What problem does it solve? Building programmatic video content requires coordinating HTML structure, GSAP animation timing, captions, audio, and scene transitions — a process prone to layout overlaps, broken timelines, and contrast failures. This Skill provides the complete rule set and reference library for authoring HyperFrames compositions where HTML is the source of truth for video. ## Core Features & Use Cases - HTML Video Compositions: Build scenes with data-attribute timing, GSAP timelines registered to window.__timelines, and deterministic animation rules (no Math.random, no repeat: -1, no async timeline construction). - Captions, TTS, and Audio-Reactive Visuals: Sync per-word captions to transcripts, generate Kokoro-82M voiceovers, and map pre-extracted frequency bands to GSAP properties. - Visual Identity System: Enforce a DESIGN.md gate before writing HTML, with 8 named visual styles (Swiss Pulse, Velvet Standard, Data Drift, and more) plus palette and typography references. - Quality Verification: Run animation-map.mjs to audit choreography, collisions, and dead zones, and contrast-report.mjs for WCAG AA compliance. - Use Case: Ask for a 30-second product launch video with animated stats, synced captions, and scene transitions, and receive a complete lint-clean composition ready for rendering. ## Quick Start Create a HyperFrames video composition with an animated title card, three data stats, and captions synced to my voiceover audio.