agentleFS
nicojunknicojunk/claude-ig/skills/ig-hook/SKILL.md

ig-hook

Skill17 starsChanged 8 months agoMITView on GitHub

Generate and score Instagram hooks. Produces 3 hook variants across 5 categories (Correction, Identity Trigger, Bold Claim, Curiosity Gap, Relatable Pain), scores each on a 25-point rubric, and outputs the top 2 as A/B variants with text overlay, spoken text, and visual action. Supports compound hooks for higher scores. Use when user says "hook", "hook schreiben", "hooks generieren", "reel hook", "ig hook", "hook ideen", "hook varianten".

1 other project uses a version of this, 1 of them edited.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 0

    None found.

    Edited · 1

    1. workflownicojunk/claude-skills★ 099% same

    Wrote their own version · 0

    None found.

    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 repository13 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 6 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If not configured, ask the user about their niche, audience, and voice.

    line 26
  2. Distinctive:

    Default: Reel (if not specified)

    line 42
  3. Distinctive:

    If the user provides all three in one message, skip asking and proceed directly.

    line 46
  4. Very specific:

    Check references/hook-library.md for proven templates in each category

    line 64
  5. Distinctive:

    Write all hook text in German

    line 78
  6. Distinctive:

    Every hook must work independently (no context from the rest of the Reel needed)

    line 83
  7. Distinctive:

    Tie-breaking: If two hooks have the same total score, prefer the one with the higher Curiosity Gap score.

    line 169

≈ “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

IG Hook Generator -- First 3 Seconds That Stop the Scroll

Generates, scores, and presents Instagram hook variants optimized for the target audience defined in references/account-baseline.md. Every hook is designed to maximize watch time by winning the first 3 seconds.

Load account context (niche, audience, voice) from references/account-baseline.md. If not configured, ask the user about their niche, audience, and voice.

Key references:

  • references/hook-library.md -- 50+ templates, 5 categories, compound hook rules
  • references/scoring-system.md -- Hook Strength sub-criteria (25pts)
  • references/format-specs.md -- First 3 seconds requirements per format

Workflow

Phase 1: Context Gathering

Collect the following from the user (ask once if missing):

  1. Topic -- What is the Reel/Carousel/Story about?
    • Example: "Intervallfasten Fehler", "Proteinbedarf ab 40", "Bauchfett loswerden"
  2. Format type -- Reel, Carousel, or Story?
    • Default: Reel (if not specified)
  3. Content angle -- What is the core insight, myth, or transformation?
    • Example: "Die meisten essen ZU WENIG Protein", "Bauchfett hat nichts mit Sit-Ups zu tun"

If the user provides all three in one message, skip asking and proceed directly.

Phase 2: Category Selection

Select hook categories from the 5 available types. Load references/hook-library.md for templates and examples within each category.

#CategoryMechanismWhen to use
1Correction"Du machst X falsch" / "Das stimmt nicht"Myth-busting, common mistakes, misconceptions
2Identity Trigger"Frauen ab 40, die..." / "Wenn du zu denen gehorst, die..."Audience self-selection, personal relevance
3Bold Claim"X ist der grosste Fehler" / "Vergiss alles, was du uber Y weisst"Strong takes backed by science, pattern interrupts
4Curiosity Gap"Was passiert, wenn du X machst..." / "Das hat niemand auf dem Schirm"Unexpected facts, counterintuitive insights
5Relatable Pain"Kennst du das..." / "Wenn du abends vor dem Kuhlschrank stehst..."Shared struggles, emotional connection, daily situations

Selection rules:

  • The 3 generated hooks MUST use at least 2 different categories
  • Compound hooks (combining 2+ categories) are encouraged and score higher
  • Check references/hook-library.md for proven templates in each category
  • Match the category to the content angle (myth = Correction, personal = Identity Trigger, etc.)

Phase 3: Hook Generation

Generate exactly 3 hook variants. Each hook MUST include all three elements:

ElementWhat it isExample
Text OverlayBold text shown on screen in the first 3 seconds"DAS zerstort deinen Stoffwechsel"
Spoken TextWhat the creator says out loud in the first 3 seconds"Wenn du das morgens als Erstes machst, sabotierst du deinen gesamten Fettstoffwechsel."
Visual ActionWhat the viewer sees the creator doingCreator schaut direkt in die Kamera, schuttelt den Kopf, halt eine Kaffeetasse hoch

Generation rules:

  • Write all hook text in German
  • At least 1 hook should be a compound hook (combining 2+ categories)
  • Text overlay: max 6-8 words, must be readable in under 2 seconds
  • Spoken text: max 2 sentences, must be completable in 3 seconds
  • Visual action: must create movement or contrast that stops the scroll
  • Every hook must work independently (no context from the rest of the Reel needed)
  • Tag each hook with its category (or categories for compound hooks)

Output format per hook:

### Hook [A/B/C]: [Category] (+ [Category] if compound)

**Text Overlay:** [bold on-screen text]
**Spoken Text:** "[what the creator says]"
**Visual Action:** [what the viewer sees]

Phase 4: Hook Scoring

Score each of the 3 hooks using the Hook Strength rubric (25 points max). Load references/scoring-system.md for detailed sub-criteria.

Sub-criterionMaxWhat it measures
Curiosity Gap5 ptsDoes the hook create an open loop the viewer needs to close?
Pattern Interrupt5 ptsDoes it break the scroll pattern visually or verbally?
Identity Trigger5 ptsDoes the viewer think "das bin ich" or "das betrifft mich"?
First-3s Clarity5 ptsIs the hook's promise clear within 3 seconds? No ambiguity.
Compound Mechanics5 ptsDoes it layer 2+ hook categories for stronger pull? Single-category hooks cap at 3.

Scoring rules:

  • Score each sub-criterion with a whole number (0-5)
  • Single-category hooks can score max 3/5 on Compound Mechanics
  • Compound hooks that cleanly integrate 2+ categories score 4-5 on Compound Mechanics
  • Forced or awkward category combinations score lower than clean single-category hooks
  • Reference references/scoring-system.md for per-point thresholds within each sub-criterion

Score each hook in a table:

| Sub-criterion | Hook A | Hook B | Hook C |
|---------------|--------|--------|--------|
| Curiosity Gap | X/5 | X/5 | X/5 |
| Pattern Interrupt | X/5 | X/5 | X/5 |
| Identity Trigger | X/5 | X/5 | X/5 |
| First-3s Clarity | X/5 | X/5 | X/5 |
| Compound Mechanics | X/5 | X/5 | X/5 |
| **Total** | **X/25** | **X/25** | **X/25** |

Phase 5: A/B Variant Output

Present the top 2 scoring hooks as the final A/B variants.

For each of the 2 selected hooks, output:

  1. The full hook (text overlay, spoken text, visual action) from Phase 3
  2. The score breakdown from Phase 4
  3. A 2-3 sentence explanation of WHY this hook works, referencing the specific mechanisms that drive its score (e.g., "The identity trigger 'Frauen ab 40' immediately filters for the target audience, while the correction 'das stimmt nicht' creates a curiosity gap the viewer needs to resolve.")

Output format:

## A/B Hook Variants for: [Topic]

### Variant A: [Category] -- [Score]/25
**Text Overlay:** [text]
**Spoken Text:** "[text]"
**Visual Action:** [description]

**Why this works:** [2-3 sentences explaining the mechanisms]

---

### Variant B: [Category] -- [Score]/25
**Text Overlay:** [text]
**Spoken Text:** "[text]"
**Visual Action:** [description]

**Why this works:** [2-3 sentences explaining the mechanisms]

---

### Recommendation
[1-2 sentences on which variant to test first and why, or when to use each]

Tie-breaking: If two hooks have the same total score, prefer the one with the higher Curiosity Gap score. If still tied, prefer the compound hook.

Quality Gates

These gates are inherited from the parent ig skill. A hook that violates any gate MUST NOT be delivered.

#GateRule
G1Score before deliveryNEVER deliver hooks without scoring all 3 variants
G2Em dashesNEVER use em dashes in hook text, spoken text, or any output
G4Controversial opinionsNEVER generate hooks built on controversial opinions (high comments, zero saves). This includes: diet shaming, attacking other fitness approaches, medical claims without scientific backing, divisive social commentary.
G7Safe zonesNEVER place text overlay in format-specific unsafe zones (check references/format-specs.md)

Anti-Patterns

Anti-PatternWhy
All 3 hooks from the same categoryReduces variety, misses stronger alternatives
Text overlay longer than 8 wordsUnreadable in 2 seconds, kills the hook
Spoken text longer than 3 secondsExceeds the hook window, loses viewers
Vague hooks ("Wusstest du das?")No specificity, no curiosity gap, no identity trigger
Clickbait without payoffDestroys trust with the creator's audience
Copy competitor hooks verbatimOriginality Score (2026 algorithm) penalizes recycled formats
Controversial opinion as hookGenerates comments but zero saves, algorithm ignores (G4)
Missing visual actionA hook is not just words; the visual must contribute to the stop

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.