Full pipeline for systematically building a UI design system. An agent team collaborates to create design tokens, React/Vue component libraries, Storybook, WCAG accessibility verification, and documentation. Use this skill for requests like 'build a design system', 'component library', 'UI kit', 'design tokens', 'Storybook setup', 'accessibility verification', 'WCAG verification', 'UI components', 'create a button component', etc. Note: Figma plugin development, design system SaaS constructio...
Scanned 9/8/2026
Install to Claude Code
npx -y skills add revfactory/harness-100 --skill design-system --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Design System?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/revfactory-design-system-harness-100)More formats (shields.io, HTML) on the badges page.
---
name: design-system
description: "Full pipeline for systematically building a UI design system. An agent team collaborates to create design tokens, React/Vue component libraries, Storybook, WCAG accessibility verification, and documentation. Use this skill for requests like 'build a design system', 'component library', 'UI kit', 'design tokens', 'Storybook setup', 'accessibility verification', 'WCAG verification', 'UI components', 'create a button component', etc. Note: Figma plugin development, design system SaaS construction, and native app components (SwiftUI/Jetpack Compose) are outside the scope of this skill."
---
# Design System — UI Design System Full Pipeline
An agent team collaborates to build design tokens, components, Storybook, accessibility verification, and documentation.
## Execution Mode
**Agent Team** — Five agents communicate directly via SendMessage and perform cross-validation.
## Agent Composition
| Agent | File | Role | Type |
|-------|------|------|------|
| token-designer | `.claude/agents/token-designer.md` | Design token design | general-purpose |
| component-developer | `.claude/agents/component-developer.md` | Component development | general-purpose |
| a11y-auditor | `.claude/agents/a11y-auditor.md` | Accessibility verification | general-purpose |
| storybook-builder | `.claude/agents/storybook-builder.md` | Storybook setup | general-purpose |
| doc-writer | `.claude/agents/doc-writer.md` | Documentation | general-purpose |
## Workflow
### Phase 1: Preparation (performed directly by the orchestrator)
1. Extract the following from user input:
- **Brand information**: Brand colors, logo, tone and manner
- **Framework**: React/Vue/Angular/Web Components
- **Component scope**: Required component list (or "all")
- **Accessibility level**: WCAG AA (default) / AAA
- **Existing system** (optional): Existing design system to migrate from
2. Create the `_workspace/` directory and subdirectories
3. Organize the input and save it to `_workspace/00_input.md`
4. If pre-existing files are available, copy them to `_workspace/` and skip the corresponding phase
5. **Determine the execution mode** based on the scope of the request
### Phase 2: Team Assembly and Execution
| Order | Task | Owner | Dependencies | Deliverable |
|-------|------|-------|-------------|-------------|
| 1 | Token design | token-designer | None | `01_design_tokens/` |
| 2 | Component development | component-developer | Task 1 | `02_components/` |
| 3a | Storybook setup | storybook-builder | Tasks 1, 2 | `03_storybook/` |
| 3b | Accessibility verification | a11y-auditor | Tasks 1, 2 | `04_a11y_report.md` |
| 4 | Documentation | doc-writer | Tasks 1, 2, 3a, 3b | `05_docs/` |
Tasks 3a (Storybook) and 3b (accessibility) run **in parallel**.
**Inter-agent communication flow:**
- token-designer completes > passes token import methods to component-developer, contrast verification to a11y-auditor
- component-developer completes > passes props types to storybook-builder, ARIA status to a11y-auditor
- a11y-auditor > requests fixes from component-developer for P0/P1 issues (up to 2 rounds)
- storybook-builder completes > passes Storybook structure to doc-writer
- a11y-auditor completes > passes accessibility guide to doc-writer
- doc-writer performs final consistency verification across all deliverables
### Phase 3: Integration and Final Deliverables
1. Verify all files in `_workspace/`
2. Confirm all accessibility P0 issues are resolved
3. Present the final summary to the user:
- Token system — `01_design_tokens/`
- Components — `02_components/`
- Storybook — `03_storybook/`
- Accessibility report — `04_a11y_report.md`
- Documentation — `05_docs/`
## Execution Modes by Request Scope
| User Request Pattern | Execution Mode | Agents Deployed |
|---------------------|---------------|----------------|
| "Build a complete design system" | **Full pipeline** | All 5 agents |
| "Just design tokens" | **Token mode** | token-designer only |
| "Build components from these tokens" | **Component mode** | component-developer + storybook-builder |
| "Accessibility verification only" | **Verification mode** | a11y-auditor only |
| "Just add Storybook" | **Storybook mode** | storybook-builder only |
| "Just write documentation" | **Doc mode** | doc-writer only |
| "Add just a Button component" | **Single component** | component-developer + a11y-auditor + storybook-builder |
**Reusing existing systems**: If existing tokens are available, skip token-designer. If only adding Storybook to existing components, deploy storybook-builder only.
## Data Transfer Protocol
| Strategy | Method | Purpose |
|----------|--------|---------|
| File-based | `_workspace/` directory | Code, configuration, documentation |
| Message-based | SendMessage | Key information, fix requests, verification results |
## Error Handling
| Error Type | Strategy |
|-----------|----------|
| Brand color not provided | Start with neutral palette (slate); request colors from user |
| Framework unspecified | Default to React + TypeScript; change after user confirmation |
| Contrast ratio not met | Auto-adjust and report both original and adjusted values |
| Accessibility P0 unresolved | Block release; request re-fix from component-developer |
| Agent failure | Retry once; if still failing, proceed without that deliverable |
## Test Scenarios
### Normal Flow
**Prompt**: "Build a React design system with blue brand colors. I need Button, Input, Card, and Modal"
**Expected result**:
- Tokens: Blue-based color scale, semantic colors, typography, spacing, dark mode
- Components: Button (4 variants, 3 sizes), Input (text/password/search), Card, Modal + shared hooks
- Storybook: Token visualization, Default/AllVariants/Playground stories per component
- Accessibility: WCAG AA compliance verification, keyboard tests, contrast matrix
- Documentation: Design principles, getting started guide, per-component Do/Don't, contribution guide
### Existing File Reuse Flow
**Prompt**: "Add a DatePicker component to the existing design system"
**Expected result**:
- component-developer implements DatePicker using existing tokens
- a11y-auditor verifies date picker accessibility (keyboard, screen reader)
- storybook-builder adds DatePicker stories
- token-designer and doc-writer have minimal involvement
### Error Flow
**Prompt**: "Build a design system" (no colors or framework specified)
**Expected result**:
- token-designer starts with a neutral palette while requesting brand colors from user
- component-developer defaults to React + TypeScript
- a11y-auditor verifies contrast ratios of the default palette
- doc-writer emphasizes the customization guide (how to change colors)
## Agent Extension Skills
Skills that enhance each agent's domain expertise:
| Skill | File | Target Agent | Role |
|-------|------|-------------|------|
| wcag-checker | `.claude/skills/wcag-checker/skill.md` | a11y-auditor, component-developer | WCAG 2.1 accessibility checklist, contrast calculation, ARIA patterns, keyboard matrix |
| token-generator | `.claude/skills/token-generator/skill.md` | token-designer, component-developer | Design token 3-tier structure, color scale algorithm, typography/spacing/motion system |
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!