Top-shortlisted hooks from a 60-candidate run for a LinkedIn post about hiring mistakes.
This is a trimmed sample. The full run produces 50-100 candidates, a full shortlist, A/B pairings, and a lifecycle deployment plan.
Five moves. Four gates. No generation before the voice is locked.
The most common failure of AI-generated hooks is volume without voice. The skill is built to prevent that. Brief locks first. Voice locks second. Only then does generation begin.
Capture the hook task.
The platform, the audience, the conversion goal, the lifecycle. What this hook actually has to do.
Lock the voice.
Builds or loads the brand's voice fingerprint. Non-negotiable. No generation begins without it.
Run the drills. 50 to 100 candidates.
Pulls inputs in conversation. Generates the hook pool. Sorts and shortlists. Halts so you can redirect.
Score and rank.
Tests against the 10 qualities of a great hook. Drops what fails. Keeps the shortlist tight.
Final hooks plus A/B plan.
The shortlist with A/B test pairings, lifecycle plan, and scale-out logic for repurposing across channels.
Pick the drill that matches the channel.
Six drill drawers. Each one handles a specific stage of the hook job. The skill reaches into the right combination for the platform and lifecycle.
Audience.
Drills for naming who the hook has to stop. The specific person, the moment they're in, what they're scrolling away from.
When "everyone on LinkedIn" is the answer.
Concept.
Drills for finding the angle. The If/Then formula. The Pattern Interrupt. The Curiosity Gap. The Named Pattern reveal.
When the hook keeps sounding like an opinion.
Generation.
Drills for producing 50 to 100 candidates fast. Variation across angle, structure, length, and tone.
When ten hooks is "all you can think of."
Templates.
Drills using viral post templates as structural beat patterns. Not for copying. For pressure-testing the candidates against proven shapes.
When the format isn't working.
Test.
Drills for scoring against the 10 qualities of a great hook. Tested formulas, comparison drills, and the "would I scroll past this myself" check.
When everything looks fine and nothing is great.
Scale.
Drills for cross-channel reuse. One hook, six platform-shaped variants. Lifecycle plan for when each one retires.
When one good hook needs to feed a content engine.
Four sequential artefacts. Each one a gate.
Brief. Voice. Pool. Final. Approval before each next step.
The Hook Brief
The real task, named.
Platform, audience, conversion goal, lifecycle, constraints. Halts the work.
The Voice Profile
Locked before generation.
Fingerprint of the brand voice. Signature phrases, rhythm, banned words. Non-negotiable.
The Hook Pool
50 to 100 candidates.
Sorted by angle. Pre-scored. Shortlisted into the top 10 by quality. Halts so you can pick or redirect.
The Final Hooks
Ready to test.
Final shortlist with A/B pairings, lifecycle plan, and platform-shaped variants for scale-out.
The brief-and-voice gate, in action.
Built for the people doing the work, not watching it.
- Founders posting on LinkedIn who keep getting "fine" engagement and want "wait, this stopped me" engagement.
- Content teams running channels and tired of the open landing soft.
- Marketers running paid social where the hook is the entire CTR.
- Podcast hosts naming episodes and writing video thumbnails.
- Anyone with a body of work that gets ignored because the opener is weak.
- ×People who want "viral hooks" without doing the brief work first.
- ×Anyone hoping Claude will reverse-engineer engagement from no inputs.
- ×Brands that don't have a voice yet.
- ×People who want clickbait. The skill generates hooks for actual posts you can deliver on.
- ×Anyone who wants to skip the voice profile.