pexo-skills

Open-source Agent Skills for AI content creation — pick the level of control you need.

A collection of independent skills for AI content creation, organized by the problem they solve. Use one, use all, or combine them however you want. Each skill works on its own — no dependencies between them.

Compatible with OpenClaw, Claude Code, Cursor, and any agent supporting the Agent Skills standard.

Skills

🧠 Agent — Your AI Creative Partner

For when you want a finished result and don't want to think about how it gets made. Describe what you're imagining — Pexo thinks with you, picks the best AI models, and delivers a polished, ready-to-share piece.

Skill	Description
pexo-agent	AI creative partner that turns conversation into complete, publish-ready content. No prompts to write, no models to choose, no editing skills required. Just describe your idea.

✍️ Creative — Story & Prompt Craft

For when you want to shape the creative direction before anything gets generated. Plan the narrative, design the storyboard, engineer the prompts.

Skill	Description
videoagent-director	AI creative director. Takes an idea, breaks it into a shot-by-shot storyboard, writes all prompts internally, and executes production. You approve the plan; it handles the rest.
seedance-2.0-prompter	Expert prompt engineering for Seedance 2.0. Transforms rough ideas and multimodal assets into structured, optimized generation prompts.

🔧 Studios — Model-Level Generation

For when you want direct control over a specific generation step — choose the model, craft the prompt, set the parameters.

Skill	Description	Models
videoagent-image-studio	Generate images — photos, illustrations, logos, artwork. One command, 8 models.	Midjourney · Flux · Ideogram · Recraft · Gemini
videoagent-video-studio	Generate video clips — text-to-video, image-to-video, reference-based. 7 models, auto-routing or manual selection.	Kling · Veo · Grok · Seedance · MiniMax · Hunyuan · PixVerse
videoagent-audio-studio	Generate audio — TTS, music, sound effects, voice cloning. One skill for every audio need.	ElevenLabs · CassetteAI

Pick What Fits

You need...	Use
A finished piece from a conversation	`pexo-agent`
A storyboard and directed production	`videoagent-director`
An optimized Seedance 2.0 prompt	`seedance-2.0-prompter`
A single image	`videoagent-image-studio`
A single video clip	`videoagent-video-studio`
Speech, music, or sound effects	`videoagent-audio-studio`

Every skill solves a different problem. None requires another to work.

Quick Install

OpenClaw / npx

npx playbooks add skill pexoai/pexo-skills

Claude Code

/plugin marketplace add pexoai/pexo-skills

ClaWHub

clawhub install pexoai/pexo-agent
clawhub install pexoai/videoagent-director
clawhub install pexoai/seedance-2.0-prompter
clawhub install pexoai/videoagent-image-studio
clawhub install pexoai/videoagent-video-studio
clawhub install pexoai/videoagent-audio-studio

Cursor

Copy any skill folder into ~/.cursor/skills/.

Key Properties

Minimal setup — One API key unlocks every capability. No juggling credentials across providers.
Multi-model — Studios integrate multiple providers and pick the best model for the job, or let you choose explicitly.
Independent — Every skill works standalone. Use one or combine several — no coupling, no orchestration required.
Open source — Every skill, proxy, and tool is inspectable. Deploy your own infrastructure if you need full control.

Skill Structure

Each skill follows the Agent Skills Specification:

skill-name/
├── SKILL.md          # Agent instructions and metadata
├── tools/            # Executable scripts the agent calls
├── proxy/            # Serverless API proxy (deploy on Vercel)
└── references/       # Knowledge base — models, prompts, guides

About

Built and maintained by @pexoai.

Website: pexo.ai
Issues / Feedback: GitHub Issues
ClawHub: clawhub.ai/pexoai

License

MIT

videoagent-director