Ai ui ux motion engine
Skill OpaceDigitalAgency/skills/skills/ai-ui-ux-motion-engine
Open AI Agent Skills for website design, UI/UX, motion graphics, 3D web, accessibility and future SEO workflows across Codex, Claude Code, Cursor, Copilot, Antigravity, Gemini, Windsurf and more.
npx -y skills add OpaceDigitalAgency/skills --skill ai-ui-ux-motion-engineAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- 11 days oldThe repository was created 11 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Design, redesign, audit and validate distinctive production websites using reference extraction, design systems, purposeful motion, scroll interactions, media pipelines and multi-pass refinement. Use for landing pages, product sites, portfolios, interactive storytelling, UI/UX improvement, video- or screenshot-led recreation, cinematic scroll effects, motion graphics, hero refinement, design-system extraction, and anti-generic visual polish in Astro, React, Next.js, Vue, Svelte or static HTML/CSS. Preserve the existing stack and require accessibility, responsive, performance and regression evidence.
SKILL.md
7.8 KB, ~1.4k tokens by cl100k_base, as published. Nobody here has run it
AI UI/UX Motion Engine
Create a project-specific interface from evidence, not a generic component recipe. Treat references as structural and interaction inputs, never as permission to clone branding, copy or protected assets.
Non-negotiable rules
- Read every applicable
AGENTS.md, project source-of-truth/status document, design system and current working-tree change before editing. - Define the bounded component, acceptance criteria, dependencies and validation plan.
- Preserve the existing framework, package manager, content model and design conventions unless the user explicitly authorises a migration.
- Label facts, source observations, inferences, proposals and unknowns separately.
- Use motion to explain hierarchy, state, continuity or spatial relationships. Reject decorative motion that delays content or weakens comprehension.
- Keep primary content and actions usable without JavaScript and when
prefers-reduced-motion: reduceis active. - Never claim reference parity, WCAG conformance, performance, cross-browser support or production readiness without direct evidence.
- Never copy brand identity, proprietary copy, source code or distinctive assets from a reference.
- Never install or configure a paid third-party service without user authority. Keep credentials out of repositories and output.
Workflow
1. Establish the brief and baseline
Record:
- goal and primary conversion;
- audience and decision context;
- required routes, states and interactions;
- content voice and evidence constraints;
- target devices, input modes and browsers;
- framework, build and release boundaries;
- baseline build, test and representative screenshots.
Read workflow.md for the complete phase gates.
2. Extract references
For each screenshot, recording or site:
- map section order, viewport geometry and content density;
- identify typography roles, colour roles, spacing, shape and image treatment;
- record entrance, scroll, hover, drag and state-transition behaviour;
- distinguish reusable principles from identity-specific material;
- combine at least two independent inputs when the user requests a mashup.
For a local recording, run:
bash scripts/extract-reference-frames.sh <video> <output-directory>
Read media-pipeline.md before processing video, image sequences or generated assets.
3. Commit to one design direction
Write a seven-line direction:
- product goal and audience;
- tone in two or three concrete adjectives;
- page composition and hierarchy;
- typography strategy;
- colour, surface and image strategy;
- motion strategy and reduced-motion equivalent;
- one signature differentiator.
Reject “clean and modern” unless every line above makes it implementable. Generate alternatives only for a material decision such as the hero, navigation model or interaction language; keep the rest fixed while comparing.
4. Select the motion architecture
Choose the lightest mechanism that expresses the intended relationship:
- CSS transitions/keyframes for local state and entrance changes;
- Intersection Observer for one-shot reveals;
- native scroll-driven animation when support and fallback are acceptable;
- a small framework motion library when the project already uses it;
- GSAP only for coordinated pinning/timelines that simpler primitives cannot express;
- video scrubbing for photographic or generated camera movement;
- canvas frame sequences only when seeking quality or device support requires them;
- WebGL/3D only when real-time depth materially improves the product story.
For a watch-style or product-inspection effect made from prompts and reference images, do not default to CAD, GLB or WebGL. Route first to the generated product scrubber: approve a clean still, generate a controlled image-to-video camera move, chain clips from exact end frames when needed, then deliver it as scroll-linked video or a frame sequence. Use real-time 3D only when the visitor must freely manipulate the object or inspect arbitrary viewpoints.
Read motion-patterns.md, generated-product-scrubber.md and framework-recipes.md before implementation.
5. Implement in passes
- Architecture: semantic HTML, section order, content and primary action.
- Macro design: grid, scale, rhythm, typography, colour and media.
- Motion: one interaction at a time with a static/reduced-motion equivalent.
- Micro states: hover, focus-visible, active, loading, empty, error and success.
- Responsive composition: mobile, tablet, desktop and zoom.
- Editorial pass: remove filler, unsupported claims and repeated copy.
Keep the primary offer within the first two useful viewports when that supports the page goal; do not apply this as a universal rule to editorial or narrative experiences.
6. Validate every bounded component
After each component:
- test pointer, keyboard and touch behaviour;
- inspect reduced motion;
- check layout at the project’s required viewports;
- check console and runtime errors;
- run the narrowest relevant automated checks.
Then run the project’s complete regression baseline. Run:
node scripts/audit-motion-safety.mjs <project-directory>
node scripts/validate-package.mjs
Read accessibility-performance.md and verification.md for acceptance gates.
Reference routing
- Use workflow.md for phase outputs and decision gates.
- Use motion-patterns.md for interaction architecture and safe implementation patterns.
- Use generated-product-scrubber.md for prompt/reference-image product films, end-frame chaining, scroll scrubbing and acceptance tests.
- Use media-pipeline.md for recording capture, frame extraction, video/image generation and asset delivery.
- Use framework-recipes.md for Astro/static, React/Next, Vue and Svelte decisions.
- Use research-and-extraction.md for competitor and design-system research.
- Use tool-connections.md only when external media or research tools are needed.
- Use accessibility-performance.md for motion safety, Core Web Vitals and responsive requirements.
- Use prompts.md for reusable task payloads.
- Use source-coverage.md to understand what was derived from the four source videos and what was independently strengthened.
- Use verification.md before claiming completion.
Completion report
Report:
- selected direction and signature differentiator;
- references used and what was extracted rather than copied;
- files and components changed;
- motion added and its reduced-motion/static equivalent;
- tests, builds, viewports and interaction states checked;
- remaining unknowns, external gates and release status.
Gives 0 of the 12 instructions most css styling skills give in ~1.4k tokens
Counted across 586 of the 596 authors here whose files we hold, read 2026-08-06
- avoid excessive centered layoutsin 55 of 586, across 12 files
- bundle code into single HTML filein 54 of 586, across 14 files
- Respect prefers-reduced-motion user settingsin 52 of 586, across 35 files
- avoid purple gradientsin 51 of 586, across 11 files
- avoid uniform rounded cornersin 51 of 586, across 11 files
- avoid Inter fontin 51 of 586, across 11 files
- edit generated files to develop artifactin 50 of 586, across 10 files
- animate only transform and opacity propertiesin 43 of 586
- Make touch targets at least 44x44 pixelsin 41 of 586, across 15 files
- Ensure minimum color contrast of 4.5:1in 39 of 586, across 10 files
- use tailwind cssin 39 of 586, across 24 files
- Use SVG icons instead of emojisin 38 of 586, across 11 files
Said here and by no other author read
- read project documentation before editing
- label facts, observations, inferences and proposals separately
- keep content usable without javascript
- record baseline goals, audiences, interactions and constraints
- map references structurally and identify reusable principles
- write a seven-line design direction
Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.