Skip to main content
skill CX Codex

watch · diffusionstudio — Claude Code skill: read video and audio footage

Answers questions about a video or audio file, summarizes it, locates scenes and pulls quotes with timestamps.

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add diffusionstudio/skills --skill watch -a claude-code
Open on GitHub

What watch does

The watch skill wires your agent into Diffusion Studio so it can work through footage: find a moment from a description, recap a recording, pull lines, check a claim about what a file contains. Video and audio both.

The guidance ships inside the Diffusion Studio app, so it matches the version you have installed. Docs sit in the app bundle — Diffusion Studio.app/Contents/Resources/docs on macOS, %LOCALAPPDATA%\DiffusionStudio\app-<version>\resources\docs on Windows. Start at skills/watch.md and follow it for the session; it links to the tool reference and prompt guides in the same folder.

Tools run through the diffusion CLI (alias dapi). Doc names are MCP tool names, and each maps to a command: grab is diffusion grab, while diffusion <command> --help lists the options. Connected MCP tools behave the same way. Keep the app in the background with diffusion open --background and surface the result with diffusion window show. If diffusion or the app is missing, the skill points you at references/installation.md — worth checking before you start.

What it helps with

  • find a scene by description in a long recording
  • summarize a webinar or interview
  • pull quotes with their timestamps
  • verify what a file actually says
  • describe what happens in footage and when

Example prompts for the agent

  • Watch this video and find the moment where the settings panel appears
  • Give me a short recap of the meeting recording
  • Pull every line about the budget with timestamps

How to install watch

1Skills CLI (Vercel)

npx skills add diffusionstudio/skills --skill watch -a claude-code

2Manually: put the files in place

~/.claude/skills/watch/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/watch/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about watch

How do I install watch in Claude?

Run npx skills add diffusionstudio/skills --skill watch -a claude-code. The skill file then lands at ~/.claude/skills/watch/SKILL.md.

What does the watch skill do?

It works through video and audio: answers questions about the content, summarizes it, finds scenes and moments, pulls quotes, and describes what happens and when.

What do I need for it to work?

The Diffusion Studio app and the diffusion CLI (alias dapi), since tools are called through it. If either is missing, the skill points to references/installation.md.

Source text

SKILL.md · MIT
---
name: watch
description: >-
  Watch and understand footage with Diffusion Studio: answer questions about a
  video or audio file, summarize it, find scenes and moments, pull quotes, and
  describe what happens and when. Use whenever the user asks what's in a piece
  of footage, wants a summary or recap, wants to locate a moment ("where does X
  happen", "find the scene where..."), or needs a claim about a video or audio
  file checked.
---

The guidance for this skill ships with the Diffusion Studio app, so it always matches the installed version. Read it from there and trust it over memory; it belongs to the app, so never edit it.

The docs live inside the app bundle at `Diffusion Studio.app/Contents/Resources/docs` (usually under `/Applications`); on Windows, at `%LOCALAPPDATA%\DiffusionStudio\app-<version>\resources\docs`. Start with `skills/watch.md` and follow it for the rest of the session. It links to the tool reference and prompt guides in the same folder.

Use the tools through the `diffusion` CLI (alias `dapi`). The docs use MCP tool names, and each one is a command: `grab` is `diffusion grab`, and `diffusion <command> --help` lists its options. If the MCP tools are connected, they work the same.

Work with the app in the background (`diffusion open --background`), then show the result with `diffusion window show`.

If `diffusion` or the app is missing, read [installation.md](references/installation.md).

© diffusionstudio · MIT · diffusionstudio/skills

Similar skills

Everything for AI and ML →

watch by other authors comes first.

watch bradautomates Analyzes a video from a URL or local path, collects frames and a transcript, then lets the agent answer questions about the footage. Useful for video review inside an IDE. CX Codex +1 17,200installs watch oxbshw watch examines linked or local videos, extracts frames and speech, then answers questions from an index. Use it to review a recording or check your own interface. CX Codex +1 583installs videoagent-audio-studio wells Routes text-to-speech, music creation, sound effects, and voice cloning to ElevenLabs and fal.ai. Returns ready audio URLs straight inside your coding agent. CX Codex +1 3,900installs video-to-sfx sonilo-ai This skill generates sound effects matched to what is on screen: footsteps, impacts, ambience, foley. Useful when you have footage that still needs sound design. CX Codex +1 1,400installs ai-multimodal mrgoonie A skill that runs audio, video, image and PDF tasks through the Gemini API. Useful when you need a transcript, a screenshot read, or tables pulled out of a document. CX Codex +1 530installs media-use heygen-com Finds or generates media for HyperFrames videos, including voiceovers, music, cutouts, and color grades. Stores files locally in assets. CX Codex +1 1,100,000installs Video Processor disler A skill that handles video files from the agent: pull the audio track, convert to mp4 or webm, transcribe speech with Whisper. Useful when video files show up in a project. CX Codex +1 ★ 1,554repo stars Transcription Automation Claude Office Skills A skill for automatic audio and video transcription: meeting notes, subtitles, podcast show notes. Runs through the media-mcp server. CX Codex +1 4,800installs

From the same set2

Project sponsors

Companies that keep this analytics open
Gold sponsor

Impressions, clicks and code copies — open statistics. We keep the numbers public for the same reason NeBlask itself is public. Become a sponsor