- Home
- Skills
- Image & Video
- ComfyUI Agent Skill
Works with the AI Agents you already use
ComfyUI Agent Skill
Run registered ComfyUI workflows, or import your own reviewed workflow, for image, video, music, speech, matting, and voice cloning.
Free
ComfyUI Agent Skill
Example session with this skill installed
Generate an image with my local ComfyUI: a cute cat sitting on a windowsill at golden hour. Use the built-in text-to-image workflow. If the server is down, tell me whether I should start it locally or point you at another machine.
- Read your context and instructions
- Compiled the comfyui agent
Summary
The skill checked the local ComfyUI server, ran the registered workflow z_image_turbo, and saved one PNG. It did not upload a reference image and it did not run an unreviewed workflow file.
What it ran
comfyui-skill generate -p "a cute cat sitting on a windowsill at golden hour"
Result
{
"success": true,
"workflow": "z_image_turbo",
"outputs": [
{
"type": "image",
"path": "C:\\Users\\you\\AppData\\Roaming\\comfyui-skill\\results\\20261004\\204100_abc123\\output.png"
}
]
}
The image is the file in outputs[].path. On Windows that directory is under %APPDATA%\comfyui-skill\results. On macOS it is ~/Library/Application Support/comfyui-skill/results. On Linux it is $XDG_DATA_HOME/comfyui-skill/results or ~/.local/share/comfyui-skill/results.
Next steps
- Open that PNG.
- To cut a subject out of it, ask for
sam3_mat_imageand describe the object. - To turn it into a short video, ask for
ltx_23_i2v_distilledand give a motion prompt. - To use your own ComfyUI workflow, import the API-format file, review the generated config, then rename it to
workflow.config.json.
Connects securely to your tools. The creator never sees your data.
About this skill
This skill runs ComfyUI on your own machine or a trusted self-hosted server. It generates images, video, music, and speech, and it also does text-driven matting, face and expression transfer, and voice cloning.
It only runs workflows that are already registered, or a workflow you already validated in ComfyUI and then imported into your private registry. Import does not run the file. You review the generated config and rename it before it becomes active. It will not execute an unreviewed workflow JSON.
It can check that the server is up, preflight a workflow's nodes and models, and save a server URL when you ask. If the server is down, it says so and asks whether ComfyUI is local or on another machine.
Built-in workflows cover text-to-image, anime and illustration, artistic and product images, image edit, image-to-video, text-to-video, music, voice design, voice cloning from a sample, matting, and expression transfer. You bring the models and custom nodes. The skill does not host ComfyUI for you.
Does not host ComfyUI or download models. You need a running server plus the models and custom nodes for the workflow you pick. It will not run an unreviewed workflow JSON. UI-format workflows must be converted before import. Image size flags only work on workflows that map both width and height.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
6 uses
Times used to date
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 1 month ago
- Free with an Agensi account
Permissions
File Scopes
The skill runs a local CLI, reads your images, video, and audio, writes results under the per-user data folder, and calls your ComfyUI server over HTTP. It reads COMFYUI_URL when you set it. It does not use a browser.