← Back to Skills Marketplace
Comfy Skill
by
edward-hsiao
· GitHub ↗
· v1.0.0
· MIT-0
700
Downloads
0
Stars
6
Active Installs
1
Versions
Install in OpenClaw
/install comfy-skill
Description
Send user prompts to local ComfyUI at http://127.0.0.1:8188 using workflow.json to generate images.
README (SKILL.md)
ComfyUI 生成技能
描述
此技能可將提示詞發送至本地 ComfyUI 並生成圖片。
指令
generate_image(prompt: string): 接收使用者的描述並發送到 ComfyUI。
設定
- endpoint: http://127.0.0.1:8188
- workflow: workflow.json
Usage Guidance
This skill appears coherent and limited: it will send prompts to a ComfyUI instance on localhost using the included workflow.json. Before installing, ensure you actually run and trust a ComfyUI server at 127.0.0.1:8188 (otherwise requests will fail). Review ComfyUI's configuration — if your local ComfyUI is configured to call out to remote services or expose its API beyond localhost, your prompts could be forwarded outside your machine. Because SKILL.md is terse and doesn't specify exact HTTP calls, confirm the agent's network permissions and how it will format the request so it doesn't accidentally hit a different endpoint. Finally, ensure the referenced model/checkpoint files exist locally if you expect successful image generation.
Capability Analysis
Type: OpenClaw Skill
Name: comfy-skill
Version: 1.0.0
The skill bundle is a standard implementation for interacting with a local ComfyUI instance (127.0.0.1:8188) to generate images. The workflow.json contains a boilerplate Stable Diffusion 1.5 configuration, and SKILL.md defines a simple generate_image function that aligns with the stated purpose without any signs of malicious intent or prompt injection.
Capability Assessment
Purpose & Capability
Name/description claim to send prompts to a local ComfyUI at 127.0.0.1:8188 and a workflow.json is included; there are no unrelated env vars, binaries, or installs requested. The workflow references a local checkpoint name (expected for an image-generation workflow).
Instruction Scope
SKILL.md is minimal and scoped to generate_image(prompt) → send to local ComfyUI using workflow.json. It does not instruct reading unrelated files or exfiltrating data. However the instructions are vague about the exact API route/HTTP method and payload format, giving the agent discretion in how to call the local endpoint; that vagueness is a minor risk because it could lead to unexpected requests if the agent is misconfigured.
Install Mechanism
No install spec and no code files beyond workflow.json and SKILL.md. Instruction-only skills are lowest-risk because nothing is written to disk by an installer.
Credentials
The skill requests no environment variables, credentials, or config paths. That is proportionate for a local ComfyUI integration.
Persistence & Privilege
always is false and the skill is user-invocable; it does not request permanent elevated presence or modify other skills/configs.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install comfy-skill - After installation, invoke the skill by name or use
/comfy-skill - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
comfy-skill 1.0.0
- 初始版本發布
- 支援將提示詞發送至本地 ComfyUI 產生圖片
- 新增 generate_image 指令
- 可設定 ComfyUI 端點與 workflow
Metadata
Frequently Asked Questions
What is Comfy Skill?
Send user prompts to local ComfyUI at http://127.0.0.1:8188 using workflow.json to generate images. It is an AI Agent Skill for Claude Code / OpenClaw, with 700 downloads so far.
How do I install Comfy Skill?
Run "/install comfy-skill" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Comfy Skill free?
Yes, Comfy Skill is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Comfy Skill support?
Comfy Skill is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Comfy Skill?
It is built and maintained by edward-hsiao (@edward-hsiao); the current version is v1.0.0.
More Skills