Watch
Analyze the contents of a video. Use whenever the user provides a YouTube/Vimeo/TikTok/X/Instagram URL (including Reels) or local video file (.mp4/.mov/.mkv/.webm) and asks a question about what's in it — what was said, what was shown, who appears, when something happens, etc. Pulls the transcript from platform captions when available, otherwise transcribes locally with whisper.cpp, and extracts frames for visual context.
Install
npx skills add khou/watch-video --skill watchgit clone --depth 1 https://github.com/khou/watch-video.git /tmp/watch-video
mkdir -p ~/.claude/skills
cp -r /tmp/watch-video/. ~/.claude/skills/watchPersonal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.
From the SKILL.md
# Watch a video
This skill prepares a video for analysis. It pulls a transcript and extracts a small number of frames for visual context. **No remote transcription API is used.** Captions come straight from the platform when available; videos without captions are transcribed locally with whisper.cpp (~140MB model, runs on CPU). If both fail, the script falls back to vision-only analysis from frames.
## When to use
Trigger when the user shares a video URL or file path and asks anything about its contents. Examples:
- "Summarize this YouTube video: <url>"
- "What does the speaker say about <topic> in this video?"
- "At what timestamp does <X> happen?"
- "Describe what's on screen at 2:30"
- "Who's in this clip?"
Do NOT trigger for non-video URLs, image-only files, or audio files.
## How to run
The script lives at `scripts/watch.py` inside this skill's directory. Resolve it to an absolute path (the directory containing this `SKILL.md`) before invoking. After running `install.sh`, the canonical path is `~/.claude/skills/watch/scripts/watch.py`.
```bash
python3 <skill-dir>/scripts/watch.py "<source>" [options]
```
Options (all optional):
- `--mode {fast,balanced,accurate}` — token vs. accuracy preset. Default `balanced`. See the table below.
- `--start <time>` and `--end <time>` — trim to a range. Use when the user asks about a specific segment ("what happens in the last 2 minutes").
- `--max-frames N` — override the mode's frame cap.Details
- Skill name
watch- Repository
- khou/watch-video
- License
- MIT
- Repo updated
- 3 May 2026
FAQ
What does the Watch skill do?
Analyze the contents of a video. Use whenever the user provides a YouTube/Vimeo/TikTok/X/Instagram URL (including Reels) or local video file (.mp4/.mov/.mkv/.webm) and asks a question about what's in it — what was said, what was shown, who appears, when something happens, etc. Pulls the transcript from platform captions when available, otherwise transcribes locally with whisper.cpp, and extracts frames for visual context.
How do I install the Watch skill?
skills CLI (Claude Code, Cursor, Codex and more): npx skills add khou/watch-video --skill watch. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.
Is Watch popular?
It lives in khou/watch-video, which has 6 GitHub stars and 1 forks.
Related skills
Suno Song Creator
5jayweiler/suno-song-creator
Co-creative songwriting and Suno music generation skill. Use this skill whenever the user wants to: write a song, create lyrics, make music, generate a Suno prompt, work on a track, express feelings through music, write something for someone, turn an experience into a song, create a style prompt, remix or rework lyrics, or mentions 'Suno', 'song', 'lyrics', 'track', 'chorus', 'verse', 'hook', 'melody', 'music', 'anthem', 'ballad', 'jam', or any music-creation-related term. Also trigger when the user describes an emotion, experience, or theme and seems to want creative musical output, even if t
Pixeltamer
6gabelul/pixeltamer-gpt-image-skill
Generate, edit, and compose images with GPT Image 2.5, including transparent PNGs with a real alpha channel. Two backends — OpenAI API key or codex CLI (uses your ChatGPT subscription, no key needed). Four modes — one-shot generate, edit or inpaint an existing image, multi-reference composition (up to 16 inputs blended into one), and multi-image batch with verification. Use when the user asks to make an image, generate a poster, draw a mockup, design an icon or sticker, create an ad creative, build an infographic, blend reference images, edit or inpaint an existing image, or produce a cutout a
Moltdj
4polaroteam/moltdj-skill
SoundCloud for AI bots. Generate tracks and podcasts, share on Moltbook, and earn from tips + royalties.
Motion Film
6tonyprots/motion-film-skill
Voice-led motion-graphics film (30 s – 4 min) built in code: story and script first, then an AI voice or own recordings, scenes on springs along the voice grid, real interface, captions, a score, independent critic rounds, 16:9 + 9:16. Use when asked to "make a video / film / explainer", for a promo, a product or research film, a video summary, a vertical Reels/Shorts film, turning a report or deck into a video. Also graphics over a finished recording and a 4–14 s loop without a voice (GIF, MP4). Not for screen recordings or music clips.
3dprint Scout
5madrobotnet/3dprint-scout
Designs 3D-printable parts reference-first: finds existing models, then builds in Fusion, OpenSCAD, Blender or Python, verifies the mesh, saves STL/3MF. Use to 3D print an object. Not for STL edits or resin.
Geometry Dsl
6shand001/geometry-dsl
Draw math geometry diagrams as PNG or SVG images using the Geometry DSL compiler. Use when the user asks to draw, shade, label, illustrate, or visually explain geometry, recreate a reference diagram, produce or repair Geometry DSL, or diagnose a diagram that looks wrong. Write DSL source, validate it with the bundled compiler, render it to an image, and visually inspect complex diagrams before delivering.
Browse the most popular Claude skills, all images, video & audio skills, or write your own with the skills generator.