agentleFS
gokulsvisiongokulsvision/Crewm8-Social-Media-Manager-Skill-Graph/skills/creation/hook-write/SKILL.md

hook-write

Skill17 starsChanged 7 months agoMITView on GitHub

Generate 5-10 hook variations for any post, thread, or caption across X, LinkedIn, Instagram, and TikTok — using different styles (question, contrarian, statistic, story, observation, promise) so the best one can be tested or selected.

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    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
  1. Distinctive:

    Before writing any post or thread (hook-first writing)

    line 29

≈ “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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Hook Write

Generate high-performing hooks for any content type and platform. The hook is the single most important element — if nobody stops scrolling, the rest doesn't matter. This skill produces multiple variations so the best one can be A/B tested or selected.

Purpose

90% of the post's success is determined before the reader finishes the first sentence. Most failed posts aren't bad posts — they have bad hooks. This skill separates the hook from the body so each can be optimized independently.

When to Use

  • Before writing any post or thread (hook-first writing)
  • A/B testing variations (deploy 2 versions, measure which wins)
  • Existing post underperformed — diagnose if the hook was the problem
  • Content calendar needs hooks drafted in advance
  • Founder says "I have an idea" but doesn't know how to open

Inputs Required

  • Core idea (one sentence — what this content communicates)
  • Target platform
  • Desired emotional reaction (curiosity, recognition, urgency, validation)
  • Brand voice constraints (any hook styles that are off-limits)

Quick Reference

Hook StylePatternBest PlatformExample
Promise"I [did X]. Here's what I found"X, LinkedIn"I analyzed 100 pricing pages. Found 3 patterns."
Contrarian"[Common belief] is wrong"X, LinkedIn"Everyone says niche down. Here's when you shouldn't."
Statistic / LedeLead with a surprising numberX, LinkedIn"97% of cold outreach gets no reply. The 3% share one trait."
Story HookStart mid-narrativeX, LinkedIn, IG"We hit $10K MRR. Then almost lost it all."
QuestionAsk a provocative questionLinkedIn, IG"What's the one metric that actually changed your decisions?"
ObservationState the unstatedX, LinkedIn"The best founders I know are terrible at celebrating wins."
How-to Promise"How to [X] without [Y]"LinkedIn, IG, TikTok"How to price your SaaS without guesswork."
Pattern InterruptUnexpected format/openerIG, TikTokVisual hook saying the opposite of what you'd expect

Procedure

  1. Define the core idea. One sentence: "This post is about [X] and the reader should feel [Y]."
  2. Select the target emotion:
    • Curiosity ("I need to know more")
    • Recognition ("That's exactly me")
    • Intrigue ("Wait, is that true?")
    • Urgency ("I need this now")
    • Validation ("Someone finally said it")
  3. Generate 5-10 hooks across 5-6 styles (selecting the most suitable for platform and topic):
    • At least 1 Promise-style
    • At least 1 Contrarian-style
    • At least 1 Story-style (if applicable)
    • At least 1 Question-style (for LinkedIn/IG)
    • At least 1 Statistic/Number-led (if data available)
    • At least 1 Observation-style (for X)
  4. Adapt each hook to the target platform:
    • X: First 3-5 words are everything. Max 280 chars total.
    • LinkedIn: First 2 lines (before "see more"). Tease without clickbait.
    • IG: Hook works with the visual. First line of caption is critical.
    • TikTok: Text overlay + first 1.5 seconds of audio/visual.
  5. Score each hook on:
    • Curiosity gap: does it make you need to know more? (1-10)
    • Specificity: is it concrete vs. generic? (1-10)
    • Brand-fit: does it sound like the brand? (1-10)
    • Platform-fit: does it follow platform rules? (1-10)
  6. Rank and recommend. Top 2 hooks: one primary pick + one A/B challenger.

Output Format

# Hook Variations: [Post Topic / Platform]

## Core Idea
[One sentence — what this post is about]

## Hook Variations

### 1. [Style: Promise/Contrarian/etc.]
[Hook text, formatted for platform]
**Curiosity:** [1-10] **Specificity:** [1-10] **Brand-fit:** [1-10] **Platform-fit:** [1-10]

### 2. [Style]
...

## Top Picks
1. **Recommended:** [Hook + why]
2. **A/B challenger:** [Alternate + why]

## Platform Adaptations
- X version: [shorter, punchier]
- LinkedIn version: [slightly more professional framing]
- IG version: [visual-led version]
- TikTok version: [text overlay + hook timing]

Done Criteria

The skill is complete when:

  1. At least 5 hook variations exist across different styles
  2. Each hook is formatted for the target platform
  3. Hooks are scored on all 4 dimensions
  4. A primary and challenger hook are recommended
  5. Platform adaptations are provided (if applicable)

Pitfalls

  • Hooks that overpromise ("This one trick will 10x your revenue") — kills trust
  • Hooks that are all sizzle, no steak (the post doesn't deliver on the promise)
  • Question hooks where the audience already knows the answer
  • Clickbait that would embarrass the founder
  • Using the same hook style for every post (pattern fatigue)

Verification

For each top hook: what does it promise? Does the actual post deliver that exact promise? If the answer is "sort of," rewrite either the hook or the post until it's a clean match.

Example

Input: Topic = "Why most SaaS startups fail at cold email." Platform = X. Emotion = curiosity + recognition.

Output:

  1. (Promise) "I read 50,000 cold emails. Here are the 3 that booked meetings." — Curiosity: 9, Specificity: 9, Brand-fit: 8, Platform-fit: 10
  2. (Contrarian) "Everyone tells you to personalize cold emails. That's wrong." — Curiosity: 8, Specificity: 8, Brand-fit: 7, Platform-fit: 9
  3. (Statistic) "97% of cold emails get ignored. The 3% that work share 1 pattern." — Curiosity: 9, Specificity: 9, Brand-fit: 8, Platform-fit: 10
  4. (Story) "I sent 1,000 cold emails. Got 2 replies. Here's what I changed." — Curiosity: 7, Specificity: 8, Brand-fit: 8, Platform-fit: 9
  5. (Observation) "The cold emails that work don't look like cold emails." — Curiosity: 7, Specificity: 7, Brand-fit: 9, Platform-fit: 10

Recommended: #3 (statistic-led, highest combined score). Challenger: #2 (contrarian, different angle).

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.