Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Landing page conversion optimization with layout rules, hero section design, and CTA psychology. Covers above-the-fold formula, social proof placement, mobile design, and F-pattern reading. Use for: startup landing pages, product pages, SaaS marketing, conversion optimization. Triggers: landing page, hero section, above the fold, conversion optimization, landing page design, cta button, hero image, landing page layout, saas landing page, product page design, conversion rate, landing page best pr
.claude/skills/aiskillstore-landing-page-design/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 81% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 38% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 20% | 0% |
| case-18 | ✗→✓ | ▲ Improved | 411% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 125% | 0% |
Build landing pages that convert AND captivate. This skill combines conversion-focused structure with distinctive visual design to create pages that stand out in an AI-saturated world. The goal: pages worth $50-100 that you'd be proud to sell.
BEFORE writing any code, you MUST run the Vibe Discovery process. This isn't a lookup table - it's a creative prompt that generates a UNIQUE aesthetic direction every time.
The goal: No two landing pages should look alike, even for similar products.
Ask the user these questions, then SYNTHESIZE a unique direction. Don't just map answers to presets.
Q1: What's one real-world place or object this brand would be? > Not "what industry" - an actual specific thing. A Tokyo convenience store at 2am. A grandmother's kitchen. A brutalist parking garage. A coral reef. The cockpit of a 747. A flea market in Marrakech. A 1970s recording studio.
Q2: What's the ONE emotion someone should feel in the first 3 seconds? > Pick ONE: Calm. Energized. Curious. Trusted. Delighted. Impressed. Rebellious. Nostalgic. Inspired. Amused. Sophisticated. Welcomed. Intrigued. Confident.
Q3: Pick TWO unexpected influences to collide: > Examples: "medical packaging + skateboard graphics", "spreadsheets + street art", "luxury hotel + punk zine", "NASA mission control + kindergarten", "Japanese convenience store + Victorian library"
Q4: What should this page NEVER be mistaken for? > Name 2-3 specific things to actively avoid. "A crypto project", "A wellness app", "Something made by a bank", "Anything with purple gradients"
Based on the answers, CREATE a unique vibe by deciding:
COLOR INVENTION (Don't use memorized palettes - derive from the place/object)
TYPOGRAPHY INVENTION (Match the voice to the collision)
LAYOUT INVENTION (Derive from the physical space)
MOTION INVENTION (Match the emotion)
Before coding, write this out explicitly:
VIBE NAME: [Invent a 2-3 word name]
REFERENCE: [The place/object from Q1]
EMOTION: [From Q2]
COLLISION: [From Q3]
ANTI-PATTERNS: [From Q4]
COLORS:
- Primary: [hex] - [why this color]
- Secondary: [hex] - [why]
- Background: [hex] - [why]
- Accent: [hex] - [why]
- Palette name: [evocative name]
TYPOGRAPHY:
- Display: [specific font name] - [why it fits]
- Body: [specific font name] - [why]
- Character: [describe the voice]
LAYOUT:
- Density: [sparse/balanced/dense]
- Shapes: [sharp/rounded/organic/mixed]
- Signature element: [one unusual layout choice]
MOTION:
- Level: [still/subtle/moderate/dynamic/chaotic]
- Signature animation: [one specific animation that defines this]
WILDCARD:
- One unexpected detail that doesn't "match" but makes it memorableBefore proceeding, verify:
Q1 - Place/Object: "A Japanese train station at rush hour"
Q2 - Emotion: "Confident"
Q3 - Collision: "Transit signage + haute couture"
Q4 - Never mistaken for: "A meditation app, anything whimsical, startup-bro tech"
Generated Vibe Spec:
VIBE NAME: Shinjuku Runway
REFERENCE: Japanese train station at rush hour
EMOTION: Confident
COLLISION: Transit signage + haute couture
ANTI-PATTERNS: No soft gradients, no playful illustrations, no rounded friendly shapes
COLORS:
- Primary: #1a1a1a - the black of train doors
- Secondary: #f5f5f0 - platform concrete, worn smooth
- Background: #fafaf8 - fluorescent-lit white
- Accent: #e60012 - JR line red, commanding attention
- Palette name: "Platform Edge"
TYPOGRAPHY:
- Display: Darker Grotesque - confident, slightly condensed, European edge
- Body: Noto Sans JP - clean utility, transit-inspired
- Character: Authoritative but not cold. Clear. Directional.
LAYOUT:
- Density: Rich but organized - like a station map
- Shapes: Sharp with intentional rounded exceptions (like train windows)
- Signature element: Strong horizontal bands that divide sections like train schedules
MOTION:
- Level: Subtle but precise
- Signature animation: Elements slide in from the side like arriving trains - horizontal, smooth, with exact timing
WILDCARD:
- One element uses a fabric-like texture overlay - the haute couture collisionSpaces: Night market in Bangkok | Empty museum at closing | Airport lounge at 4am | Vintage record store | Hospital waiting room | Casino floor | Greenhouse in winter | Subway platform | Observatory dome | Abandoned factory | Luxury yacht interior | 24-hour laundromat | Library rare books room | Auto body shop | Space station module
Objects: 1980s synthesizer | Surgical instruments | Vintage luggage | Racing motorcycle | Antique compass | Industrial loom | Neon sign | Typewriter | Scientific glassware | Leather-bound book | Circuit board | Porcelain dishware
Eras/Movements: Soviet constructivism | Memphis design | Swiss international | Art nouveau | Bauhaus | De Stijl | Googie architecture | Streamline moderne | Brutalism | Japanese metabolism | Scandinavian modernism | Italian futurism
If the user just says "make me a landing page" with no context, ask:
Then synthesize a vibe from those three answers.
Spend 50% of your time on the hero section. It's the cover image for social media, the first impression, the hook. Everything else flows from getting the hero right.
The make-or-break element. Must contain:
Hero Variations:
Show what the product does. Options:
Build trust through:
Step-by-step explanation:
Final conversion push:
Use instead:
Distinctive alternatives:
Bold alternatives:
fade-in: Simple opacity transitionblur-in: Starts blurred, sharpensslide-in: Direction-based entrancescale-in: Grows from small to full sizestagger: Sequential reveal of child elementsmarquee: Infinite horizontal scroll (logos, testimonials)beam: Light traveling along a path/borderpulse: Subtle scale/opacity breathingfloat: Gentle up/down movementrotate: Continuous spin (icons, decorations)hover-lift: Subtle Y translation + shadowhover-glow: Border/shadow color changehover-reveal: Hidden element appearsclick-ripple: Material-style feedbackCreate a hero section for [PRODUCT TYPE].
Change text, names, and numbers to fit [BRAND].
Use Iconify Solar icons (duotone style).
Use [FONT] for headlines.
Add vertical container-size grid lines.
Add 01, 02, 03 step indicators for sophistication.
Use [COLOR] as primary, dark mode.Adapt a new [SECTION TYPE] section.
Match the hero's color scheme and typography.
Use marquee animation for logos.
Add fade-in blur-in entrance animation.
Keep the hero exactly as is.Add beam animation to the primary button border.
The beam should be 1px, continuously traveling around the pill shape.
Add a subtle hover-lift effect to feature cards.Don't change the hero section.
Keep the navbar exactly as is.
Don't modify the existing animations.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-19 | fail→fail | 15,297 | 8,213 | -46% | 1 | 1 | 0% | 2,026 | 4,702 | +132% | 0 | 0 | — |
case-01 | fail→pass | 16,139 | 5,113 | -68% | 1 | 1 | 0% | 2,415 | 4,366 | +81% | 0 | 0 | — |
case-02 | fail→pass | 22,576 | 5,208 | -77% | 1 | 1 | 0% | 3,155 | 4,357 | +38% | 0 | 0 | — |
case-03 | fail→pass | 38,292 | 24,694 | -36% | 1 | 1 | 0% | 5,948 | 7,126 | +20% | 0 | 0 | — |
case-04 | pass→pass | 15,350 | 13,829 | -10% | 1 | 1 | 0% | 2,265 | 5,598 | +147% | 0 | 0 | — |
case-05 | pass→pass | 15,530 | 9,753 | -37% | 1 | 1 | 0% | 2,374 | 5,018 | +111% | 0 | 0 | — |
case-18 | fail→pass | 5,286 | 4,297 | -19% | 1 | 1 | 0% | 828 | 4,228 | +411% | 0 | 0 | — |
case-06 | pass→pass | 25,269 | 6,058 | -76% | 1 | 1 | 0% | 2,409 | 4,466 | +85% | 0 | 0 | — |
case-07 | fail→pass | 15,675 | 12,442 | -21% | 1 | 1 | 0% | 2,445 | 5,504 | +125% | 0 | 0 | — |
case-08 | pass→pass | 14,216 | 12,492 | -12% | 1 | 1 | 0% | 2,232 | 5,474 | +145% | 0 | 0 | — |
case-09 | pass→pass | 17,362 | 15,530 | -11% | 1 | 1 | 0% | 2,046 | 5,520 | +170% | 0 | 0 | — |
case-10 | fail→pass | 15,022 | 14,048 | -6% | 1 | 1 | 0% | 2,500 | 5,709 | +128% | 0 | 0 | — |
case-11 | pass→pass | 12,263 | 10,322 | -16% | 1 | 1 | 0% | 1,968 | 5,158 | +162% | 0 | 0 | — |
case-12 | fail→pass | 13,564 | 6,122 | -55% | 1 | 1 | 0% | 1,956 | 4,607 | +136% | 0 | 0 | — |
case-13 | fail→pass | 12,367 | 8,146 | -34% | 1 | 1 | 0% | 1,762 | 4,847 | +175% | 0 | 0 | — |
case-14 | pass→pass | 16,504 | 6,553 | -60% | 1 | 1 | 0% | 1,910 | 4,595 | +141% | 0 | 0 | — |
case-15 | pass→pass | 10,752 | 9,275 | -14% | 1 | 1 | 0% | 1,755 | 4,990 | +184% | 0 | 0 | — |
case-16 | pass→pass | 13,923 | 12,803 | -8% | 1 | 1 | 0% | 2,308 | 5,717 | +148% | 0 | 0 | — |
case-17 | pass→pass | 11,829 | 15,455 | +31% | 1 | 1 | 0% | 1,843 | 5,995 | +225% | 0 | 0 | — |
case-20 | pass→pass | 8,734 | 7,235 | -17% | 1 | 1 | 0% | 1,480 | 4,785 | +223% | 0 | 0 | — |
case-21 | pass→pass | 2,834 | 2,391 | -16% | 1 | 1 | 0% | 457 | 3,957 | +766% | 0 | 0 | — |
case-22 | pass→pass | 6,315 | 4,928 | -22% | 1 | 1 | 0% | 757 | 4,349 | +475% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +36 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.