yt-shorts
Skill10 starsChanged 30 days agoMITView on GitHub
Find the moments in a long video that work as Shorts on their own, and write each one a new opening. Use for "cut this into shorts", "clip this", "repurpose this video", "what should I clip".
7 other projects use a version of this, 4 of them edited.
Adopted by
Most recently taken up by
- uplika/uplika-plugintheir own version, first seen Oct 25, 2026
- Rohit-Kumar-AI/youtube-creator-agenttheir own version, first seen Oct 20, 2026
- swayamdhawan/youtube-agentedited, first seen Oct 15, 2026
- ahmedsakri/youtube-agent-proedited, first seen Oct 15, 2026
- nessalazne/youtube-agent-skilledited, first seen Sep 23, 2026
- Jakeschincariol/youtube-agent-skilledited, first seen Sep 16, 2026
Other versions7 projects use a version of this.
Word for word · 0
None found.
Edited · 4
- Jakeschincariol/youtube-agent-skill★ 4723% same
- nessalazne/youtube-agent-skill★ 520% same
- ahmedsakri/youtube-agent-pro★ 114% same
- swayamdhawan/youtube-agent★ 023% same
Wrote their own version · 3
- uplika/uplika-plugin★ 3same name
- Rohit-Kumar-AI/youtube-creator-agent★ 0same name
- tylerprogramming/claude-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 repository11 other files it gives its agents.
- yt-auditSkill
- yt-chaptersSkill
- yt-commentSkill
- yt-editSkill
- yt-packageSkill
- yt-planSkill
- yt-researchSkill
- yt-retentionSkill
- yt-scriptSkill
- yt-seoSkill
- yt-viralSkill
The file
We split this skill into the rules it gives an agent
3 rules. 0 are common, 1 is distinctive, and 2 are very specific to this project.
0
Common
Most of the 7 edited versions we read keep these.
1
Distinctive
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 35
Read ~/.claude/youtube/voice.md if it exists.
- Very specific: line 40
Run shorts.py and work from its shortlist.
- Distinctive: line 42
It starts on a fresh thought: If the first sentence needs the minute before it, it is not a Short.
“Common” is measured: kept by at least 6 in 10 of the 7 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.
Files
yt-shorts
A good Short from a long video is not its best moment. It is a moment that makes sense with nothing around it, and those are rarer.
python3 ${CLAUDE_SKILL_DIR}/shorts.py transcript.srt # top 5, 20-55s each
python3 ${CLAUDE_SKILL_DIR}/shorts.py transcript.srt --top 8 --max 90 --jsonThe tools live in this skill's folder. ${CLAUDE_SKILL_DIR} is that folder in Claude Code; anywhere
it does not expand (claude.ai, Codex), use the folder this SKILL.md was loaded from.
It uses ../yt-edit/deadair.py and ../yt-script/hookscore.py, so install all three together.
To fetch a video's own captions:
yt-dlp --skip-download --write-auto-subs --sub-langs en --sub-format vtt -o transcript "<video url>"Auto-captions have no punctuation, so sentence ends are guesses; the tool says when that is the
case. A whisper transcript gives cleaner cuts.
Before you start
Read ~/.claude/youtube/voice.md if it exists. The new opening line is spoken or captioned in the
creator's voice, and the "will not claim" list still applies.
Picking
Run shorts.py and work from its shortlist. Each Short needs all three:
- It starts on a fresh thought. If the first sentence needs the minute before it, it is not a Short.
- Something turns - a claim, then what complicates or proves it.
- It ends on a line, not a trailing thought.
The tool checks the first and the third and scores the opening line. Only you can check the turn: read each candidate, drop those without one, and show the best five with their timecodes and first lines so the creator can reject one at a glance.
Length: Shorts can run to three minutes. 20-55 seconds is a choice: short spans finish and
loop better. Use --max for longer only when the turn needs the room. Past 60 seconds, any
Content ID claim blocks the Short everywhere, so a long span over someone else's music is not
an option.
Writing each one
- A new first line. The long video's line leans on context this viewer never saw. Write the
replacement and score it with
${CLAUDE_SKILL_DIR}/../yt-script/hookscore.py. - On-screen text for the first two seconds, in different words from the spoken line.
- A loop: what the last line sets up so the first line answers it.
- A vertical note: what the 9:16 crop loses from a 16:9 frame, and whether it matters.
Before you hand it over
This skill drafts; you publish. End with the copy-ready block and one question: post it as is, or another pass?
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
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.