viewer-panel
Skill23 starsChanged 18 days agoMITView on GitHub
Six different viewers watch the first 3 seconds of a video and each says swipe or stay, and why. Use to pressure-test a hook from the point of view of the doom-scroller, the sceptic, the beginner, the niche expert, the buyer and the rival creator.
In this repository8 other files it gives its agents.
- four-failures-checkSkill
- hook-formatsSkill
- hook-masterSkill
- lock-inSkill
- niche-tunerSkill
- the-councilSkill
- viral-scoutSkill
- visual-matchSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
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
- Distinctive: line 39
Run the panel again on the new hook.
- Distinctive: line 45
Don't let all six agree by default.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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.
Files
Viewer Panel
A hook is judged by strangers in a few seconds. Simulate six of them. Each one watches only what is actually in the first 3 seconds: the first shot, the on-screen text and the spoken words. They know nothing else.
This is a thinking tool, not data. Never present the panel as a survey, a prediction or a percentage of real viewers.
The six viewers
| Viewer | Who they are | What makes them stay | What makes them swipe |
| Doom-scroller | Thumb moving, half watching, sound maybe off | A clear picture of something surprising or wanted, readable text | Talking head with nothing happening, a greeting, a logo |
| Sceptic | Has seen a thousand hooks, smells clickbait | A specific, checkable claim with proof on screen | "Nobody tells you", "secret", round stats with no source, hype |
| Beginner | New to the topic, doesn't know the words | Their own problem in their own words | Jargon, insider terms, assumes knowledge |
| Niche expert | Knows the topic well | A real detail or a claim that's new to them | The obvious advice they already know |
| Buyer | Could pay for this (book, sign up, buy) | Their fear or want named, and a hint the creator can solve it | Brand-first, "about us", nothing about their situation |
| Rival creator | Makes content in the same niche | A format or angle they haven't seen done this way | The same opening everyone in the niche uses |
Pick the two who matter most for this video and say why (a local business needs the buyer and the beginner; an educational creator needs the doom-scroller and the niche expert).
How to run it
For each viewer, one line:
[Viewer]: SWIPE at [second / word]. "[the exact thing that loses them]"
[Viewer]: STAY. "[the exact thing that holds them]", and what they now want to know.Then:
- Who matters most: the two viewers, and why.
- Pattern: what the swipes have in common (usually one failure from
four-failures-check). - What would keep the two who matter: one sentence.
Run the panel again on the new hook. A rewrite that keeps the sceptic but loses the beginner is not finished.
Rules
- Tie every verdict to something in the opening. "Feels boring" is not a reason. "Desk and white coat for 2 seconds, nothing moves" is.
- The sceptic is the honesty alarm. If they would call a line clickbait or a lie, fix it, even if the doom-scroller loves it.
- Don't let all six agree by default. Real viewers differ; if they all say stay, look harder.
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.