slideshow-builder

name	slideshow-builder
description	Use this skill when the user asks Claude to convert a single hook from the hook bank into a complete TikTok slideshow draft — hook slide + 4-6 body slides + soft CTA final slide, with image direction per slide and TikTok-native text styling.

Slideshow Builder

Purpose: Take one hook → produce a complete slideshow draft (text + image direction per slide) ready for posting.

---

SLIDESHOW STRUCTURE

Slide 1 — HOOK (the only slide that determines stop rate)
  - Single line text, large font, high contrast
  - Image: most arresting visual in the deck

Slides 2-6 — BODY (4-6 slides — vary by depth needed)
  - One point per slide
  - Each slide is a single sentence or bulletized claim
  - Image: supports the point but doesn't fight the text

Slide 7 — SOFT CTA (final slide)
  - One nudge to action
  - "Link in bio" / "See how it works" / "Follow for more"
  - Brand handle visible
  - Image: cleanest brand-aligned visual in the deck

Total: 6-8 slides ideal. Don't go past 10 — engagement drops.

---

SLIDE-BY-SLIDE CONSTRUCTION

Slide 1 (Hook)

• Copy: pull verbatim from hook bank entry
• Image: select from image-collections/<niche>/hook-images/ OR generate via Higgsfield with brand brief
• Text styling: 60-80pt, brand font, high contrast on dark/light background
• Composition: text in top-third OR bottom-third; image is the focal point

Slides 2-6 (Body)

• Copy: pull "Body slide direction" from hook bank entry → expand each into a single slide
• Image: select from image-collections/<niche>/body-images/
• Text styling: 40-50pt (smaller than hook), brand font
• Composition: text overlay; image is contextual

Slide 7 (Soft CTA)

• Copy: pull "Soft CTA hint" from hook bank entry → finalize
• Image: brand-clean visual; could be a product shot OR a lifestyle moment
• Text styling: same as body
• Add brand handle/logo bottom-right

---

TIKTOK-NATIVE TEXT STYLING

Element	Rule
Font	Sans-serif, bold or black weight
Size (hook)	60-80pt
Size (body)	40-50pt
Color	High-contrast against image — white on dark, black on light, NEVER mid-tone on mid-tone
Background	Either: (a) image with text overlay, (b) solid color block with text, (c) text on blurred image
Padding	Generous — text shouldn't touch slide edges
Lines	Hook slides = 1-2 lines max; body slides = 1-3 lines max
Mobile-first	Test at 9:16 vertical, 1080×1920

---

IMAGE SELECTION RULES

Slide	Image source priority
Hook	Highest-stop visual — Pinterest/branded library OR Higgsfield generation
Body	Contextual but not distracting — Pinterest collection OR product detail shots
CTA	Brand-clean — product hero OR signature lifestyle shot

If using Pinterest-sourced images:

• Save to image-collections/<niche>/ for reuse
• Confirm reuse rights (no commercial use of copyrighted images without license)

If generating via Higgsfield:

• Use higgsfield-image-gen skill from static-ad-generator mode
• Apply brand aesthetic rules

---

OUTPUT FORMAT

outputs/slideshows/<hook-slug>-<date>.md:

# Slideshow — <Hook Name> — <Date>
**Brand:** <name>
**Hook category:** <from bank>
**Target avatar:** <avatar>
**Status:** DRAFT (pending review)

## Slide 1 — Hook
- Copy: "[Verbatim from bank]"
- Image: [path or Higgsfield brief]
- Text styling: 70pt, Oswald Bold, white on dark, top-third

## Slide 2
- Copy: "[Body point 1]"
- Image: [path]
- Text styling: 45pt, Inter Bold, black on light, bottom-third

## Slide 3
...

## Slide 7 — Soft CTA
- Copy: "[CTA]"
- Image: [path — brand-clean]
- Text styling: same as body + brand handle
- Brand handle position: bottom-right

## Production Notes
- [Anything special for the design pass]
- [Coach review needed before posting]

## Publish Plan
- Caption: "[3-5 sentence caption with hashtags]"
- Hashtags: <list — niche-relevant, max 5 strong tags>
- Posting time: <per posting cadence doc>
- TikTok title: <short, hook-aligned>

## Compliance Pre-Check
- [ ] No medical claim
- [ ] No copied competitor copy
- [ ] No copyrighted music called out (audio is separate file)
- [ ] No fabricated stats
- [ ] Brand voice consistent across slides

---

ANTI-PATTERNS

Anti-pattern	Why it's wrong
12-slide slideshow	Engagement drops past slide 8-10
Same image style across all slides	Looks like AI-generated mush; vary contextually
Hard CTA on slide 1	Reads as ad — algorithm and audience both punish
Body slides that contradict the hook	Confuses the viewer; they drop off
Brand handle only on slide 7	Add subtle brand presence on body slides too (corner watermark)

---

INTEGRATION

• Hook source → hook-bank-generator skill
• Image source → Pinterest library OR higgsfield-image-gen from static-ad-generator mode
• Performance tracking → dtc-second-brain/raw/ads/ after posting
• Winning slideshow hooks → translate to static-ad-generator/ad-concept-generator for paid amplification