agentleFS
Sign inStart free
VanKyle00VanKyle00/claude-skills/skills/youtube-learn/SKILL.md

youtube-learn

Skill0 starsChanged 5 months agoMITView on GitHub

Use when the user shares a YouTube link (youtu.be / youtube.com) and wants you to learn from, summarize, analyze, explain, fact-check, react to, or pull information out of the video's ACTUAL content — talks, tutorials, lectures, reviews, demos, stock/market or news videos. Triggers on "what can you learn from this video", "summarize this YouTube video", "watch this", "what does this video say", or a bare video URL paired with an analysis ask. The page itself only exposes the title; this skill captures scene-change + interval keyframes and the transcript so you can read what is shown and said.

  • Installs packages

2 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions2 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 2

    1. dotanminh/youtube-learn-skill★ 29same name
    2. carlosbuelvas1/openclaw-skills★ 0same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository2 other files it gives its agents.

The file

Create a free account to mount
Files
SKILL.md

youtube-learn

Overview

YouTube does not expose a video's transcript or audio to the fetch tool — scraping the page returns only the title. This skill turns a YouTube URL into things you can consume: keyframe images (so you can see slides, charts, demos, faces) plus a timestamped transcript (so you can read what was said), bundled into one manifest.md.

When to use

  • The user wants the substance of a specific YouTube video, not just its metadata.
  • Skip it for pure-metadata questions ("who posted this?"), or music videos where the audio isn't informational.

Run it

From this skill's directory:

python scripts/extract.py "<youtube_url>"

Useful flags (--help for all):

  • --interval 20 force a frame at least every N seconds (safety net for slow-changing slides)
  • --scene 0.4 scene-change sensitivity (lower = more frames)
  • --max-frames 40 cap total frames (evenly subsampled if exceeded)
  • --height 720 download resolution cap
  • --no-whisper skip the local transcription fallback

It writes to ~/Downloads/yt-learn/<video-id>/ and prints the path to manifest.md.

Then consume the output

  1. Read manifest.md — it interleaves each keyframe with the transcript spanning that moment.
  2. Read the referenced frames/*.jpg images — you are vision-capable, so actually open them.
  3. Answer from frames and transcript together. Cite [mm:ss] timestamps from the manifest.

Notes & troubleshooting

  • Transcript source appears in the manifest: youtube captions (default, instant) or local whisper (fallback when a video has none — first run downloads a ~140 MB model and needs faster-whisper, which the script tries to pip install automatically; on success it transcribes on CPU, which is slow for long videos).
  • Long videos: raise --interval and/or lower --max-frames to keep the frame set readable.
  • Download fails: age/region-restricted or members-only videos may be inaccessible to yt-dlp.
  • Requirements: yt-dlp and ffmpeg on PATH (or set the YT_DLP / FFMPEG env vars). No third-party Python packages are needed unless the Whisper fallback runs.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free