What problem does it solve?
It helps teams avoid generic, “AI-slop” frontend output by enforcing brief-first design direction, consistent systems, and pre-flight checks so the result reads like a real, intentional design.
Core Features & Use Cases
- Brief-first design inference: reads page kind, audience, vibe words, and constraints to set the correct design direction before generating UI.
- Anti-default guardrails: prevents common templated patterns (AI-purple glows, duplicated CTA intent, centered-hero clichés, eyebrow overuse, and unsafe typography/color choices).
- Design system selection logic: chooses official component libraries when the brief maps to one; otherwise builds with native CSS/Tailwind conventions.
- Engineering-ready conventions: enforces Next.js/RSC + client-component isolation rules, icon-family consistency, responsiveness mechanics, and accessibility/contrast checks.
- Visual asset strategy: prioritizes generating or sourcing real images and logos (rather than fake screenshots or text-only trust walls) to keep the page authentic.
Quick Start
Use the design-taste-frontend skill to generate a landing page UI from a brief describing the target audience, page type, and desired vibe, while following its pre-flight and anti-slop rules.