← Back to Skills Marketplace
2951
Downloads
2
Stars
7
Active Installs
1
Versions
Install in OpenClaw
/install comfyui-request
Description
Send a workflow request to ComfyUI and return image results.
README (SKILL.md)
comfyui-request
Purpose
Send a workflow request to a running ComfyUI instance and return the generated image URL or base64 data.
Configuration
COMFYUI_HOST: Host/IP of the ComfyUI server (default192.168.179.111).COMFYUI_PORT: Port of the ComfyUI server (default28188).COMFYUI_USER: Optional username for basic auth.COMFYUI_PASS: Optional password for basic auth.
These can be set via environment variables or a .env file in the skill directory.
Usage
{
"action": "run",
"workflow": { ... } // JSON workflow object
}
The skill will POST to http://{host}:{port}/run and return the response JSON.
Example
{
"action": "run",
"workflow": {
"nodes": [ ... ],
"edges": [ ... ]
}
}
Notes
The skill expects the ComfyUI server to expose the /run endpoint and return a JSON object containing an image field with a URL or base64 string.
Usage Guidance
This skill's code appears to perform the expected ComfyUI request/poll flow, but the documentation and declared requirements disagree with the implementation. Before installing: (1) confirm you want the default host/port (192.168.179.111:28188) or set COMFYUI_HOST/PORT to your own server; (2) note the SKILL.md says it posts to /run but the code uses /prompt and /history — if your ComfyUI API differs, the skill may fail; (3) the package declares curl as required but the script does not use it — you can remove that requirement or verify why it was declared; (4) if you plan to provide COMFYUI_USER/PASS, ensure credentials are only stored in a secure place; (5) test the CLI in a safe environment against your ComfyUI endpoint to confirm behavior, and consider asking the skill author to correct the documentation or code so endpoints and expectations match.
Capability Analysis
Type: OpenClaw Skill
Name: comfyui-request
Version: 1.0.0
The skill's behavior is fully aligned with its stated purpose of interacting with a ComfyUI server. It reads configuration (host, port, optional username/password) from environment variables and uses them to make HTTP requests to the specified ComfyUI endpoint. There is no evidence of data exfiltration, malicious execution, persistence mechanisms, or prompt injection attempts against the agent. The `SKILL.md` and `bin/cli.js` files clearly define and implement the intended functionality without any high-risk behaviors.
Capability Assessment
Purpose & Capability
The code implements sending a prompt to a local ComfyUI-like HTTP API (/prompt), polling /history/{prompt_id}, and returning image info and view URLs — this matches the stated purpose of requesting ComfyUI workflows. However the SKILL.md text claims the skill POSTs to /run and expects an 'image' field, which does not match the implementation. The skill also hardcodes a default LAN IP (192.168.179.111) and port (28188) which is plausible for local testing but may be surprising to users.
Instruction Scope
The SKILL.md instructs POSTing to /run and expecting an 'image' field (or base64) but the CLI actually POSTs to /prompt, expects a prompt_id in the response, polls /history/{prompt_id}, and extracts images from outputs. This mismatch could lead to unexpected behavior. The SKILL.md also suggests using a .env file in the skill directory; the code reads environment variables but there is no install step that creates or documents such a file.
Install Mechanism
There is no install spec and no remote downloads. The skill is instruction + a small CLI script (Node) — nothing writes arbitrary archives to disk or pulls external code at install-time.
Credentials
The skill does not require any secrets by default. It supports optional COMFYUI_HOST/PORT/USER/PASS and timeout/poll env vars which are reasonable. However the skill metadata and registry list required binaries as node and curl; the included Node script does not call curl, so declaring curl as required is unnecessary and disproportionate. Also the default COMFYUI_HOST is a specific private IP — verify that this default is intended for your environment.
Persistence & Privilege
The skill is not always-enabled and does not request elevated privileges or modify other skills or system-wide settings. It executes as a CLI invoked by the agent and only makes HTTP calls to the configured host/port.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install comfyui-request - After installation, invoke the skill by name or use
/comfyui-request - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release of comfyui-request.
- Send workflow requests to a running ComfyUI instance.
- Receive generated image URL or base64 data in the response.
- Supports host, port, and optional basic authentication configuration via environment variables or .env file.
- Communicates with the ComfyUI server via the /run endpoint.
Metadata
Frequently Asked Questions
What is Comfyui anfrage?
Send a workflow request to ComfyUI and return image results. It is an AI Agent Skill for Claude Code / OpenClaw, with 2951 downloads so far.
How do I install Comfyui anfrage?
Run "/install comfyui-request" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Comfyui anfrage free?
Yes, Comfyui anfrage is completely free (open-source). You can download, install and use it at no cost.
Which platforms does Comfyui anfrage support?
Comfyui anfrage is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Comfyui anfrage?
It is built and maintained by Xtopher86 (@xtopher86); the current version is v1.0.0.
More Skills