0xS4N4TG/papercut-explainer / skills/papercut-hook
papercut-hook
VerifiedWrite or fix the hook of a papercut reel. Produces the SAY / WRITE / SHOW / ALIGN block, five variants scored with the hook check, the frame-0 cover title and the motion out of the hook. Use for "write a hook for...", or when a reel's first three seconds lose people.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 4.9 KB |
| references/hook-patterns.md | markdown | 4.0 KB |
references/hook-patterns.md
View on GitHub# Hook patterns for tech and developer-tool reels
Seven patterns that recur in outlier reels about AI and developer tools. Use them in the hook workflow (`../SKILL.md`): write the five hook variants from at least three different patterns, then score them with the hook check. Every pattern still has to pass SAY / WRITE / SHOW alignment and stay true to the facts sheet. The examples below use invented products.
## Topic first: first on the right repo wins
When several creators cover the same trending repo with the same hook, the topic carries all of them, and the order matters more than the edit. The first reel takes most of the views, and each later one gets less. When a repo is climbing fast, ship within a day or two, and pick an angle the earlier reels didn't use. Being fifth on a topic people want still beats being first on a topic nobody wants.
## 1. Borrowed authority + giveaway
"This hackathon winner just released his entire coding-agent setup for free."
In 12 words it gives a credible source, "just" (it's news), "entire" (it's complete) and "free".
- The source must be named and credible: a hackathon winner, a platform engineer, a company's security team, a repo with tens of thousands of stars. "Someone just released..." is the version that fails. It is anonymous, and it takes too many words to reach the number.
- Stack the three signals: who, how complete, what it costs.
## 2. Name the viewer, then a number
"5 plugins every Codewright user needs right now."
The first words tell the algorithm and the viewer who it's for. The number keeps people watching until #5.
- Name the tool the viewer uses (Codewright users), not a vague "developers".
- Be careful with list counts. A bare "here are N things" opener tends to flop. A number that counts things you then show (5 plugins) works better than one that promises tips.
## 3. One object, one big number, one twist
"One single file on GitHub with over 100,000 stars", then the flip: the file tells the AI what *not* to do.
- One concrete object (a file, a repo, a command), one big number, then the twist the viewer didn't expect.
- The twist is the gap. Don't answer it in line 1.
## 4. Price shock
"Codewright just became completely free." "(No credit card.)" Taking the cost away is the most reliable hook in this niche.
- Only when it's true for the viewer's real setup. If part of it costs money (a paid engine, an API key), say so in the catch beat.
- Don't copy a competitor's exact price line on the same topic.
## 5. Pain, and the fear of looking amateur
Pain fits cut-paper comedy well, because the picture can act out the symptom.
- "Why AI coding tools keep breaking things."
- "...so your app doesn't get sued."
- A skit that opens with a character asking why everyone else's AI-built app looks like a $10K job while theirs looks like 2006.
- "If your AI keeps giving you bad answers, make 100 copies of it fight to the death."
Same shape as the template "If you have [symptom], you probably have [cause]". Name a symptom the viewer has felt this week.
## 6. Verdict or conflict
"65% chose wrong." Tier lists. People come into the comments to argue, and comments push the reel.
- Works best when there's a real split (legal vs not, tool A vs tool B). Put the verdict in the payoff, not the hook.
- With a comment keyword, the argument and the keyword both land in the comments, which is fine.
## 7. Own authority
The creator speaks from their own use ("I tested...", "I pointed Codewright at...") and shows their own result, not the README's.
- For a faceless series, the narrator tells it in the first person ("I gave Codewright an app with no source code..."). Say it only if it's true. Run the test yourself, or word the line as what the docs and public examples show. Whatever the narrator says the tool did must match the tool's docs and showcases (the facts sheet), so commenters can't catch it out.
- It's also the best way to stand out on a topic other creators already covered. They read out the README, and you tell it as your own test.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| add-product-e2e-evalpaperclipai/paperclipAdd or extend a Paperclip full-stack runner E2E workflow, fixture, matcher, or report evidence path for local or Daytona execution. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| add-runner-evalpaperclipai/paperclipAdd or extend a Paperclip Runner protocol evaluation definition, roster, assertion, or report fixture with provenance and narrow validation. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| diagnose-why-work-stoppedpaperclipai/paperclipCompatibility entry point for failed, stalled, looping, or over-recovered Paperclip work. Use fix-failed-paperclip-run for the full diagnosis, regression, repair, real verification, and green-PR workflow; honor an explicitly requested diagnosis-only scope. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| fix-failed-paperclip-runpaperclipai/paperclipDiagnose and fix failed, stuck, slow, or incorrectly blocked Paperclip tasks and runs. Use with a task/run URL or failure report to trace the root cause, prove a regression red then green, check second-order effects, verify the real workflow, and deliver a PR with passing checks. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |