agentleFS
Sign inStart free
gladosAagmangladosAagman/youtube-agent-skill/skills/yt-voice/SKILL.md

yt-voice

Skill1 starsChanged 5 days agoMITView on GitHub

Measure how the creator actually talks on camera from their own transcripts and write voice.md from evidence, then check any script for drift before they read it out loud. Use for "write my voice.md", "learn how I talk", "this script doesn't sound like me", "make it sound like me", first-time setup of the pack, or when the user shares transcripts or caption files of their own videos.

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 1

    1. ahmedsakri/youtube-agent-pro★ 1same 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 repository14 other files it gives its agents.

The file

Create a free account to mount
Files
SKILL.md

yt-voice

On YouTube the creator has to say the words. A script in the wrong voice is not just off-brand, it is unreadable on camera: the sentences are the wrong length for their breath, the words are not theirs, and it shows in the take. This measures the voice from what they have already said.

python3 voice.py video1.srt video2.srt video3.srt               # caption files: one sample each
python3 voice.py transcripts.txt --save ~/.claude/youtube/voice.json   # text: --- between samples
python3 voice.py --profile ~/.claude/youtube/voice.json --check script.txt

Input

Three to ten transcripts of the creator's own videos: .srt or .vtt caption files (timestamps are stripped automatically, each file is one sample), or plain text with a --- line between transcripts. Use their best-performing videos and their most recent ones, not an old style they have moved away from.

Auto-captions have no punctuation, which makes sentence-length numbers meaningless. Use the creator's edited captions, a Whisper transcript, or the transcript YouTube generates after they have corrected it. Say this if the profile shows one 300-word "sentence".

What it measures

Sentence length and spread (how long they talk before a breath), contractions, I / we / you (does the channel talk to someone), questions, numbers per 100 words, the words that recur across videos, and how videos open. Each is a mean and a spread.

Then write voice.md

  1. Measure and save the profile.
  2. Fill in templates/voice.md: who they are talking to, what they will not claim, their format.
  3. Paste the Measured voice block under "How I actually talk". Point out where it disagrees with what the creator said about themselves.
  4. Save to ~/.claude/youtube/voice.md.

Checking a script

--check script.txt scores the script against the profile and names the metric furthest from them. /yt-preflight runs it automatically when the profile exists. A script that scores DRIFT gets read out loud once before anyone fixes it: the ear finds the wrong line faster than the numbers.

Say this honestly

Stylometry on a few transcripts. It catches a script that drifted into generic YouTube voice. It does not make a script good, and it is for the creator's own voice only.

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