21:11:50 +08:00

30 lines
1.1 KiB
Plaintext

# ComfyUI Skills for OpenClaw
Project: ComfyUI_Skills_OpenClaw
Primary keywords: OpenClaw, ComfyUI, Skills, 小龙虾
Language: English + Chinese
License: MIT
## One-line summary
An OpenClaw-oriented skill layer that turns ComfyUI API-format workflows into callable agent skills for local or remote image generation.
## Core capabilities
- Discover enabled workflows from a local registry.
- Map natural-language-derived args into ComfyUI workflow node fields via schema.
- Queue prompt jobs to ComfyUI servers and wait for completion.
- Download generated images to local outputs.
- Support multi-server workflow namespaces (`<server_id>/<workflow_id>`).
## Entry points
- Skill spec: `SKILL.md`
- Workflow registry CLI: `scripts/registry.py`
- Generation runner: `scripts/comfyui_client.py`
- Server manager CLI: `scripts/server_manager.py`
- Local web dashboard: `ui/app.py` and `ui/static/`
## Typical use
1. Add/configure ComfyUI servers.
2. Register workflows and schema mappings.
3. Call a workflow with `--workflow <server_id>/<workflow_id> --args '{...}'`.
4. Return local generated image path(s).