GitHub radar
reelbench-skills: Video Skill Analysis for AI Agents
Wesley Wang’s Claude Code & Codex can auto-generate shot reports from videos, which is valuable for content factories to replicate success.
reelbench-skills: a collection of skills developed by Wesley Wang for Claude Code & Codex. The first skill parses a given video into a shot table with time, shot type, camera motion, scene composition, edit pace. This is measured by ffmpeg and all subjective interpretations are delegated to the model. The second skill visualizes a video and its corresponding shot table simultaneously. Shots will be highlighted during playback. This tool will generate an offline HTML report with a player, shot table search, and keyframe display. This tool does not require any npm package, and it does not require an API key. There are some Chinese descriptions in the repo description, please ask your agents to translate the English version of the README if necessary.
Why a vibe-coder should care
This is a great tool for a content factory to understand what works in the industry. Simply analyse a reel in your field with the shortest shot length and cut timing, and you can train your video editors/generators accordingly. All this is done on your existing agent subscription and is free! You can also build upon this idea to create customised reports for your niche (e.g. compare the last 10 videos you made vs. the top 10 for your topic).
How to install
Copy this and send it to your agent — Claude Code, Codex, any of them:
Install the video analysis skills from https://github.com/eternityspring/reelbench-skills — follow the README for Claude Code, then break my video down shot by shot and open the report for me.
A regular laptop: only Node.js and ffmpeg needed — the agent installs them for you.
Open on GitHub▌ More finds