Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add allcolor/PawFlow-Agents --skill operate-comfyuigit clone --depth 1 https://github.com/allcolor/PawFlow-AgentsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/allcolor/pawflow-agents/operate-comfyui)<a href="https://agentmods.dev/skills/allcolor/pawflow-agents/operate-comfyui"><img src="https://agentmods.dev/badge/skills/allcolor/pawflow-agents/operate-comfyui/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/allcolor/pawflow-agents/operate-comfyui"><img src="https://agentmods.dev/badge/skills/allcolor/pawflow-agents/operate-comfyui.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00087 | $0.04526 |
| Opus 5 | $0.00044 | $0.02263 |
| Sonnet 5 | $0.00017 | $0.00905 |
| Haiku 4.5 | $0.00009 | $0.00453 |
Grade A, and why
operate-comfyui scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 382 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Operate ComfyUI through PawFlow
Use this skill for every self-hosted ComfyUI task: installation, relay setup, models, custom nodes, API workflows, reference-driven image/video/audio generation, authorized voice continuity, monitoring, QA, troubleshooting, and delivery.
Package orchestration contract
This skill is the intelligent entry point for the package. Keep user-facing requests simple: infer the operation, bootstrap missing preferences, select the versioned flow, and translate the final result. Users do not need to know ComfyUI nodes, graph formats, model folders, relay routing, or PawFlow flow IDs.
The package provides these flow templates:
pawflow.comfyui.ensure-ready:1.0.0probes the selected target and durably asks for automatic, manual, or cancelled recovery when it is not ready.pawflow.comfyui.provision-assets:1.0.0validates a model, LoRA, or custom node plan and durably waits for explicit approval.pawflow.comfyui.generate-video:1.0.0normalizes and validates a video request, gates estimated partner cost, calls the active video service, and checks that a result reference exists.pawflow.comfyui.validate-video:1.0.0is the reusable structural validation stage. It does not replace the technical and visual QA required below.
Flows containing durableWait must be deployed, started, and invoked through
manage_flow.invoke; never run them with the synchronous batch action. Reuse
an existing compatible running instance when possible. Persist its instance ID
and the invocation/job correlation in working state. If it parks for a user
answer or a long job, schedule a passive continuation and end the turn; never
poll in a loop.
First-use bootstrap
Run this idempotently whenever the skill is first used in a conversation:
- Resolve the relay in this order: agent binding, conversation binding,
user variable
comfyui.default_relay, then ask the user. - Verify that the relay exists and is connected before storing it.
- Read
comfyui.config_versionand the target entry incomfyui.targets. Ask only for values that are missing for the requested operation. - Store durable preferences with
manage_variableat user scope. Store conversation-only choices at conversation scope. Never put secrets in variables. - Deploy the selected flow on that relay, start it, and invoke its
inputport with one UTF-8 JSON object.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 10d ago First seen · 382 lines · 87 tokens per session scan A 9ed74384904f
operate-comfyui is a skill published in the GitHub repository allcolor/PawFlow-Agents (25 stars, last pushed 2d ago), licensed MIT. It adds 87 tokens to every session and 4,526 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
illustration-style
Define an illustration style guide — visual language, colour usage, and application rules. Use when commissioning or standardising illustration. For icons, use icon-system (design-systems).
ai-provider-elevenlabs
ElevenLabs voice AI SDK patterns for TypeScript/Node.js -- text-to-speech, streaming, voice cloning, speech-to-speech, pronunciation control, and conversational AI.
scroll-video
Build the foundation of an Apple-style scroll-driven video website — extract frames from a video, render them on canvas, and bind playback to scroll progress with a circle-wipe hero reveal.
make-a-deck
Use when building 16:9 presentation slide decks, visual storyboards, or pitch decks.
animated-video
Use when creating animated product launch videos, GSAP timelines, or visual motion assets.
motion-animations
Use when building website animations, scroll effects, entrance reveals, layout transitions, gesture interactions, parallax, hero choreography, or any animated UI component. Auto-trigger on any website build or frontend feature that involves motion, transitions, or animated elements.