agentleFS
fahdAghribifahdAghribi/youtube-channel-skills/skills/youtube-script-writer/SKILL.md

youtube-script-writer

Skill2 starsChanged 12 days agoMITView on GitHub

Write a retention-first YouTube script for long-form or Shorts: hook for the first 30 seconds, open loops, value beats with pattern interrupts, mid-roll CTA placement and an ending that feeds the next video. Outputs a two-column script (spoken / on screen) with timing. Not for titles and thumbnails (use youtube-title-thumbnail) or repurposing an existing video (use youtube-shorts-repurposer).

7 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use content-repurposer.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.

The file

We split this skill into the rules it gives an agent

2 rules. 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 6 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

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

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

    Retention pass: run the checklist below. Mark each risky moment with ⚠ drop risk and a fix.

    line 32
  2. Distinctive:

    Never fabricate experiences, results, testimonials or sources.

    line 65

“Common” is measured: kept by at least 6 in 10 of the 6 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

YouTube Script Writer

Scripts written for the ear and for the retention curve. Packaging comes first: if the video has no title + thumbnail concept yet, write 3 title candidates before the script (or call youtube-title-thumbnail), because the hook has to continue that promise.

When to use

  • "Write the script for...", "give me an outline", "improve my intro"
  • Rewriting a script whose retention data shows an early drop

Input

  • Idea + chosen title/thumbnail (or ask for them)
  • Format: talking head, tutorial, test, story, list, comparison, Short
  • Target length (long-form default 8-12 min ≈ 1,200-1,800 spoken words at ~150 wpm; Short ≤ 60 s default, ≤ 180 s max)
  • Receipts: numbers, stories, demos the creator has (profile §Assets or the user)

Steps

  1. Load the profile (voice, language, persona, recurring segments).
  2. Lock the promise: one sentence, "By the end, the viewer will ___". Everything not serving it is cut.
  3. Pick a hook pattern from ../../references/hook-patterns.md that pairs with the title formula.
  4. Outline with the structure in references/script-structures.md for the chosen format. Place the main open loop by 0:30 and close it in the last third.
  5. Write the script in two columns: Spoken and On screen (b-roll, text, zoom, demo). Add a visual change every 5-8 s in the first minute and every 15-30 s after.
  6. Place CTAs: subscribe ask only after the first payoff (never in the hook); product/newsletter CTA once, where it solves the problem just discussed; end screen CTA pointing to one specific next video.
  7. Retention pass: run the checklist below. Mark each risky moment with ⚠ drop risk and a fix.
  8. Return the script plus a 1-line summary, word count and estimated duration.

Output

TITLE: ...            THUMBNAIL: ...           PROMISE: ...
EST. DURATION: 10:40  WORDS: 1,610             HOOK: H4

| Time | Spoken | On screen |
|------|--------|-----------|
| 0:00 | ...    | ...       |

Then: Retention checklist (passed / fixed), CTA map, End screen target.

Retention checklist

  • First line continues the title; no greeting, no channel intro
  • Promise restated or shown by 0:10, open loop planted by 0:30
  • No "before we start" detours; sponsor (if any) after the first payoff
  • Every section starts with why the viewer should care (stakes) before how
  • At least one re-hook every ~2 minutes ("but this is where it went wrong")
  • No section > 90 s without a visual change or a new micro-question
  • The ending does not say "that's it for today": it hands off to the next video
  • Receipts are real (profile or user). Unknown numbers marked [NUMBER?]

Hard rules

Global voice rules: see root SKILL.md §Voice rules.

  • Spoken language: short sentences, contractions, second person. Read-aloud test.
  • Shorts: hook in the first second (visual + words), one idea, payoff by the end, loop the last line into the first when possible.
  • Never fabricate experiences, results, testimonials or sources. If the user wants a claim, they supply it.

Files

  • references/script-structures.md — structures per format
  • ../../references/hook-patterns.md
  • ../../references/voice-rules.md

Related skills

  • youtube-title-thumbnail — packaging the script must continue
  • youtube-metadata-optimizer — chapters from the script's section times
  • youtube-shorts-repurposer — cut Shorts from the finished video

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.