メニュー
Create clear, production-grade frontend interfaces with high design quality. Use when building web components, pages, or applications that need polished, professional UI. Focuses on clarity, consistency, and functional beauty over flashiness.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Run the GitHub issue triage tool for Redux repos. This skill should be used when triaging GitHub issues, generating triage reports, or understanding triage output for redux-toolkit, react-redux, reselect, redux, or immer repositories.
Create architecture review documents to understand and document existing codebases. Use when onboarding to an unfamiliar codebase, conducting architecture audits, or creating reference documentation for team members.
Document coding conventions, patterns, and guidelines for an existing codebase. Use when onboarding to a new project to understand its coding style, or when creating style documentation for team reference.
Design patterns for data-dense dashboard UIs with dark themes. Use when building tables, charts, status panels, and data visualization components. Focuses on readability, information density, and consistent design systems.
Create detailed development plans with task lists and progress tracking for major projects. Use for multi-phase work requiring detailed task breakdown, progress tracking, and living documentation of decisions and learnings.
Manage tasks via dex CLI. Use when breaking down complex work, tracking implementation items, or persisting context across sessions.
| name | frontend-design |
| description | Create clear, production-grade frontend interfaces with high design quality. Use when building web components, pages, or applications that need polished, professional UI. Focuses on clarity, consistency, and functional beauty over flashiness. |
Adapted from Impeccable by Paul Bakaus (Apache 2.0 license). Modified for dashboard/tool UI: emphasis shifted from bold creative expression to clarity, information density, scanability, and consistent design systems.
This skill guides creation of polished, production-grade frontend interfaces that avoid generic "AI slop" aesthetics. Implement real working code with careful attention to design details and intentional choices.
Establish a clear design direction before writing code:
CRITICAL: Choose a clear direction and execute it with precision. The key is intentionality: every color, spacing value, and typographic choice should trace back to a deliberate decision, not a default.
Then implement working code that is:
→ Consult typography reference for scales, pairing, and loading strategies.
Choose fonts appropriate to the interface's purpose. For data-heavy tools, readability and numeric clarity matter more than personality.
DO: Use a modular type scale with fluid sizing (clamp)
DO: Vary font weights and sizes to create clear visual hierarchy
DO: Use font-variant-numeric: tabular-nums for data columns
DON'T: Use overused fonts—Inter, Roboto, Arial, Open Sans, system defaults — unless they genuinely serve the purpose (system fonts are fine for tools)
DON'T: Use monospace typography as lazy shorthand for "technical/developer" vibes
DON'T: Put large icons with rounded corners above every heading—they rarely add value and make sites look templated
→ Consult color reference for OKLCH, palettes, and dark mode.
Commit to a cohesive palette with clear semantic roles. In data-dense UIs, color discipline is even more critical — every color should mean something specific.
DO: Use modern CSS color functions (oklch, color-mix, light-dark) for perceptually uniform, maintainable palettes DO: Tint your neutrals toward your brand hue—even a subtle hint creates subconscious cohesion DO: Define semantic color tokens (positive, negative, warning, info, neutral) and enforce strict usage DON'T: Use gray text on colored backgrounds—it looks washed out; use a shade of the background color instead DON'T: Use pure black (#000) or pure white (#fff)—always tint; pure black/white never appears in nature DON'T: Use the AI color palette: cyan-on-dark, purple-to-blue gradients, neon accents on dark backgrounds DON'T: Use gradient text for "impact"—especially on metrics or headings; it's decorative rather than meaningful DON'T: Default to dark mode with glowing accents—it looks "cool" without requiring actual design decisions DON'T: Reuse semantic colors (green/red) for unrelated meanings — green for "profit" AND "enabled" AND "fresh" creates confusion
→ Consult spatial reference for grids, rhythm, and container queries.
Create visual rhythm through varied spacing. In dense UIs, consistent spacing systems are even more important — they prevent the "wall of data" effect.
DO: Create visual rhythm through varied spacing—tight groupings, generous separations DO: Use fluid spacing with clamp() that breathes on larger screens DO: Use a consistent spacing scale (4pt base: 4, 8, 12, 16, 24, 32, 48, 64) DON'T: Wrap everything in cards—not everything needs a container DON'T: Nest cards inside cards—visual noise, flatten the hierarchy DON'T: Use identical card grids—same-sized cards with icon + heading + text, repeated endlessly DON'T: Use the hero metric layout template—big number, small label, supporting stats, gradient accent DON'T: Center everything—left-aligned text with asymmetric layouts feels more designed DON'T: Use the same spacing everywhere—without rhythm, layouts feel monotonous
DO: Use intentional, purposeful decorative elements that reinforce the interface's identity DO: Use subtle elevation differences to create depth hierarchy on dark themes DON'T: Use glassmorphism everywhere—blur effects, glass cards, glow borders used decoratively rather than purposefully DON'T: Use rounded elements with thick colored border on one side—a lazy accent that almost never looks intentional DON'T: Use sparklines as decoration—tiny charts that look sophisticated but convey nothing meaningful DON'T: Use rounded rectangles with generic drop shadows—safe, forgettable, could be any AI output DON'T: Use modals unless there's truly no better alternative—modals are lazy
→ Consult motion reference for timing, easing, and reduced motion.
In tool UIs, motion should be functional — communicating state changes and providing feedback, not creating spectacle.
DO: Use motion to convey state changes—entrances, exits, feedback DO: Use exponential easing (ease-out-quart/quint/expo) for natural deceleration DO: For height animations, use grid-template-rows transitions instead of animating height directly DON'T: Animate layout properties (width, height, padding, margin)—use transform and opacity only DON'T: Use bounce or elastic easing—they feel dated and tacky; real objects decelerate smoothly
→ Consult interaction reference for forms, focus, and loading patterns.
Make interactions feel fast. Use optimistic UI—update immediately, sync later.
DO: Use progressive disclosure—start simple, reveal sophistication through interaction (basic options first, advanced behind expandable sections; hover states that reveal secondary actions) DO: Design empty states that teach the interface, not just say "nothing here" DO: Make every interactive surface feel intentional and responsive DON'T: Repeat the same information—redundant headers, intros that restate the heading DON'T: Make every button primary—use ghost buttons, text links, secondary styles; hierarchy matters
→ Consult responsive reference for mobile-first, fluid design, and container queries.
DO: Use container queries (@container) for component-level responsiveness DO: Adapt the interface for different contexts—don't just shrink it DON'T: Hide critical functionality on mobile—adapt the interface, don't amputate it
→ Consult ux-writing reference for labels, errors, and empty states.
DO: Make every word earn its place DON'T: Repeat information users can already see
Critical quality check: If you showed this interface to someone and said "AI made this," would they believe you immediately? If yes, that's the problem.
A well-designed interface should make someone ask "how was this made?" not "which AI made this?"
Common AI fingerprints to avoid — review the DON'T guidelines above. They are the telltale signs of AI-generated work from 2024-2025:
Match implementation complexity to the design's needs. Data-heavy tools need precise alignment, consistent token usage, and clear information hierarchy. Every CSS value should reference a token. Every color should have a semantic role.
Build design systems, not pages. Components should be consistent, composable, and token-driven so theming and iteration are cheap.
