skills.rest
allcolor
allcolorCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

operate-comfyui

Install, operate, and troubleshoot self-hosted ComfyUI through PawFlow relays for media generation.

26|3|Updated Apr 1, 2026
Search Tags:#workflow automation#image generation#video generation#relay#model management#comfyui#pawflow
One-click install
npx skills add https://github.com/allcolor/PawFlow-Agents --skill operate-comfyui-allcolor
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: operate-comfyui
Source: https://github.com/allcolor/PawFlow-Agents/tree/main/packages/pawflow.comfyui-operator.pfpdir/content/skills/operate-comfyui
Command: npx skills add https://github.com/allcolor/PawFlow-Agents --skill operate-comfyui-allcolor

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Running self-hosted ComfyUI for image, video, and audio generation involves complex installation, model management, API workflow construction, and result validation that is error-prone when done manually across remote hosts. ## Core Features & Use Cases - Installation and Repair: Automates ComfyUI Desktop or portable setup on relay hosts, including PyTorch environment configuration, loopback binding, and smoke testing. - Model and Custom Node Management: Downloads models with checksum verification, installs reviewed custom nodes with pinned revisions, and validates discovery through /object_info. - Workflow Generation and QA: Builds and submits versioned API graphs for image, video, and audio generation, then validates outputs with ffprobe, frame extraction, and audio analysis before FileStore delivery. - Use Case: A user asks to generate a 10-second video from a reference image with voice continuity from a prior clip. The skill resolves the relay, ensures ComfyUI is ready, submits the validated workflow, monitors the job, QA-checks frames and audio, and delivers the result to FileStore. ## Quick Start Install ComfyUI on my relay host and generate an image from this prompt using the configured video generation service.

Frequently Asked Questions about operate-comfyui

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I install ComfyUI on a remote host automatically?▼

Use the ensure-ready flow to probe the target, then run automatic installation which clones the official repository, creates a virtual environment, installs the hardware-specific PyTorch build, and starts ComfyUI on 127.0.0.1:8188 with durable logs. Verify with /system_stats before production use.

How to submit a ComfyUI API workflow for video generation?▼

Export the graph using File > Export Workflow (API), never POST a UI workflow to /prompt. Validate node IDs, class types, and inputs against /object_info, patch only intended values, then POST to /prompt and persist the returned prompt_id with a submission manifest.

What is the difference between local=true and relay_local=true in PawFlow?▼

Tool local=true executes the action on the relay host itself, while service relay_local=true makes ComfyUI HTTP requests originate from the relay host helper. Prefer base_url=relay://RELAY_NAME/localhost:8188 with relay_local=true for host-local ComfyUI.

Can ComfyUI generate video with voice continuity from a prior clip?▼

Yes, load the prior video and split it with a node like GetVideoComponents, routing video and audio to the model's reference inputs. Use only voices the user owns or is authorized to use, and human-listen when exact wording or speaker identity matters.

Why does ComfyUI prompt validation fail after submission?▼

Validation fails when graph inputs do not match the installed node classes. Compare the graph against /object_info, confirm required models and custom nodes are discoverable, and check /history for the specific node error before resubmitting.

Is it safe to expose ComfyUI port 8188 to the internet?▼

No, ComfyUI must stay bound to loopback and never expose port 8188 publicly. Access it through the PawFlow relay route instead, and treat workflow JSON and custom nodes as executable code requiring source review.

Related Skills

Explore More Software Engineering Skills→
PrimotionStudioPrimotionStudio

prisma-postgres-setup

Provisions a Prisma Postgres database via the Management API and connects it to a local project.

Community
Advanced
artokunartokun
530

director

Orchestrate a full short film production pipeline from a text story into a final video.

Community
Advanced
cdalsoniiicdalsoniii

Dependency Health Check

Analyze manifest files for CVE exposure, version freshness, maintainer risk, and license compatibility.

Community
Advanced
FutureAtomsFutureAtoms
1

personaplex-deploy-hybrid

Design hybrid AI deployment strategies integrating self-hosted and cloud GPU infrastructure.

Community
Advanced