In the Skill's Own Terms
This skill enforces award-winning frontend design engineering that breaks standard LLM layout biases.
It uses Python-driven randomization to prevent repetitive layouts, mandates AIDA page structure with massive section spacing, limits hero headings to 2-3 lines using ultra-wide containers, builds mathematically perfect gapless bento grids with grid-flow-dense, implements advanced GSAP ScrollTriggers for pinning, stacking, and scrubbing animations, embeds inline micro-images directly in typography, and bans meta-labels like 'SECTION 01' and narrow containers that cause 6-line text wraps. Every page requires a design_plan verification block before outputting UI code.
What it produces
- React/UI code with GSAP ScrollTrigger animations and mathematically perfect grids
How It Works
- 01Simulate Python Randomization from Prompt Seed
Before writing any UI code, simulate a Python script using the character count of the user prompt to deterministically select one hero architecture, one typography stack, three component architectures, and two GSAP paradigms.
- 02Structure Page with AIDA Framework
Build the page following Attention (cinematic hero), Interest (high-density bento grid), Desire (GSAP scroll effects), and Action (massive CTA and footer).
- 03Build Gapless Bento Grids with Mathematical Precision
Use Tailwind's grid-flow-dense on every bento grid and mathematically verify that col-span and row-span values interlock perfectly with zero empty voids or missing corners.
- 04Implement Advanced GSAP ScrollTrigger Animations
Write real GSAP with ScrollTrigger for scroll pinning with split layouts, image scale and fade effects, scrubbing text reveals with sequential opacity changes, and dynamic card stacking.
- 05Output Mandatory Design_plan Verification
Before writing any React or UI code, output a design_plan block containing Python RNG execution, AIDA check, hero math verification for 2-3 line guarantee, bento density proof, and label sweep confirmation.