design-intelligence
design-intelligence
AI-powered web design skill that guides both generation and review of web interfaces. Works for any framework (HTML/CSS, React, Vue, Svelte, Next.js, Nuxt, etc.).
Modes
This skill operates in two modes. Identify the request type and apply the corresponding instructions.
GENERATE Mode
Trigger: User wants to design, build, create, or scaffold a web page or UI.
Approach:
- Ask the 3 essential questions (or infer from context)
- Produce a Page Strategy Summary
- Define Section Architecture
- Apply UX Rule Check from
references/WEB_RULES.md - Flag optional enhancements (only after base is solid)
REVIEW Mode
Trigger: User wants to review, audit, critique, or check an existing page or design.
Approach:
- Score each of the 12 rubric categories (1-5)
- Classify issues: Critical / Important / Nice to Improve
- Issue final recommendation: approve / revise / redesign
- Reference
references/REVIEW_RUBRIC.mdfor detailed category guidance
Decision Tree
User request
│
├─ "tasarla / oluştur / build / create" → GENERATE Mode
│ └─ What page type? (landing, dashboard, e-commerce, blog, portfolio, pricing, about)
│
└─ "incele / audit / review / kontrol et / eleştir" → REVIEW Mode
└─ Is this a live URL or code/snippet?
├─ Live URL → navigate, snapshot, score
└─ Code/snippet → analyze directly
GENERATE Mode Instructions
Step 1 — Essential Questions
If not already clear, ask or infer:
- Page type: landing / dashboard / e-commerce / blog / portfolio / pricing / about / other
- Primary goal: sell / inform / convert / onboard / dashboard / profile / other
- Target audience: developers / consumers / B2B / internal / other
Step 2 — Page Strategy Summary
Output:
## Page Strategy Summary
- Page type: [type]
- Page goal: [what the page must accomplish]
- Primary audience: [who uses this page]
- Primary message: [the one thing this page must communicate]
- Primary CTA: [the main action — can be "none" for purely informational pages]
Step 3 — Section Architecture
List ordered sections for the page type. Example for landing page:
1. Hero — single message + primary CTA
2. Trust indicators — logos, stats, social proof
3. Feature bento grid — modular card cluster
4. Results or stats — measurable impact
5. Testimonial or proof — human evidence
6. FAQ — objection handling
7. Final CTA — last conversion push
Example for dashboard:
1. Navigation shell — sidebar or top bar
2. KPI cards — top-level metrics
3. Primary data view — main table/chart
4. Secondary panels — detail, filters, activity
5. Action bar — bulk actions, exports
Step 4 — UX Rule Check
Confirm each rule is satisfied. Reference references/WEB_RULES.md for the full rule text.
| Rule | Status | Notes |
|---|---|---|
| One Primary Message | ✅ | |
| No Hero Slider | ✅ | |
| Modular Layout | ✅ | |
| Mobile-First | ⚠️ | Side nav collapses to hamburger on mobile |
| Performance Targets | ✅ | LCP < 2.5s, INP < 200ms, CLS < 0.1 |
| Accessibility Baseline | ✅ | Semantic HTML, focus states, contrast |
| Functional Motion Only | ✅ | |
| Forbidden Defaults | ✅ | No slider, no carousel, no decoration-only animation |
Step 5 — Optional Enhancements
Only after the base solution is solid. Apply references/WEB_RULES.md Section "Advanced Effects":
- scrollytelling — only for guided narrative (product launches, brand stories)
- 3D/WebGL — only when it adds real explanatory value + has fallback
- AI personalization — only as enhancement layer, never as structural replacement
- video background — only if meaningful + lightweight + contrast-safe
REVIEW Mode Instructions
Step 1 — Gather Information
- Live URL: Navigate, take snapshot, analyze
- Code snippet: Read directly, identify framework and structure
- Image: Describe visually, identify layout patterns
Step 2 — Score 12 Categories
Score each category 1-5 and record justification:
| # | Category | Score (1-5) | Justification |
|---|---|---|---|
| 1 | Message Clarity | ||
| 2 | Hero Focus | ||
| 3 | CTA Clarity | ||
| 4 | Modular Structure | ||
| 5 | Visual Hierarchy | ||
| 6 | Mobile Usability | ||
| 7 | Touch Comfort | ||
| 8 | Performance Safety | ||
| 9 | Accessibility Quality | ||
| 10 | Motion Usefulness | ||
| 11 | Visual Consistency | ||
| 12 | Advanced Effects Justification |
Reference references/REVIEW_RUBRIC.md for detailed scoring guidance per category.
Step 3 — Issue Classification
Classify every issue into one of three tiers:
Critical — damages clarity, usability, accessibility, performance, or conversion path:
- Hero slider hiding primary value
- Unreadable contrast in hero
- Primary CTA unclear or missing
- Broken mobile flow
- No keyboard accessibility
- CLS-causing layout shifts
Important — weakens quality without breaking the page:
- Inconsistent spacing system
- Weak hierarchy in secondary sections
- Unnecessary media weight
- Mixed card patterns
Nice to Improve — polish opportunities:
- Better proof ordering
- Cleaner CTA wording
- Tighter typography rhythm
Step 4 — Final Recommendation
## Final Recommendation: [approve / revise / redesign]
[approve] — No critical issues, page is clear and usable
[revise] — No structural failure, but important issues weaken effectiveness
[redesign] — Structurally confused, hero strategy broken, mobile poor, systemic problems
Page Type Matrix
Rules adapt based on page type. The default rules in references/WEB_RULES.md apply to all types. This matrix highlights per-type emphasis.
| Page Type | Primary Focus | Key Adaptation |
|---|---|---|
| Landing page | Conversion, single message | Hero is critical, CTA placement paramount |
| Dashboard / Admin | Data density, task efficiency | High information density OK, hierarchy by importance |
| E-commerce | Product discovery, conversion | Product hero, gallery, clear pricing, CTA priority |
| Blog / Article | Readability, content first | Typography-first, minimal chrome, distraction-free |
| Portfolio | Work showcase, personality | Visual-first, project cards, clean layout |
| SaaS Pricing | Plan comparison, decision support | Table or grid layout, feature clarity, CTA per plan |
| About / Contact | Trust, human connection | Team, story, clear contact channels |
| Error / Empty states | Clarity, next action | Friendly message, clear CTA, recovery path |
Per-Type Rule Modulations
Dashboard / Admin panel:
- Rule 1 (One Primary Message): Relaxed — data density is expected, but primary KPIs must be immediately visible
- Rule 3 (Modular Layout): Strongly encouraged — dense data must be chunked into cards/tables/panels
- Rule 4 (Hierarchy by Size): Critical — larger blocks for primary metrics, smaller for secondary
E-commerce product page:
- Rule 1 (One Primary Message): Product = hero, single product focus
- Rule 9 (Mobile-First): Critical — mobile commerce dominates
- Rule 2 (No Hero Slider): Product gallery is OK (thumbnails/video), carousel is not
Blog / Article:
- Rule 1 (One Primary Message): Article headline = primary message
- Rule 18 (Typography as Structure): Critical — readability is the product
- Rule 3 (Modular Layout): Article body is linear, sidebar/modules are modular
Forbidden Defaults
Unless explicitly justified and explainable, these are never acceptable:
- Hero sliders / auto-rotating carousels
- Multiple competing primary CTAs in the same focal area
- Decorative-only animation
- Desktop-first layout logic
- Random asymmetry or broken reading order
- Dense text walls without modular structure
- Heavy WebGL on basic pages
- AI-driven navigation changes that reduce predictability
- Unreachable touch targets under 44×44 px on mobile
- Content that breaks below 4.5:1 contrast ratio
- Pages without reduced-motion fallback
Exception Policy
A forbidden default may be broken only if the agent:
- Explains why the exception is necessary
- Describes the concrete benefit it creates
- Confirms it does not harm clarity, performance, or accessibility
If that explanation cannot be made convincingly, keep the default rule.
Quick Reference
GENERATE workflow:
1. Identify page type → 2. Ask essential questions → 3. Page Strategy Summary
→ 4. Section Architecture → 5. UX Rule Check → 6. Optional Enhancements
REVIEW workflow:
1. Gather info (URL or code) → 2. Score 12 categories (1-5)
→ 3. Classify issues → 4. Final recommendation (approve/revise/redesign)
Core reading:
references/WEB_RULES.md— 27+ rules with per-type adaptationsreferences/REVIEW_RUBRIC.md— 12 scoring categories with detailed guidancereferences/EXAMPLES.md— concrete examples per page type
More from fatih-developer/fth-skills
task-decomposer
Break down large, complex, or ambiguous tasks into independent subtasks with dependency maps, execution order, and success criteria. Plan first, then execute step by step. Triggers on 'how should I do this', 'where do I start', 'plan the project', 'break it down', 'implement' or whenever a task involves multiple phases.
24context-compressor
Compress long conversation histories, large code files, research results, and documents by 70% without losing critical information. Triggers when context window fills up, when summarizing previous steps in multi-step tasks, before loading large files into context, or on 'summarize', 'compress', 'reduce context', 'save tokens'.
17multi-brain-debate
Two-round debate protocol where perspectives challenge each other before consensus. Round 1 presents independent positions, Round 2 allows counter-arguments and rebuttals. Produces battle-tested decisions for high-stakes choices.
17multi-brain-score
Confidence scoring overlay for multi-brain decisions. Each perspective rates its own confidence (1-10) with justification. Consensus uses scores as weights, flags low-confidence areas, and surfaces uncertainty explicitly.
15checkpoint-guardian
Automatic risk assessment before every critical action in agentic workflows. Detects irreversible operations (file deletion, database writes, deployments, payments), classifies risk level, and requires confirmation before proceeding. Triggers on destructive keywords like deploy, delete, send, publish, update database, process payment.
14parallel-planner
Analyze multi-step tasks to identify which steps can run in parallel, build dependency graphs, detect conflicts (write-write, read-write, resource contention), and produce optimized execution plans. Triggers on 3+ independent steps, 'speed up', 'run simultaneously', 'parallelize', 'optimize' or any task where sequential execution wastes time.
14