More screenshots
Works with the AI tools you already use
Design System & UI Component Architect
Level 0 — Unsystematic: One-off styling, inconsistent components, no shared foundation.
Secure checkout via Stripe
See it in action
You say
I have 5 different blue buttons and inconsistent padding across my React dashboard. Turn these screenshots into a semantic token system and a standardized button component spec.
Your agent does
### Semantic Color Tokens - `color.action.primary.default`: #0070F3 (from Brand Primary) - `color.action.primary.hover`: #0051B0 - `color.action.primary.focus`: #0070F3 with 2px offset ### Button Component Spec - **Anatomy**: Container, Label, Icon (Optional) - **States**: Default, Hover, Focus, Pressed, Disabled, Loading - **Spacing**: `space.control.md` (8px 16px) - **Radius**: `radius.control.sm` (4px)
What you get
About this skill
Design System & UI Component Architect is a premium product-design, UI architecture, token-system, component-library, accessibility, design-governance, and frontend handoff skill for teams that need to turn inconsistent interfaces into a scalable system.
It can build a design system from:
Existing Digital Products Brand Guidelines Screenshots Mockups Legacy Interfaces Figma Libraries Component Inventories SaaS Products Mobile Apps Enterprise Software Dashboards Admin Panels Ecommerce Products Marketplaces Fintech Interfaces EdTech Products Health Applications Multi-Brand Ecosystems Cross-Platform Products
The skill is designed to solve common problems such as:
Too Many Button Styles Inconsistent Spacing Typography Drift Hard-Coded Colors Duplicated Components Missing Interaction States Missing Focus States Inconsistent Tables Inconsistent Navigation Inconsistent Error Handling Dark Mode Built Through Simple Color Inversion Arbitrary Border Radii Uncontrolled Shadows Figma and Production Mismatches Accessibility Added Too Late Feature Teams Creating Their Own Components Uncontrolled Variants No Deprecation Strategy No Contribution Workflow No Design-System Ownership No Token Architecture No Migration Plan
The skill treats a design system as an operational product framework rather than a decorative component gallery.
Its architecture is:
Brand → Foundations → Primitive Tokens → Semantic Tokens → Component Tokens → Core Components → Product Patterns → Templates → Product Experiences → Governance → Design QA
The skill begins with an audit.
It identifies:
Existing Components Duplicate Components Near-Duplicate Components Visual Drift Behavior Drift Missing States Accessibility Gaps Off-Token Values One-Off Styles Legacy Patterns Product-Specific Components Candidates for Consolidation Candidates for Deprecation
Each component can be classified as:
Keep Merge Split Rename Rebuild Deprecate Core System Product-Specific Extension
The skill can assess design-system maturity.
Level 0 — Unsystematic: One-off styling, inconsistent components, no shared foundation.
Level 1 — Shared Styles: Shared colors and typography exist, but system behavior is limited.
Level 2 — Component Library: Reusable components exist, but governance and tokenization remain incomplete.
Level 3 — Tokenized Design System: Semantic tokens, component states, responsive rules, accessibility, and documentation exist.
Level 4 — Governed Multi-Product System: Contribution workflows, versioning, deprecation, multi-theme support, design-code parity, and controlled product extensions exist.
The skill creates layered token architecture.
Primitive Tokens: Raw values.
Examples:
color.blue.600 space.16 radius.8 font.size.16
Semantic Tokens: Purpose-based values.
Examples:
color.text.primary color.surface.default color.action.primary color.border.subtle color.feedback.error
Component Tokens: Component-specific values only where necessary.
Examples:
button.primary.background.default input.border.focus table.row.background.selected
The preferred architecture is:
Primitive → Semantic → Component
Product interfaces should use semantic meaning rather than raw visual values whenever possible.
The skill creates semantic naming conventions.
Strong:
color.action.primary.hover
Weak:
blueButtonHover
The skill separates brand colors from product roles.
A marketing brand palette may contain:
Signature Colors Editorial Colors Gradients Decorative Accents
but application UI needs semantic roles such as:
Primary Action Secondary Action Focus Error Warning Success Informational Surface Border Text
The skill prevents a brand color from being reused indiscriminately for every action, status, and decorative element.
The color system can define:
Brand: Primary Secondary Accent
Text: Primary Secondary Muted Inverse Disabled Link
Surface: Canvas Default Subtle Raised Overlay Inverse
Border: Subtle Default Strong Focus Error
Action: Primary Secondary Tertiary Destructive
Feedback: Success Warning Error Information
Data Visualization: Categorical Sequential Diverging
Chart colors can remain separate from action colors to avoid semantic confusion.
The skill can build tonal ramps only where the system actually requires them.
Dark mode is designed semantically rather than through mechanical inversion.
Dark-mode review can include:
Surfaces Text Borders Elevation Focus Status Charts Disabled States Illustrations
High-contrast mode can define:
Stronger Borders Explicit Focus Reduced Decorative Surfaces Non-Color Status Communication Improved Data Visualization Differentiation
The skill supports multi-brand architectures.
Possible structure:
Core Foundations → Brand Theme → Product Semantic Layer → Component Layer
This allows shared behavior while preserving brand variation.
The skill defines complete typography architecture.
It can specify:
Font Family Fallback Stack Weights Type Scale Line Heights Tracking Paragraph Spacing Maximum Reading Width Responsive Adjustments
Semantic typography roles may include:
Display Page Title Section Title Card Title Body Large Body Body Small Label Helper Metadata Button Code Numeric Display
The skill avoids excessively granular type scales with nearly identical values.
Text scaling and accessibility are incorporated into the foundation.
Mobile or accessibility-sensitive interfaces should support:
Wrapping Reflow Larger Text Expanded Controls Dynamic Type or Equivalent Scaling
The skill creates systematic spacing.
Example scale:
4 8 12 16 24 32 48 64
or a denser scale when the product requires it.
Semantic spacing can distinguish:
Inline Spacing Control Spacing Component Spacing Group Spacing Section Spacing Layout Spacing
The skill supports density modes.
Comfortable: Suitable for consumer or lower-density interfaces.
Compact: Suitable for enterprise and data-heavy workflows.
Density should be tokenized where possible rather than creating duplicated component libraries.
The skill defines grid systems.
Possible decisions include:
Container Widths Columns Gutters Page Margins Nested Grids Breakpoints Maximum Width
Responsive grids may use patterns such as:
Mobile — 4 columns Tablet — 8 columns Desktop — 12 columns
but the system should follow actual product requirements rather than arbitrary conventions.
Breakpoints should respond to content and layout behavior, not specific device brand names.
The skill creates controlled radius systems.
Example:
0 4 8 12 16 Full/Pill
Potential semantic roles:
Control Card Dialog Pill
It avoids arbitrary corner-radius variation.
The skill creates border systems.
Potential roles:
Subtle Default Strong Focus Error Selected
The skill creates elevation architecture.
Potential levels:
Base Raised Dropdown Popover Modal Toast
Elevation communicates hierarchy rather than decoration.
The skill creates controlled shadow systems and can account for dark-mode differences.
It creates z-index architecture to prevent arbitrary escalation.
Potential layers:
Base Sticky Dropdown Popover Modal Toast Critical Overlay
The skill creates iconography standards.
It can define:
Icon Family Stroke Style Fill Style Sizes Alignment Optical Adjustments Platform Variants Accessible Naming Rules
The skill creates motion systems.
Potential dimensions include:
Duration Easing Fade Scale Slide Expansion Collapse Navigation Transition
Motion should communicate:
Causality State Change Hierarchy Progress
It should not delay task completion.
Reduced-motion alternatives are defined.
The skill creates a unified focus system.
Focus behavior applies consistently to:
Buttons Inputs Links Tabs Menu Items Custom Controls Table Actions
The skill performs component architecture systematically.
Every component can define:
Purpose Anatomy Variants Sizes States Content Rules Behavior Responsive Rules Accessibility Do / Don't Implementation Notes
Before creating a new component, the skill asks:
Is this truly a new pattern? Can an existing component solve it? Would a variant be enough? Would that variant create long-term complexity? Is the pattern domain-specific rather than universal?
The skill prevents uncontrolled variant proliferation.
Strong variants represent meaningful semantic differences.
Example:
Primary Secondary Tertiary Destructive
Weak variants represent accidental visual differences.
Example:
Blue Dark Blue Light Blue Blue With Shadow Blue Rounded
The skill can define a complete button system.
Possible variants:
Primary Secondary Tertiary Destructive Icon Button
Possible sizes:
Small Medium Large
Possible states:
Default Hover Pressed Focus Disabled Loading
Rules can include:
Keep one dominant primary action within a local context where possible. Do not use destructive styling for ordinary secondary actions. Icon-only buttons require accessible names. Loading states should preserve layout stability where practical.
The skill creates link systems.
Possible types:
Inline Link Standalone Link Navigation Link Destructive Link
The skill creates complete input systems.
Components may include:
Text Input Textarea Password Search Number Email Phone
URL
Input anatomy can define:
Label Required Indicator Control Prefix Suffix Helper Error Counter
Input states can include:
Default Hover Focus Filled Disabled Read-Only Error Success
Persistent labels are preferred.
Placeholder text should not function as the only label.
The skill defines selects, dropdowns, and comboboxes.
It considers:
Search Keyboard Interaction Clear Behavior Groups Selected Values Disabled States Errors
The skill avoids custom interaction models that reduce keyboard accessibility.
The skill defines:
Checkboxes for independent choices. Radio buttons for one selection among visible alternatives. Switches for immediate binary settings. Segmented controls for compact mode switching.
It avoids using switches for actions that still require a separate Save action unless that behavior is clearly communicated.
The skill creates date and time picker rules.
It can define:
Locale Timezone Formatting Allowed Ranges Keyboard Alternatives Desktop Behavior Mobile Behavior
The skill creates search-field architecture.
Potential states include:
Idle Focused Loading Suggestions Results No Results Error
The skill creates form systems.
It defines:
Field Groups Vertical Rhythm Section Grouping Single vs Multi-Column Layout Responsive Collapse Validation Error Summary Recovery Success
It avoids unnecessary multi-column forms.
The skill creates card systems.
Possible variants include:
Static Interactive Selectable Summary Metric Media
It avoids "card for everything" interface architecture.
Card anatomy can include:
Header Title Subtitle Content Status Actions Footer
The skill treats tables as a dedicated system rather than a generic card variant.
Table architecture can define:
Header Rows Columns Sorting Filtering Selection Bulk Actions Row Actions Pagination Sticky Behavior Density Empty State Loading Error Responsive Behavior
Table tokens can include:
Row Height Header Height Cell Padding Selected Background Hover Background Borders Sticky Shadow
Column rules can define:
Text Alignment Numeric Alignment Truncation Wrapping Priority Minimum Width Maximum Width
Responsive table strategies may include:
Priority Columns Stacked Records Expandable Detail Horizontal Scroll Where Necessary
The skill creates data-visualization systems.
Potential areas include:
Categorical Colors Sequential Colors Diverging Colors Axes Gridlines Tooltips Legends Annotations Accessible Alternatives
Chart palettes should not casually reuse interaction colors.
The skill creates navigation systems.
Potential components include:
Top Navigation Sidebar Bottom Navigation Tabs Breadcrumbs Command Palette Pagination
Sidebar architecture can define:
Expanded Width Collapsed Width Groups Active State Badges Nested Items Scrolling Mobile Adaptation
Top navigation can define:
Logo Workspace Search Primary Actions Notifications Account
Tabs can define:
Default Hover Focus Selected Disabled Overflow Responsive Behavior
Breadcrumbs can define:
Separator Truncation Current Item Overflow
Pagination can define:
Previous / Next Page Numbers Rows Per Page Mobile Adaptation Accessibility Labels
The skill creates a feedback system.
Components may include:
Toast Snackbar Banner Inline Alert Notification Status Badge Progress Skeleton
Alert variants may include:
Informational Success Warning Error Critical Where Required
Alert anatomy can define:
Icon Title Message Action Dismiss Persistence
The skill avoids using transient toast notifications as the only confirmation for critical outcomes.
Persistent banners are reserved for appropriate product-wide states rather than becoming permanent clutter.
The skill creates badge and status architecture.
Status should not rely on color alone.
The skill creates tooltip and popover systems.
Tooltips supplement information.
They should not contain essential instructions required for task completion.
Popovers can support:
Contextual Detail Short Action Groups Compact Controls
The skill creates menu architecture.
Menu items can define:
Label Icon Shortcut Destructive State Disabled State Grouping Separators
The skill creates modal/dialog systems.
Dialog anatomy can include:
Title Body Primary Action Secondary Action Close Destructive Variant
Behavior can define:
Focus Trap Scrolling Keyboard Responsive Adaptation
Nested modals should be avoided.
The skill creates drawer and side-sheet systems for:
Quick Details Editing Filtering Inspectors
It can define width, closing behavior, action placement, and responsive adaptation.
For mobile products, the skill can create bottom-sheet behavior.
It can define:
Grabber Snap Points Where Appropriate Safe Areas Actions Keyboard Interaction
The skill creates a complete empty-state system.
Types may include:
First Use Filtered Empty Completed Permission Restricted Error Empty
Anatomy can include:
Optional Illustration Title Explanation Primary Action Secondary Action
The skill creates loading systems.
Possible patterns include:
Spinner Skeleton Progress Bar Determinate Progress Indeterminate Progress Local Loading Full-Screen Loading
Skeletons are recommended only when content structure is predictable.
The skill creates error systems.
Potential levels include:
Field Error Component Error Module Error Page Error Network Error Permission Error Destructive Failure
The skill creates success systems.
Potential patterns include:
Inline Success Toast Dedicated Confirmation Receipt Completed State
The skill creates permission-state architecture.
Potential strategies:
Hidden Disabled + Explanation Access Request Restricted Page
The skill creates supporting components such as:
Avatar Chip Tag Accordion Stepper File Upload Media Components
File-upload systems can define:
Idle Drag Selected Uploading Success Error Retry Remove
The skill distinguishes components from product patterns.
Patterns combine multiple components into repeatable workflows.
Potential patterns include:
Application Shell Dashboard Settings Record Detail Form Checkout Onboarding Search Results Master-Detail Split Pane
The application shell can define:
Global Navigation Content Area Utility Area Responsive Behavior Loading System Status
Dashboard patterns can define:
Page Header Controls KPI Zone Task Zone Analytics Zone Alerts Responsive Collapse
Settings patterns can define:
Category Navigation Section Layout Form Hierarchy Save Behavior Destructive Section
Record-detail patterns can define:
Identity Status Primary Actions Tabs or Sections Activity Metadata
The skill creates responsive rules.
Mobile considerations include:
Safe Areas Touch Targets Bottom Navigation Sheets Keyboard Mobile Density Mobile Tables Mobile Filters
Desktop considerations include:
Hover Keyboard Focus Wide Layouts Sidebars High Data Density Resizable Panes
Tablet considerations include:
Split View Multi-Pane Adaptive Navigation Touch and Keyboard Interaction
A responsive component matrix can define:
Mobile Tablet Desktop Large Desktop
The skill supports cross-platform design systems.
Shared layers may include:
Brand Semantic Colors Terminology Information Hierarchy Product Rules
Platform-specific adaptation may include:
Navigation Pickers Sheets Back Behavior System Prompts Typography Metrics Gestures
The skill builds accessibility into components rather than adding it later.
Every component can define:
Semantic Role Accessible Name Keyboard Behavior Focus Behavior Touch Target Contrast Text Scaling Reduced Motion Error Association Screen-Reader Announcement
The skill can perform WCAG-oriented checks for:
Contrast Focus Visibility Keyboard Access Labels Headings Status Messages Errors Target Sizes Color Independence
It does not claim formal accessibility compliance without implementation testing.
The skill incorporates localization.
It can account for:
Long Translations Pluralization Date Formats Number Formats Currencies
RTL
Wrapping Truncation
The skill creates content-design rules.
Potential standards include:
Capitalization Punctuation CTA Style Terminology Error Voice Status Language Empty-State Voice
A product terminology glossary can prevent inconsistent terms from appearing across different screens.
The skill creates naming conventions.
Possible design naming:
Component / Variant / State / Size
Possible token naming:
category.role.state
Design and code names should align where practical.
The skill creates Figma architecture.
Potential library structure:
Foundations Tokens Components Patterns Templates Examples Deprecated
Each component documentation page can include:
Purpose Anatomy Variants States Behavior Accessibility Content Rules Responsive Rules Examples Do / Don't
Where supported, Figma variables can map:
Primitive Tokens Semantic Tokens Themes Modes
The skill creates design-to-code parity architecture.
Recommended model:
Design Component ↔ Code Component ↔ Documentation ↔ Token Source
Names should not drift unnecessarily between design and engineering systems.
Frontend token outputs may be prepared conceptually for:
JSON
CSS Variables TypeScript Platform-Specific Token Files
The implementation format should follow the user's engineering environment.
The skill creates Design QA processes.
Checks can identify:
Off-Token Colors Off-Token Spacing Unsupported Typography Unsupported Radii One-Off Shadows Invalid Variants Missing Focus Missing Error State Inconsistent Icon Sizes Component Drift
A component consistency check can define:
Expected Pattern Observed Pattern Deviation Severity Required Action
A token consistency check can define:
Property Expected Token Observed Value Off-Token Status Recommendation
The skill performs variant audits.
It can define:
Current Variants Duplicate Variants Missing Variants Deprecated Variants Recommended Final Set
The skill creates governance because design systems degrade without ownership.
Governance can define:
Ownership Contribution Review Approval Documentation Release Versioning Deprecation Migration Quality Assurance
Possible ownership roles include:
Design-System Lead Design Owner Engineering Owner Accessibility Reviewer Content Reviewer Product Representative
A contribution workflow can follow:
Need Identified → Existing Component Check → Proposal → Design Review → Accessibility Review → Engineering Review → Implementation → Documentation → Release
New component proposals can require:
Problem User Need Existing Workaround Proposed Anatomy Variants States Accessibility Responsive Behavior Product Usage Maintenance Cost
The skill discourages adding new variants just because one screen needs a visual exception.
Legitimate exceptions can be documented with:
Reason Scope Owner Review Date Expiration Where Appropriate
The skill creates deprecation strategy.
Deprecated components can define:
Replacement Migration Guidance Removal Timeline Where Appropriate
The skill creates versioning and release strategy.
Potential mechanisms include:
Semantic Versioning Release Notes Component Changelog
Breaking changes may include:
Removed Variant Changed Prop Changed Behavior Changed Spacing Renamed Token
Migration impact should be documented.
Release notes can include:
Added Changed Deprecated Removed Fixed Accessibility
The skill can define system metrics such as:
Component Adoption Off-System Components Token Adoption Duplicate Variants Accessibility Issues Design-Code Parity Migration Progress
It does not invent metric values.
The skill creates migration strategies for legacy products.
Possible phases:
Phase 1 — Audit Phase 2 — Foundations and Tokens Phase 3 — Core Components Phase 4 — High-Traffic Screens Phase 5 — Product Patterns Phase 6 — Deprecation Cleanup Phase 7 — Governance Phase 8 — QA Automation
Migration can prioritize:
High-Use Components High-Risk Inconsistencies Accessibility Problems Duplicate Components Components Blocking Redesign
The skill supports brownfield systems.
It does not require complete product rewrites.
Possible strategies include:
Compatibility Layers Phased Token Adoption Component Wrappers Incremental Replacement
For greenfield systems:
Foundations First Core Components Second Patterns Third Product Screens Fourth
The skill supports multi-product ecosystems.
Recommended model:
Core System → Product Extensions → Brand Themes
Product-specific components remain separate where they encode real domain behavior.
Examples include:
Trading Ticket Medical Result Card AI Prompt Composer Order Fulfillment Row Learning Progress Card
The skill prevents domain-specific complexity from contaminating universal components.
Supported operating modes include:
Full Design System Architect: Creates foundations, tokens, components, patterns, responsiveness, accessibility, governance, and migration.
Screenshot-to-System Reconstructor: Extracts repeated visual patterns from screenshots and converts them into a provisional system while labeling inferred values.
Design Token Architect: Focuses on primitive, semantic, component, theme, density, and platform tokens.
UI Component Library Architect: Focuses on reusable controls, navigation, data components, feedback, overlays, and interaction states.
Design-System Consistency Auditor: Finds duplicated components, off-token styling, missing states, and UI drift.
Figma Library Architect: Structures variables, pages, naming, variants, documentation, and contribution workflows.
Design-to-Code System Architect: Maps UI foundations and components to frontend architecture.
Multi-Brand Design System Architect: Creates shared product architecture with theme-level brand adaptation.
Accessibility-First Design System Architect: Builds focus, contrast, keyboard, semantics, text scaling, and reduced-motion requirements into the system.
Legacy UI Migration Architect: Creates phased migration from fragmented UI to standardized components and tokens.
The core commercial promise is: transform scattered UI decisions into a scalable design system with semantic foundations, reusable components, complete states, responsive rules, accessibility, design-code alignment, and governance so the product can grow without continuously accumulating visual and interaction inconsistency.
How to install
Drop the file into your AI Agent. Works with Claude, Cursor, ChatGPT, and 20+ more.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill passes our trust checks below.
Security scanned
Passed our 8-point scan before listing
Fresh listing
Recently published to Agensi
30-day refund
Not a fit? Get your money back
Trust & safety
Security scanned
Verified clean today
- 30-day refund guarantee
- One-time purchase, yours forever
- Secure checkout via Stripe