skills.rest
agentspace-so
agentspace-soOFFICIAL
Explore all skills by this author →
GitHub

controlnet-pose

Generate pose-conditioned images or videos using OpenPose, depth, or canny controls.

31|9|Updated Apr 30, 2026
Search Tags:#image generation#controlnet#motion control#depth map#runcomfy cli#openpose#pose transfer
One-click install
npx skills add https://github.com/agentspace-so/runcomfy-agent-skills --skill controlnet-pose-agentspace-so
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: controlnet-pose
Source: https://github.com/agentspace-so/runcomfy-agent-skills/tree/main/controlnet-pose
Command: npx skills add https://github.com/agentspace-so/runcomfy-agent-skills --skill controlnet-pose-agentspace-so

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Controlnet-pose helps you recreate consistent characters and compositions by conditioning image or video generation on a pose, skeleton, or motion reference instead of relying on guesswork from prompts alone.

Core Features & Use Cases

  • Video pose transfer (motion control): Transfers a reference performance’s motion and blocking onto a target character image for outputs like choreography re-shots and sports motion stylizations.
  • Image pose-conditioned generation (ControlNet via LoRA): Generates pose-locked images from a control image such as an OpenPose/DWPose skeleton, depth map, or canny edge while following your prompt.
  • Automatic routing by input type and intent: Picks the right RunComfy model route for video vs still and stylized vs photoreal needs, based on your request keywords and provided inputs.
  • Works with pose-style conditioning references: Handles common “controlnet/pose control/openpose/depth/canny” style intents and routes to the appropriate Model API endpoint.

Quick Start

Use the controlnet-pose skill to generate a pose-conditioned output by asking it to route to pose control and then running the RunComfy CLI with your reference video or control image URL plus your target character or prompt.

Frequently Asked Questions about controlnet-pose

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

FAQPage Schema
How do I transfer motion from a reference video to a character image?▼

To transfer motion from a reference video, you condition the generation on the reference's pose using the RunComfy CLI. You provide the reference video URL and target character image to route the request to the correct motion control model endpoint.

Can I use a depth map or OpenPose skeleton to control image generation?▼

Yes, you can use a depth map or OpenPose skeleton to generate pose-locked images. The system applies ControlNet conditioning via LoRA, following your prompt while maintaining the structural constraints of the provided control image.

What is pose-conditioned image generation and when do I need it?▼

Pose-conditioned image generation recreates consistent characters and compositions by using a pose, skeleton, or motion reference instead of relying on prompts alone. You need it when exact character positioning or motion transfer is required.

Does pose control routing work automatically based on my input type?▼

Yes, the system automatically routes your request based on input type and intent. It analyzes your request keywords and provided inputs to pick the right RunComfy model route for video versus still and stylized versus photoreal needs.

Do I need the RunComfy CLI to generate pose-locked outputs?▼

Yes, you need the RunComfy CLI to generate pose-locked outputs because the skill requires a CLI invocation that posts the correct JSON input schema. The environment allows only Bash commands with the runcomfy prefix and downloads outputs to a specified directory.

What control image formats work with pose transfer for image generation?▼

Pose transfer works with common control image formats including OpenPose/DWPose skeletons, depth maps, and canny edge images. These references allow the system to lock the structural pose while generating the target character.

Related Skills

Explore More Design & Creative Skills→
runcomfy-comruncomfy-com
12

controlnet-pose

Generates pose-conditioned images or video motion transfers using RunComfy CLI and JSON inputs.

Official
Intermediate
doany-aidoany-ai
5

controlnet-pose

Routes pose-conditioned generation tasks to RunComfy model APIs via CLI inputs and outputs to a user-specified directory.

Official
Intermediate
prime-skillsprime-skills
31

controlnet-pose

Route pose-conditioned image and video generation through RunComfy models.

Official
Advanced
camillanapolescamillanapoles

eachlabs-video-generation

Generate videos from text prompts or images via the EachLabs Predictions API.

Community
Advanced