Claude
Skills
Sign in
โ† Back

Plugin Guide & Help

Included with Lifetime
$97 forever

Use this skill when users ask specifically about the Storybook Assistant plugin - "what can the storybook assistant do", "storybook plugin features", "storybook assistant commands", "help with storybook assistant", "storybook plugin capabilities". ONLY triggers when user explicitly mentions "storybook assistant", "storybook plugin", or asks about Storybook-specific features. Does NOT trigger on generic help questions.

General

What this skill does


# Plugin Guide & Help

## Overview

This skill helps users discover and understand all Storybook Assistant Plugin capabilities through natural language queries **that specifically mention the Storybook Assistant or Storybook plugin**.

## When to Use This Skill

**IMPORTANT:** This skill should ONLY trigger when the user explicitly mentions the plugin or Storybook-specific features.

**โœ… DO Trigger on:**
- "What can the **Storybook Assistant** do?"
- "Show me **Storybook plugin** features"
- "Help with **Storybook Assistant**"
- "List **Storybook** commands"
- "**Storybook Assistant** capabilities"
- "What **Storybook** features are available?"
- "How do I use the **Storybook plugin**?"
- "Tell me about **Storybook Assistant**"

**โŒ DO NOT Trigger on:**
- "help" (generic - goes to Claude Code help)
- "what can you do" (generic - not plugin-specific)
- "show features" (generic - could be anything)
- "list commands" (generic - could be any commands)
- General questions that don't mention Storybook or the plugin

## Response Strategy

### For Storybook-Specific Feature Questions

When user asks about the Storybook Assistant plugin specifically, provide a well-organized, scannable overview:

```markdown
# ๐ŸŽจ Storybook Assistant - What I Can Do

I'm your complete AI-powered Storybook development assistant with 17 specialized features and 11 commands.

## ๐Ÿš€ Most Popular Features

### Vision AI: Design-to-Code ๐ŸŽจ
Upload a screenshot and I'll generate pixel-perfect React components with:
- Exact spacing, colors, and typography
- Design tokens automatically extracted
- Multiple variants detected
- Full TypeScript support

**Try:** `/design-to-code` or "Generate a component from this design"

### Natural Language Components ๐Ÿ’ฌ
Describe a component in plain English and I'll build it:
- Complete TypeScript component
- Props with types
- Storybook stories
- Interaction tests

**Try:** `/generate-from-description` or "Create a user profile card with..."

### AI-Powered Accessibility โ™ฟ
Automatically find and fix WCAG 2.2 violations:
- Context-aware fix suggestions
- One-click application
- Ranked recommendations
- Real-time detection

**Try:** `/fix-accessibility Button.tsx` or "Check accessibility of my Card component"

### Visual Regression Testing ๐Ÿ”
Intelligent screenshot comparison with 90% fewer false positives:
- AI understands intentional vs accidental changes
- Auto-approves theme updates
- Detects real layout bugs
- CI/CD integration

**Try:** `/setup-visual-testing`

### Design Token Sync ๐Ÿ”„
Keep Figma and code perfectly synchronized:
- Bidirectional sync (Figma โ†” Code)
- Automatic drift detection
- Conflict resolution
- Multi-format output

**Try:** `/sync-design-tokens --setup`

---

## ๐Ÿ“‹ All Available Commands

### Setup & Configuration
- `/setup-storybook` - Initialize Storybook 10
- `/setup-ci-cd` - Generate CI/CD pipeline
- `/setup-visual-testing` - Configure visual regression

### Component Development
- `/create-component` - Scaffold new component
- `/generate-stories` - Generate stories for existing components
- `/design-to-code` - Screenshot โ†’ Component
- `/generate-from-description` - Natural language โ†’ Component

### Quality & Maintenance
- `/fix-accessibility` - AI a11y remediation
- `/generate-dark-mode` - Auto-generate dark mode
- `/sync-design-tokens` - Figma sync
- `/analyze-usage` - Component analytics

---

## ๐Ÿ’ก Common Workflows

**New Project:**
1. `/setup-storybook`
2. `/setup-ci-cd`
3. `/setup-visual-testing`

**Design-to-Code:**
1. `/design-to-code [screenshot]`
2. `/fix-accessibility ComponentName`
3. `/generate-dark-mode`

**Improve Existing:**
1. `/generate-stories ComponentName`
2. `/fix-accessibility ComponentName`
3. `/analyze-usage ComponentName`

---

## ๐ŸŽฏ Just Ask Me Naturally!

You don't need to memorize commands. Just ask:
- "Generate a button component from this description..."
- "Fix accessibility issues in my Card component"
- "Set up visual regression testing"
- "Sync my design tokens from Figma"
- "Show me how to create a dark mode"

I understand natural language and will use the right tools automatically.

---

**Getting started?** Try `/setup-storybook` first
**Have a screenshot?** Try `/design-to-code`
**Want more details?** Just ask about any specific feature!
```

### For Specific Feature Questions

When user asks about a specific capability:

**User:** "Can you help with accessibility?"
**Response:**
```markdown
# โ™ฟ Accessibility Features

Yes! I have comprehensive AI-powered accessibility features:

## What I Can Do

1. **Automatic Detection**
   - Scans components for WCAG 2.2 violations
   - Real-time analysis via hooks
   - Covers 90% of common issues

2. **AI-Powered Fixes**
   - Context-aware suggestions (knows button purpose)
   - Ranked recommendations (Best โ†’ Good โ†’ Acceptable)
   - One-click application
   - Color contrast calculation

3. **Education**
   - Explains WHY each fix is needed
   - Teaches WCAG best practices
   - Links to relevant WCAG guidelines

## How to Use

**Quick fix for a component:**
```bash
/fix-accessibility Button.tsx
```

**Natural language:**
"Check my Card component for accessibility issues"
"Fix accessibility violations in the LoginForm"

## What Gets Checked

- Missing accessible names (buttons, links, images)
- Color contrast (WCAG AA/AAA)
- Keyboard navigation
- ARIA usage (semantic HTML first)
- Form labels
- Heading hierarchy

## Example Output

```
โŒ Accessibility issue detected in Button.tsx:12

Issue: Button has no accessible name
WCAG: 4.1.2 Name, Role, Value (Level A)
Impact: Screen reader users won't know button purpose

Recommended fixes:
1. Add visible text + icon (BEST) โญ
2. Use aria-label (GOOD)
3. Use title attribute (ACCEPTABLE)

Apply fix? [1] [2] [3]
```

**Related Features:**
- `testing-suite` - Accessibility tests in Storybook
- `component-scaffold` - New components with a11y built-in

Try `/fix-accessibility [file]` or just ask me to check accessibility!
```

### For "How Do I..." Questions

Extract the task and provide step-by-step guidance:

**User:** "How do I set up visual regression testing?"
**Response:**
```markdown
# Setting Up Visual Regression Testing

Here's how to set up AI-powered visual regression testing:

## Quick Start (Automated)

Just run:
```bash
/setup-visual-testing
```

I'll automatically:
1. Detect your framework and build tool
2. Install dependencies (Playwright, test-runner)
3. Create configuration files
4. Capture baseline screenshots
5. Set up NPM scripts
6. Configure CI/CD integration (optional)

## Step-by-Step Process

**1. Run Setup Command**
```bash
/setup-visual-testing
```

**2. Choose Your Preferences**
I'll ask about:
- Integration with Chromatic/Percy (optional)
- Auto-approval rules (conservative/balanced/aggressive)
- CI/CD integration (GitHub Actions, GitLab, etc.)

**3. Capture Baselines**
Initial screenshots will be captured automatically:
```
Capturing baseline screenshots...
  โœ“ Button/Primary
  โœ“ Card/Default
  ...
  โœ“ Captured 47 stories
```

**4. Test Your Setup**
Make a small change and run:
```bash
npm run test:visual
```

**5. Review AI Analysis**
The AI will categorize changes:
- โœ… Expected (theme updates)
- โš ๏ธ Warning (review needed)
- โŒ Error (likely bugs)

## What Makes It "AI-Powered"?

Traditional visual regression:
- Flags thousands of anti-aliasing changes โŒ
- Can't distinguish intentional from accidental โŒ
- Requires manual review of everything โŒ

AI-powered approach:
- Understands git history and design tokens โœ…
- Auto-approves expected changes โœ…
- Only flags real issues โœ…
- 90% fewer false positives โœ…

## Next Steps

After setup:
- Run `npm run test:visual` to test
- Use `npm run test:visual:interactive` for manual review
- Integrate with CI/CD to block regressions

Need more details? Type `/help visual-regression`
```

### For Version/Changelog Questions

**User:** "What's new?" or "What changed?"
**Response:**
```markdown
# ๐Ÿ†• What's New in v2.1.0

## Major Additions (3 New Featur
Files: 1
Size: 11.8 KB
Complexity: 23/100
Category: General

Related in General