Skip to main content
This guide explains how to install and run the skills.video skills for video and image generation, including first-time API key setup and what to do when credits are insufficient.

​
Why use these skills

Use these skills to standardize how your team calls skills.video endpoints. They enforce an SSE-first flow, provide polling fallback guidance, and include helper scripts for key checks and runtime error handling.

​
Install

Install the skills package from GitHub: 
npx skills add chuyun/ai-skills
This package currently includes:
  • ai-video-skills
  • ai-image-skills

​
Supported clients

These skills are not limited to Codex. Use them in any client that supports the Skills package format, including:
  • Codex
  • Claude Code
  • Other compatible Skills clients

​
First-time API key setup

The skills require SKILLS_VIDEO_API_KEY before any API call. If the key is missing, they should prompt setup instructions.

​
Get an API key

  1. Sign in at https://skills.video/dashboard/developer.
  2. Click Create API Key.
  3. Copy the key.

​
Set environment variable

export SKILLS_VIDEO_API_KEY="<YOUR_API_KEY>"
To persist in zsh: 
echo 'export SKILLS_VIDEO_API_KEY="<YOUR_API_KEY>"' >> ~/.zshrc
source ~/.zshrc

​
Use the skills

Call the skill directly in your client prompt. 
Use $ai-video-skills to create a Kling v2.6 generation request.
Use $ai-image-skills to create a Nano Banana Pro generation request.
Typical flow inside each skill:
  1. Check API key with scripts/ensure_api_key.py.
  2. Resolve endpoint and payload from OpenAPI.
  3. Use SSE endpoint as default.
  4. Fall back to polling if SSE is unavailable.
From the installed skill directory (path depends on your client), run the first-use key check: 
cd "<SKILLS_INSTALL_PATH>/ai-video-skills"
python scripts/ensure_api_key.py

​
Handle insufficient credits

When a generation call fails due to credits, the skill should stop retries and guide recharge.

​
Recharge path

  • Open https://skills.video/dashboard.
  • Go to Billing/Credits.
  • Recharge, then retry the request.

​
Check current credits

curl -X GET "https://open.skills.video/api/v1/credits" \
  -H "Authorization: Bearer $SKILLS_VIDEO_API_KEY"

​
Classify runtime error (example)

Run this from the installed skill directory. 
python scripts/handle_runtime_error.py \
  --status 402 \
  --body '{"message":"Insufficient credits"}'