agentleFS
motola10motola10/claude-code-base/.claude/skills-disabled/youtube-to-guideline/SKILL.md

youtube-to-guideline

Skill1 starsChanged 5 months agoMITView on GitHub

Use this skill whenever the user provides a YouTube URL and wants structured documentation, a guideline, reference material, or actionable notes from it. Trigger on any combination of a YouTube link plus intent words like: "make a guide", "create a guideline", "turn this into docs", "extract the workflow", "create reference material", "guideline from this video", "transcribe and document", "make notes from this", or "build a knowledge doc". Also trigger proactively when the user shares a YouTube URL in a technical or workflow context and asks for "something useful" — they almost certainly want a guideline. The skill runs two steps: transcribe with youtube-transcribe, then create a guideline with claude-guideline-creation.

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 3 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Only after confirming .claude/docs/<topic-slug>-guideline.md was written successfully and has real content (not empty, not a stub), delete …

    line 64
  2. Distinctive:

    When all three steps are done, confirm the single remaining output file:

    line 79
  3. Very specific:

    Do not re-implement transcript fetching — youtube-transcribe handles it

    line 90
  4. Very specific:

    Do not proceed to Step 2 if youtube-transcribe reported errors or agent-to-confirm failed

    line 91
  5. Distinctive:

    Do not write a guideline that summarises what was said — it must contain rules, steps, and patterns a practitioner can act on directly

    line 92
  6. Distinctive:

    Do not delete the transcript before verifying the guideline file actually exists and has real content — an empty or failed guideline with …

    line 94

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “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

What this skill does

Converts any YouTube video into a single persisted artefact by chaining two existing skills:

  1. <video_id>_transcript.md — produced by the youtube-transcribe skill (intermediate — deleted in Step 3)
  2. .claude/docs/<topic-slug>-guideline.md — produced by the claude-guideline-creation skill (the deliverable)

The transcript is scratch material for building the guideline, not something the user asked to keep. Run the three steps in sequence. Do not skip or reorder them.


Step 1 — Transcribe the video

Invoke the youtube-transcribe skill with the YouTube URL.

That skill handles everything:

  • Extracting the video ID
  • Fetching all segments via youtube-transcript-api (v1.x .fetch() API)
  • Assembling a structured, timestamped, section-headed transcript
  • Applying ≥95%-confidence ASR corrections
  • Running agent-to-confirm to verify completeness

The output is <video_id>_transcript.md saved to the project directory.


Step 2 — Create the Engineering Guideline

Invoke the claude-guideline-creation skill in Create mode:

Create a new Engineering Guideline titled "[Video Title from transcript]" based on
the transcript at <video_id>_transcript.md.
Save it to .claude/docs/<topic-slug>-guideline.md.

The guideline should distill the practical workflow, rules, and patterns from the
video into actionable reference material — not a summary of what was said.
A reader who has not watched the video should be able to follow it.
Cover all major topics identified in the transcript sections.

The guideline skill applies its own formatting rules: numbered sections, H3 sub-sections, tables over prose, * **Bold Key:** description bullets, code blocks for commands and prompts, blockquotes for standout rules.


Step 3 — Delete the intermediate transcript

Only after confirming .claude/docs/<topic-slug>-guideline.md was written successfully and has real content (not empty, not a stub), delete the transcript:

rm <video_id>_transcript.md

If guideline creation failed or produced something incomplete, do NOT delete the transcript — keep it around so the guideline step can be retried or debugged without re-fetching the video.


Completion confirmation

When all three steps are done, confirm the single remaining output file:

Guideline : .claude/docs/<topic-slug>-guideline.md
(transcript deleted — was intermediate scratch material)

What NOT to do

  • Do not re-implement transcript fetching — youtube-transcribe handles it
  • Do not proceed to Step 2 if youtube-transcribe reported errors or agent-to-confirm failed
  • Do not write a guideline that summarises what was said — it must contain rules, steps, and patterns a practitioner can act on directly
  • Do not delete the transcript before verifying the guideline file actually exists and has real content — an empty or failed guideline with the source already deleted is unrecoverable without re-fetching

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.