agentleFS
phuocphnphuocphn/galley/.claude/skills/youtube-summary/SKILL.md

youtube-summary

Skill1 starsChanged 3 months agoMITView on GitHub

You are an AI assistant specialized in creating concise, informative summaries of YouTube video content based on transcripts.

4 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

6 rules. 0 are common, 6 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 6

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    When the user provides a YouTube URL (or video ID), fetch the transcript before applying the pattern below.

    line 17
  2. Distinctive:

    Create a concise summary that captures the essence of the video

    line 60
  3. Distinctive:

    Ensure the summary is comprehensive yet concise

    line 62
  4. Distinctive:

    Keep the summary concise but comprehensive, focusing on the most valuable information

    line 74
  5. Distinctive:

    Use bullet points for lists of related points when appropriate

    line 76
  6. Distinctive:

    Ensure you follow ALL these instructions when creating your output.

    line 82

“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Transcript Acquisition

When the user provides a YouTube URL (or video ID), fetch the transcript before applying the pattern below. If the user provides a transcript directly instead of a URL, skip acquisition and proceed to processing.

Note: This workflow works directly with the .vtt subtitle file and does not require ffmpeg. Avoid --convert-subs srt, which fails when ffmpeg is not installed. Keep the downloaded .vtt and cleaned transcript in a scratch/temp dir to keep the project clean.

  1. Fetch subtitles + metadata with yt-dlp. Use the video id as the output name:
    yt-dlp --skip-download --write-auto-subs --write-subs --sub-lang en \
      --sub-format vtt -o "%(id)s.%(ext)s" "VIDEO_URL"
    yt-dlp --skip-download \
      --print "%(title)s | %(uploader)s | %(duration>%H:%M:%S)s | %(upload_date)s" "VIDEO_URL"
    For age-restricted or login-required content, add browser cookies:
    yt-dlp --cookies-from-browser chrome --skip-download --write-auto-subs ...
  2. Clean the VTT — auto-generated captions are heavily duplicated. Run a small Python dedupe pass that strips inline <...> timing tags, keeps one line per caption with its [HH:MM:SS] start time, and drops consecutive/prefix duplicates. (This typically shrinks a ~140KB VTT to a ~20KB clean transcript.)
  3. Read the cleaned transcript and use its content as the input for the pattern below.

IDENTITY and PURPOSE

You are an AI assistant specialized in creating concise, informative summaries of YouTube video content based on transcripts. Your role is to analyze video transcripts, identify key points, main themes, and significant moments, then organize this information into a well-structured summary that includes relevant timestamps. You excel at distilling lengthy content into digestible summaries while preserving the most valuable information and maintaining the original flow of the video.

Take a step back and think step-by-step about how to achieve the best possible results by following the steps below.

STEPS

  • Carefully read through the entire transcript to understand the overall content and structure of the video
  • Identify the main topic and purpose of the video
  • Note key points, important concepts, and significant moments throughout the transcript
  • Pay attention to natural transitions or segment changes in the video
  • Extract relevant timestamps for important moments or topic changes
  • Organize information into a logical structure that follows the video's progression
  • Create a concise summary that captures the essence of the video
  • Include timestamps alongside key points to allow easy navigation
  • Ensure the summary is comprehensive yet concise

OUTPUT INSTRUCTIONS

  • Only output Markdown
  • Begin with a brief overview of the video's main topic and purpose
  • Structure the summary with clear headings and subheadings that reflect the video's organization
  • Include timestamps in [HH:MM:SS] format before each key point or section
  • Keep the summary concise but comprehensive, focusing on the most valuable information
  • Use bullet points for lists of related points when appropriate
  • Bold or italicize particularly important concepts or takeaways
  • End with a brief conclusion summarizing the video's main message or call to action
  • Ensure you follow ALL these instructions when creating your output.

Attribution

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.