video-analysis — Claude Code skill: what it does, how to install
Inspects local video files using direct model ingestion for small clips or ffmpeg frame extraction and Whisper audio transcription for larger media.
npx skills add Starchild-ai-agent/official-skills --skill video-analysis -a claude-codeWhat video-analysis does
The tool processes video files depending on their size threshold, which defaults to 20MB. Smaller clips go directly to a model that handles video input natively, with google/gemini-3.1-flash-lite selected as the default. The model reviews both the image sequence and the audio stream together.
When a file exceeds 20MB, the workflow switches to extraction mode. It runs ffmpeg to capture keyframes: sampling every 2 seconds for clips under 60 seconds or picking scene-change frames on longer footage, capped at 30 images. Whisper transcribes the spoken dialogue, and the agent receives the resulting stills alongside the transcript context.
The skill operates on local workspace file paths. Direct dotted Python imports fail with a syntax error due to the hyphen in the directory name, so scripts must invoke it from within the skill directory or via dynamic loaders.
What it helps with
- summarize visual events and actions inside a local video file
- extract spoken dialogue into text using Whisper
- split long video recordings into keyframes via scene detection
- read video dimensions, runtime, and audio presence without a full review
Example prompts for the agent
- What happens in this video
output/videos/clip.mp4? - Summarize the recording at
output/videos/demo.mp4and extract the dialogue - Check the technical metadata for
media/interview.mp4
How to install video-analysis
1Skills CLI (Vercel)
npx skills add Starchild-ai-agent/official-skills --skill video-analysis -a claude-code2Manually: put the files in place
~/.claude/skills/video-analysis/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/video-analysis/
1Skills CLI (Vercel)
npx skills add Starchild-ai-agent/official-skills --skill video-analysis -a codex2Manually: put the files in place
.agents/skills/video-analysis/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add Starchild-ai-agent/official-skills --skill video-analysis -a cursor2Manually: put the files in place
.cursor/skills/video-analysis/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install Starchild-ai-agent/official-skills video-analysis2Manually: put the files in place
.github/skills/video-analysis/SKILL.md1Skills CLI (Vercel)
npx skills add Starchild-ai-agent/official-skills --skill video-analysis -a gemini-cli2Platform CLI
gemini skills install https://github.com/Starchild-ai-agent/official-skills --path video-analysisManually: put the files in place
.devin/skills/video-analysis/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add Starchild-ai-agent/official-skills --skill video-analysis -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about video-analysis
How to install video-analysis in Claude Code?
Run npx skills add Starchild-ai-agent/official-skills --skill video-analysis -a claude-code in your terminal.
What does the video-analysis skill do?
It examines video files by feeding small media directly to multimodal models and breaking down large files into keyframe stills and Whisper transcripts.
Can video-analysis process direct video URLs?
No, the execution script expects a local file path. For web video links, instructions advise fetching cloud captions or text transcripts rather than downloading media.
