GenMedia Skills - Setup Guide¶
Source: genmedia-labs/skills (30 skills · 42.5K combined installs)
Repo: github.com/genmedia-labs/skills
Category: AI Media / Video & Image Generation
First Seen: August 13, 2026
Quality Tier: 🟢 Production (video-edit 8.7K, ai-music 8.6K, ai-video-generation 8.5K installs)
GenMedia is the fastest-moving media cluster on skills.sh right now - ai-video-generation gained 372 installs in a single hour during the August 13 sweep. Five core skills cover the editing and generation workflow, while 25 model-specific skills pin the exact generation backends: FLUX 2 Klein, Kling 3.0, GPT Image 2, Seedance V2, Wan 2.7, Nano Banana 2, and more. One install command gives an agent the whole modern image/video generation surface.
Installation¶
# Full cluster
npx skills add genmedia-labs/skills
# Hermes: install individual skills by identifier
hermes skills install genmedia-labs/skills/video-edit
hermes skills install genmedia-labs/skills/ai-video-generation
If the CLI rejects the multi-skill source, clone genmedia-labs/skills and copy the skill directories into the agent's skills folder.
Core Skills¶
| Skill | Installs | Use For |
|---|---|---|
video-edit |
8.7K | Editing and assembling video from agent pipelines |
ai-music |
8.6K | Music generation and audio scoring |
ai-video-generation |
8.5K | Text-to-video and image-to-video orchestration |
ai-image-generation |
8.4K | Image generation across multiple backends |
image-to-video |
8.4K | Animating stills into motion clips |
lipsync |
1 | Lip-sync for avatar and character video |
face-swap |
1 | Face replacement in video |
relight |
1 | Scene relighting for images and video |
video-extend / video-inpainting / video-outpainting |
1 | Video extension, repair, and canvas expansion |
Model pins: flux-2-klein, flux-kontext, kling-3-0, gpt-image-2, gpt-image-edit, seedance-v2, wan-2-7, nano-banana-2, nano-banana-edit, happyhorse-1-0, codex-pet |
1 | Backend-specific generation recipes |
ai-avatar-video, image-edit, image-inpainting, image-outpainting, controlnet-pose, elevenlabs-music-generation, runcomfy-cli, ace-step |
1 | Avatar video, image editing, music via ElevenLabs, RunComfy integration |
Prerequisites¶
| Requirement | Details |
|---|---|
| Node.js + npx | For the skills.sh CLI install path |
| Generation API keys | Per-model keys (FLUX, Kling, OpenAI, etc.) only for the backends you use |
| Any agent runtime | Skills are procedural - no runtime-specific code |
CorpusIQ Use Cases¶
| Use Case | How |
|---|---|
| UGC video pipeline | ai-video-generation + image-to-video for product motion clips |
| Voiceover pairing | ai-music for background scoring, lipsync for avatar narration |
| Ad creative variants | flux-2-klein + gpt-image-2 recipes for rapid asset iteration |
| Video repair | video-inpainting / video-extend to fix or lengthen existing clips |
Integration with CorpusIQ¶
CorpusIQ is the connector layer that lets AI assistants read the systems a business already runs on - accounting, payments, CRM, email, files and analytics - and answer with the source named. It is read-only by design: it retrieves data and never writes to, or acts in, a connected system, which is what makes it safe to point at live accounts.
This page is part of the CorpusIQ directory, the reference we maintain for the MCP server and agent-skill ecosystem. Run alongside CorpusIQ, a third-party server like this one can be combined with a company's own business data in a single answer.
Limitations / Verification¶
- Model skills at 1 install each are new - treat as recipes, verify against provider docs
- API costs accrue per backend; only pin models you have keys for
- Verify install:
npx skills list | grep -E 'video|image|music'