explaingit

0xzgbot/hermes-comfyui-skills

★ 0Audience · designerComplexity · 1/5ActiveLicenseSetup · easy

TLDR

Three Hermes Agent skill bundles that load multiple ComfyUI image and video model skills with a single slash command instead of typing each one.

Mindmap

mindmap
  root((hermes-comfyui-skills))
    Inputs
      Hermes CLI
      Installed ComfyUI skills
    Outputs
      Slash command bundles
      Grouped model presets
    Use Cases
      Load all ComfyUI image skills at once
      Load video and animation skills together
      Run a full prompt session with one command
    Tech Stack
      Hermes
      YAML
      ComfyUI

Things people build with this

USE CASE 1

Load all ComfyUI image generation skills in Hermes with one slash command

USE CASE 2

Set up a video and animation prompting session for Wan, LTX, and Seedance at once

USE CASE 3

Group Flux 2, SDXL, and LTX prompt skills under a single namespace

Tech stack

HermesYAMLComfyUI

Getting it running

Difficulty · easy Time to first run · 5min

Bundles only wrap other skills, so the underlying ComfyUI and model skills must already be installed in Hermes.

MIT: do almost anything with the code as long as you keep the copyright notice.

In plain English

This repository is a small collection of preconfigured skill bundles for the Hermes Agent, aimed specifically at people using ComfyUI to generate images and video. ComfyUI is a node based interface for running open source image and video models like Flux, SDXL, LTX, Seedance, Wan, and AnimateDiff. Hermes is the agent framework these bundles plug into. A skill bundle, as the README explains it, is a Hermes YAML file that groups several individual skills together under one slash command. Without a bundle, you would have to type something like /comfyui /flux2 /ltx23 /quality_tokens one after another to load everything you need for a session. With a bundle installed, a single slash command pulls in the whole set. There are three bundles in the repo. /comfyui-prompting is the all in one set, loading nine skills covering Flux.2, two versions of LTX, Seedance 2, SDXL, Wan, plus general prompt engineering, quality token vocabularies, and iterative refinement advice. /comfyui-image is the narrower image only set, with four skills covering still image generation, img2img, ControlNet, inpainting, and upscaling. /comfyui-video focuses on video and animation with ten skills, adding camera motion vocabulary, character consistency tips, and beat based scripting on top of the underlying model skills. Installation is two commands at the Hermes command line. You tap the repository as a skill source and then either list and pick bundles or install one by name. The bundles only act as wrappers, so they assume the underlying skills are already installed on your system, things like the builtin comfyui skill, local workflow skills for Flux 2 and LTX, and various mlops helpers. If some of those underlying skills are missing, the bundle will still load the ones it can find and note the gaps. The project is MIT licensed.

Copy-paste prompts

Prompt 1
Walk me through installing hermes-comfyui-skills as a Hermes skill source
Prompt 2
Show the YAML for the /comfyui-image bundle and explain each skill it pulls in
Prompt 3
Write a new Hermes bundle that combines Flux 2 and SDXL prompting skills only
Prompt 4
List which underlying ComfyUI skills I need installed before the /comfyui-prompting bundle works
Open on GitHub → Explain another repo

Generated 2026-05-22 · Model: sonnet-4-6 · Verify against the repo before relying on details.