feat: add video-analyzer skill - MP4/video to AI-readable report

[VjAlbert]

Name:

video-analyzer

Short description (tweet-length):

Converts any video file into a full AI-readable report: transcript, character detection, visual timeline, and JSON output — because Claude Projects don t accept MP4s.

Full description:

## video-analyzer — MP4 to AI report

Upload any video and get back a complete structured report Claude can actually use.

### What you get
- Full audio transcript with timestamps (Whisper, 99 languages)
- Automatic character/scene detection (no ML model required)
- Visual timeline — frame-by-frame description via Claude Vision
- Dual output: human-readable Markdown + machine-readable JSON

### Why this exists
Claude Projects don t support MP4 uploads. This skill bridges the gap:
process the video once, get a report you can paste into any project as a document.

### Works on
- Live-action video
- AI-generated animation (SHOWRUNNER, Sora, etc.)
- Screen recordings and tutorials
- Vertical Reels/Shorts format

### Usage
Just upload a video and write: analizza questo video or analyze this video

### Dependencies
Auto-installed on first run: ffmpeg (system), openai-whisper, opencv-python-headless

Tags:

video, mp4, transcription, whisper, ffmpeg, analysis, report, json, character-detection

Category: Media & Video

License: Apache 2.0