award-design
Award Design
Build websites that score 8+ on Awwwards. AI-generated designs are immediately recognizable to experienced judges and score poorly — this skill exists to beat that.
Workflow
- Understand the brief: What is being built? For whom? What must it communicate? What's the one thing someone will remember? If
-u <url>was passed, readreferences/brand-extraction.mdfirst and reverse-engineer a DESIGN.md observation from the live site — that observation seeds the archetype recommendation in step 2, it doesn't replace the brief. - Recommend an archetype: Analyze the brief and recommend the single best archetype from the table below. Present it with:
- The archetype name and its signature trait
- Why it fits this brief specifically (not generic reasoning)
- The default Density/Variance/Motion scores
- Ask the user to validate before proceeding. Do not continue until confirmed.
- If the user wants to explore alternatives, present all 9 archetypes as a compact list:
- Minimalist — extreme whitespace, typography carries everything
- Brutalist — thick borders, flat colors, deliberate anti-polish
- Editorial — serif + sans-serif pairing, magazine grids, pull quotes
- Bold / Maximal — organized chaos, kinetic typography as art
- Immersive / Cinematic — full-screen video, WebGL 3D, dark backgrounds
- Experimental — bespoke, no template, creative coding, mixed media
- Corporate Luxury — quiet sophistication, custom serifs, generous whitespace
- Bento / Card — modular tiles, container queries, self-contained units
- Spatial Organic — dimensional depth, organic shapes, tactile textures
- The user is free to pick any archetype — the recommendation is guidance, not a constraint.
- See
references/exemplars.mdfor 2–4 real-world brands per archetype. Share 2–3 alongside the recommendation — exemplars travel faster than prose and give the user a concrete "that feel" to react to.
- Read archetype reference: Once the archetype is confirmed, read its reference file from the table below.
- Calibrate atmosphere: Set Density, Variance, and Motion scores using the Atmosphere Calibration table. Adjust ±2 from defaults based on the brief. Present the calibrated scores to the user for validation.
- Load foundations: Read
references/foundations.mdfor cross-cutting technical implementation (typography systems, color theory, animation toolkit, performance, UX quality, accessibility). - Produce DESIGN.md: If the project has no
DESIGN.md, create one following the Google DESIGN.md open standard — YAML frontmatter with design tokens (colors,typography,rounded,spacing,components) plus eight ordered prose sections (Overview, Colors, Typography, Layout, Elevation & Depth, Shapes, Components, Do's and Don'ts). Record the calibrated atmosphere scores as prose in the Overview section — the spec does not define atmosphere tokens. If the/design-systemskill is installed, follow itsreferences/design-md-spec.mdand the example files for shape; otherwise, the sections listed in the archetype reference files cover the same ground. Validate the finished file with/design-system audit DESIGN.md(preferred — human-readable report with fix proposals per finding). If/design-systemis unavailable, fall back to the raw CLI:npx @google/design.md lint DESIGN.md. Broken token references (broken-ref) and contrast violations below WCAG AA are caught here. If the user hands you an existing legacy DESIGN.md with the Stitch 9-section format, recommend/design-system migrate <path>first to port it, then resume. Every applicable section must be complete — the/design-systemskill governs this file for all future UI changes, and incomplete sections create token gaps that agents fill with defaults. - Design with intent: Every visual choice serves communication. One signature unforgettable moment outperforms scattered effects everywhere.
- Production hardening: When implementation touches video, scroll-driven cinematic reveals, or full-screen heroes on mobile browsers, read
references/production-hardening.md. Most patterns are cross-browser (viewport units, scroll-restoration, autoplay belt-and-suspenders, fail-safe reveal logic, proportional layout) with iOS Safari flagged as the sharpest test case. Each section states its scope — genuinely iOS-only rules are marked. Skip if the project is desktop-only with no video or scroll choreography. - Validate: Read
references/anti-patterns.mdand check the design against it — axiomatic rejections first (any hit is stop-and-fix), then AI tells, performance failures, UX anti-patterns. Cross-checkreferences/foundations.mdUX Quality and Accessibility sections. For a calibrated score with an actionable punch list, runreferences/audit-rubric.md(7 categories, 0–10 each, P0/P1 fixes with CSS snippets). Verify against the judging criteria below. - Visual review (optional): If
dev-browserCLI is available, screenshot key states (hero, mobile, signature interaction, dark mode) and iterate. If not available, suggest installing it fromhttps://github.com/SawyerHood/dev-browser— the skill works without it, but visual verification catches issues that code review alone misses.
Changing archetype mid-project
If the user wants to switch archetypes after initial selection (during design or even after implementation has started):
- Confirm the new archetype choice
- Read the new archetype's reference file
- Recalibrate atmosphere — the new archetype has different default scores. Present the recalibrated scores for validation
- Regenerate DESIGN.md — the entire file must be rewritten from the new archetype's foundations. Do not patch the old file — archetype tokens are deeply interconnected
- If code already exists, flag which components need updating based on the token diff between old and new DESIGN.md
Combining archetypes (remix)
If the brief refuses to pick a single archetype — "Linear rigor but Anthropic warmth", "Brutalist character for luxury clients", a creative studio serving enterprise — read references/remixing.md. It gives an arbitration framework (parent DNA percentage, 7 rules that pick one parent per dimension, one-paragraph identity declaration) so the remix reads as a third coherent brand rather than a blend. Default is still to pick one archetype; reach for a remix only when a single archetype leaves the brief unsatisfied after two attempts.
Archetype Selector
| Archetype | Signature | Ideal for | Reference |
|---|---|---|---|
| Minimalist | Extreme whitespace, 2-3 colors, typography carries everything | SaaS, luxury, architecture, portfolios | references/minimalist.md |
| Brutalist | Thick borders, flat colors, deliberate anti-polish | Creative agencies, indie tech, streetwear | references/brutalist.md |
| Editorial | Serif + sans-serif pairing, magazine grids, pull quotes | Media, fashion, cultural institutions | references/editorial.md |
| Bold / Maximal | Organized chaos, 4-6+ colors, kinetic typography as art | Entertainment, music, Gen Z brands | references/bold-maximal.md |
| Immersive / Cinematic | Full-screen video, WebGL 3D, dark backgrounds | Automotive, luxury, gaming, museums | references/immersive-cinematic.md + references/production-hardening.md |
| Experimental | Bespoke, no template, creative coding, mixed media | Developer portfolios, art institutions | references/experimental.md |
| Corporate Luxury | Quiet sophistication, custom serifs, generous whitespace | High-end fashion, hotels, jewelry, wealth | references/corporate-luxury.md |
| Bento / Card | Modular tiles, container queries, self-contained units | SaaS product pages, feature comparisons | references/bento-card.md |
| Spatial Organic | Dimensional depth, organic shapes, tactile textures, native APIs | Sustainability brands, post-2025 creative studios | references/spatial-organic.md |
Selection guide: Match the archetype to the brand's personality, not to what's trending. A luxury hotel should never be brutalist. A creative agency should never be generic minimalist. When in doubt, the brief's tone decides.
Handoff to design-system: The DESIGN.md produced by this skill becomes the single source of truth for all future UI work. Once created, the /design-system skill governs it — enforcing tokens, handling updates, and preventing drift. This skill (award-design) is for initial creation and complete re-architecting only. Token-level changes (adjust a color, tweak spacing) go through /design-system.
Atmosphere Calibration
After selecting an archetype, calibrate its atmosphere on three axes (1–10). This makes design choices measurable rather than intuitive, and prevents drift during implementation.
| Axis | 1–3 | 4–6 | 7–10 |
|---|---|---|---|
| Density | Gallery airy — generous whitespace, few elements per viewport | Balanced — clear hierarchy with moderate content | Cockpit dense — information-rich, tight spacing |
| Variance | Predictable — symmetric grids, uniform spacing, expected flow | Structured surprise — asymmetric grids, varied rhythm | Artsy chaotic — broken grids, overlapping zones, rule-breaking |
| Motion | Static — minimal transitions, opacity-only reveals | Purposeful — scroll-triggered sequences, hover states | Cinematic — continuous animation, parallax depth, WebGL layers |
Default scores per archetype (adjust ±2 based on brief):
| Archetype | Density | Variance | Motion |
|---|---|---|---|
| Minimalist | 2 | 3 | 3 |
| Brutalist | 4 | 7 | 3 |
| Editorial | 5 | 5 | 4 |
| Bold / Maximal | 6 | 8 | 8 |
| Immersive / Cinematic | 3 | 6 | 9 |
| Experimental | 5 | 9 | 7 |
| Corporate Luxury | 2 | 4 | 5 |
| Bento / Card | 7 | 4 | 4 |
| Spatial Organic | 4 | 6 | 6 |
Use these scores to resolve design ambiguity: "Should this section have more whitespace?" → check Density score. "Should I break the grid here?" → check Variance score. "Does this element need scroll animation?" → check Motion score. Record the final calibrated scores in the project's DESIGN.md.
Judging Criteria
Awwwards evaluates: Design 40% · Usability 30% · Creativity 20% · Content 10%. Honorable Mention requires 6.5+. SOTD requires ~7.5+.
What separates 8+ from 6-7:
- One signature unforgettable interaction (not scattered micro-animations everywhere)
- Mobile reconsidered, not just responsive breakpoints bolted on
- Complex visuals that load fast on mid-range devices (LCP < 1.5s)
- Real content with genuine photography — no stock
- Scroll as narrative — content unfolds with purpose and pacing
- Precise animation choreography (timing, easing, sequencing)
Strategic path: CSSDA first (most accessible, WOTD > 8.0) → FWA (rewards experimental boldly) → Awwwards (highest bar). Best submission months: Feb-Apr, Sep-Nov.
Implementation Checklist
After selecting an archetype, calibrating atmosphere, and reading its reference:
- Establish the design system first (CSS custom properties: typography scale, color palette, spacing tokens)
- Build mobile-first, then enhance for desktop
- Apply UX quality rules from foundations (touch targets, safe areas, form behavior, deep-linking)
- Add animation last — choreograph deliberately, don't scatter. Check Motion score before adding effects
- The one-interaction test: if you remove all but one animation, which one stays? That's your signature moment
- Validate: Lighthouse 90+ on Performance and Accessibility, test on mid-range devices
- If
dev-browseris available: screenshot hero, mobile viewport, and signature interaction — compare against archetype expectations and iterate
More from coroboros/agent-skills
markitdown
Convert any document to Markdown with Microsoft's `markitdown` CLI — PDF, Word, Excel, PowerPoint, HTML, CSV, JSON, XML, ZIP, EPub, images (OCR/EXIF), audio (transcription), and YouTube URLs. Use whenever the user wants to extract text from a binary document, transcribe audio, OCR an image, scrape a YouTube transcript, or pre-process a file for an LLM context window — even when they just say "convert this pdf", "what's in this docx", "transcribe this mp3", or "get the text out of this".
14design-system
Govern the DESIGN.md — Google's open standard for design tokens (YAML frontmatter + eight prose sections). Auto-activates during UI edits to enforce token-only sourcing for colors, typography, spacing, and corner radius. Also exposes six CLI-backed subcommands — audit (lint + fix proposals), diff (regression check), export (Tailwind / DTCG), spec (canonical spec emission), migrate (port from legacy Stitch format), init (minimal scaffold). When a UI/UX change is requested, DESIGN.md is updated first, audited, then code propagates.
3apex
Systematic implementation using APEX methodology (Analyze-Plan-Execute-eXamine) with parallel subagents and self-validation. Use when implementing features, fixing bugs, or making code changes that benefit from structured workflow.
1spec
Transform ideas into structured execution specs with prioritized workstreams, complexity ratings, dependencies, and acceptance criteria. Use this whenever the user needs to decompose work before building — even when they don't explicitly say "spec" (e.g. "break this down", "plan this out", "create issues for", "what are the steps"). The bridge between thinking and building.
1oneshot
Single-pass feature implementation using Explore → Code → Test. Ships focused changes at maximum speed, with a built-in circuit breaker that stops and recommends `/apex` or `/spec` when the task turns out more complex than it looked. Use this whenever the user wants a quick win on a single, focused task — even when they don't say "oneshot" (e.g. "just", "quickly", "small change", "#42", or a GitHub issue URL for a small fix).
1agent-creator
Create, configure, and orchestrate Claude Code subagents — specialized Claude instances with focused roles and limited tool access. Covers YAML frontmatter (name, description, tools, model, permissions, hooks, MCP servers), system prompt design, tool restriction patterns, background execution, and multi-agent orchestration. Use whenever the user mentions subagents, delegation, specialists, agent configs, `.claude/agents/`, the `/agents` command, or wants to parallelize work — even when they just say "background agent" or "delegate this".
1