Synthèse automatique
Génère des invites optimisées pour les outils de prototypage IA (v0.dev, Lovable, Bolt.new) à partir de spécifications ou de PRD.
Points forts
- Transforme rapidement des PRD en invites prêtes à l'emploi
- Prend en charge plusieurs outils de prototypage avec des modèles adaptés
- Inclut l'analyse du contexte (design system, parcours utilisateur)
- Produit des invites structurées et complètes
Limites
- Nécessite une compréhension claire des besoins du produit en amont
- Dépend des capacités et des limites des outils cibles
- Ne génère pas directement le prototype, seulement l'invite
Utilisez cette compétence pour accélérer le passage d'un PRD à un prototype fonctionnel en quelques minutes.
Évitez-la si le prototype nécessite des fonctionnalités backend complexes non supportées par les outils cibles.
Analyse de sécurité
SûrThe skill only generates text prompts for AI prototyping tools. It does not instruct execution of any commands, network access, or data exfiltration. The skill reads files (PRDs, design docs) as context but that is within normal AI agent operation and not risky.
Aucun point d'attention détecté
Comment cette analyse est faiteExemples
Generate a prototype prompt for the onboarding wizard from PRD-2024-05Create a Bolt.new prompt for a quick demo of the search featureGenerate a Lovable prompt for a full-stack authentication flowname: generate-ai-prototype description: Generate v0.dev, Lovable, or Bolt.new prompts for AI-powered prototyping
Generate AI Prototype Skill
Create optimized prompts for AI prototyping tools (v0.dev, Lovable, Bolt.new) to quickly build working prototypes from PRDs or feature specs.
Quick Start
- Tell me what you want to prototype (feature name, PRD link, or description)
- I'll ask which tool: v0.dev (UI only), Lovable (full-stack), or Bolt.new (quick demo)
- I check your PRD, design system, and user research for context
- I generate a copy-paste-ready prompt optimized for the tool
- You paste it, get a working prototype, and share the link
Example: "Generate a prototype prompt for the onboarding wizard from PRD-2024-05"
Output: Saved to thoughts/shared/prototypes/[feature]-prototype-prompt.md
Time: 10-15 minutes to generate prompt, 1-2 minutes to build prototype
When to Use This Skill
- Converting PRDs into functional prototypes
- Rapid iteration on UI concepts
- Demonstrating features to stakeholders
- Testing user flows before engineering starts
Supported AI Prototyping Tools
v0.dev (Vercel)
- Best for: React/Next.js apps, modern UI components
- Strengths: Clean design, shadcn/ui integration, responsive
- Limits: Frontend only, no backend logic
- Cost: Free tier, then paid
Lovable (formerly GPT Engineer)
- Best for: Full-stack apps with backend
- Strengths: Complete apps, authentication, database
- Limits: Opinionated stack (React + Supabase)
- Cost: Free tier, then $20/mo
Bolt.new (StackBlitz)
- Best for: Quick demos, isolated components
- Strengths: Instant preview, no account needed
- Limits: Simpler apps, less customization
- Cost: Free
Workflow
Step 1: Gather Requirements
Ask the PM:
- Feature scope: What are we prototyping?
- Target tool: v0.dev, Lovable, or Bolt.new?
- Fidelity level: Quick sketch or polished demo?
- Key user flows: What paths must work?
- Design system: Match existing product or generic?
Step 2: Analyze Context
Read relevant files:
- PRD or feature spec
- Design system docs (if available)
- Existing UI screenshots (if provided)
- User journey maps
Extract:
- Core user flows
- UI components needed
- Data models (if backend)
- Key interactions
- Edge cases to demonstrate
Step 3: Generate Prompt
Create a structured prompt optimized for the target tool.
v0.dev Prompt Template
Build a [feature name] interface with the following requirements:
## Core Functionality
[List 3-5 key features]
## User Flow
1. [Step 1: User action]
2. [Step 2: System response]
3. [Step 3: Next action]
## UI Components Needed
- [Component 1: Description]
- [Component 2: Description]
- [Component 3: Description]
## Design Specifications
- **Style:** [Modern, minimal, professional, etc.]
- **Colors:** [Primary: #hex, Secondary: #hex, or "use default shadcn palette"]
- **Layout:** [Desktop-first, mobile-first, responsive]
- **Components:** Use shadcn/ui components
## Key Interactions
- [Interaction 1: e.g., "Click 'Add Item' opens modal"]
- [Interaction 2: e.g., "Drag to reorder list items"]
## Data Structure (mock)
```typescript
interface [DataType] {
[field]: [type];
[field]: [type];
}
Edge Cases to Show
- [Empty state]
- [Loading state]
- [Error state]
- [Success state]
Example Content
[Provide realistic sample data]
Technical preferences:
- Use TypeScript
- Tailwind CSS for styling
- React hooks for state
- Responsive design (mobile + desktop)
#### Lovable Prompt Template
Create a [feature name] application with the following specs:
App Overview
[1-2 sentence description]
User Stories
- As a [user type], I want to [action] so that [benefit]
- As a [user type], I want to [action] so that [benefit]
- As a [user type], I want to [action] so that [benefit]
Pages/Views
-
[Page Name]
- Purpose: [What this page does]
- Components: [List key UI elements]
- Actions: [User can do X, Y, Z]
-
[Page Name]
- Purpose: [What this page does]
- Components: [List key UI elements]
- Actions: [User can do X, Y, Z]
Database Schema
-- [Table Name]
CREATE TABLE [table_name] (
id UUID PRIMARY KEY DEFAULT uuid_generate_v4(),
[field_name] [TYPE] [CONSTRAINTS],
created_at TIMESTAMP DEFAULT NOW()
);
Authentication
- [x] User signup/login required
- [x] Protected routes: [list routes]
- [ ] Public access allowed
Key Features
Business Logic
- [Rule 1: e.g., "Users can only edit their own items"]
- [Rule 2: e.g., "Max 10 items per user on free tier"]
Sample Data
[Provide realistic examples for testing]
Tech Stack Preferences:
- Frontend: React + TypeScript
- Backend: Supabase
- Styling: Tailwind CSS
- State: React hooks/context
#### Bolt.new Prompt Template
Build a simple [feature name] component:
What it does: [2-3 sentence description]
UI Layout: [Describe the visual structure]
User Interactions:
- [User does X] → [System responds Y]
- [User does A] → [System responds B]
Functionality:
Styling: [Modern/minimal/colorful/etc. + any color preferences]
Example Data: [Provide 2-3 sample items]
Keep it simple and functional. Focus on core flow.
### Step 4: Optimization Tips
**For all tools:**
- ✅ Be specific about functionality
- ✅ Provide example data
- ✅ Describe interactions clearly
- ✅ Specify edge cases (empty, loading, error states)
- ❌ Don't overload with features (start simple)
- ❌ Don't assume technical knowledge (be explicit)
**v0.dev specific:**
- Use shadcn/ui component names
- Specify responsive behavior
- Request TypeScript for better quality
**Lovable specific:**
- Define database schema clearly
- Specify authentication requirements
- Describe business logic rules
**Bolt.new specific:**
- Keep scope small (single feature/component)
- Focus on visual demo over complex logic
### Step 5: Deliver Prompt
Present the generated prompt to the PM with:
1. **The Prompt** (formatted for copy-paste)
2. **Instructions:**
- Where to paste (v0.dev/Lovable/Bolt.new)
- Expected output
- How to iterate (what to tweak if needed)
3. **Next Steps:**
- "Share prototype link with stakeholders"
- "Use as spec for engineering"
- "Iterate based on feedback"
## Advanced Techniques
### Matching Existing Design Systems
If PM provides screenshots or design system docs:
Use this design system:
Colors:
- Primary: [extracted from screenshot]
- Secondary: [extracted from screenshot]
- Text: [extracted from screenshot]
Typography:
- Headings: [font family, size]
- Body: [font family, size]
Components:
- Buttons: [describe style]
- Inputs: [describe style]
- Cards: [describe style]
Spacing:
- Use [4px/8px/16px] grid system
Reference the attached screenshot for visual style.
### Multi-Screen Flow Guidance
For features with multiple screens or states (e.g., onboarding wizard, checkout flow, settings panels), structure the prompt to produce a cohesive multi-step experience rather than isolated pages.
**Include in the prompt:**
1. **Screen/step list with descriptions** -- List every screen by name, its purpose, and what the user sees
2. **Navigation between screens** -- Specify how users move forward, back, and skip (buttons, tabs, swipe, URL routing)
3. **State management (what data carries forward)** -- Clarify which inputs persist across screens (e.g., "selections from Step 1 appear in Step 3 summary")
4. **Edge cases per screen** -- Define empty, error, loading, and validation states for each screen individually
**Prompt structure for multi-step flows:**
Build a multi-step flow with [N] screens:
Screen 1: [Name]
- Purpose: [What happens here]
- UI: [Key components]
- Data collected: [Fields, selections]
- Validation: [Required fields, format rules]
- Next: [Button label] → Screen 2
Screen 2: [Name]
- Purpose: [What happens here]
- UI: [Key components]
- Data from previous: [What carries forward]
- Next: [Button label] → Screen 3
- Back: [Button label] → Screen 1
Screen [N]: [Name - usually Confirmation/Review]
- Purpose: Review and submit
- Shows: Summary of all previous inputs
- Actions: [Submit], [Edit Step X]
- Success state: [What user sees after completion]
Global Requirements
- Progress indicator showing current step (e.g., "Step 2 of 4")
- Persist data if user navigates back
- Handle browser back button gracefully
- Mobile responsive (stack layout vertically on small screens)
**Avoid:** Generating separate, disconnected prompts for each screen. The AI tool produces better results when it understands the full flow context in a single prompt.
### Prototype Iteration
After initial prototype:
Refine the existing prototype:
What's working:
- [Keep feature X]
- [Keep interaction Y]
What to change:
- [Modify feature A: description]
- [Add feature B: description]
- [Remove feature C: reason]
Specific updates:
- [Detailed change 1]
- [Detailed change 2]
## Common Use Cases
### 1. Dashboard Prototype
Build a dashboard showing:
- [Metric 1] as a card with trend
- [Metric 2] as a line chart
- [Recent activity] as a list
- Filters for date range
### 2. Form/Wizard Prototype
Create a multi-step form: Step 1: [Collect basic info] Step 2: [Collect detailed info] Step 3: [Review & confirm] Include validation and progress indicator
### 3. Settings Page Prototype
Build a settings interface with:
- Tabbed navigation (Profile, Preferences, Billing)
- Toggle switches for boolean settings
- Dropdowns for choices
- Save/Cancel buttons
### 4. Onboarding Flow Prototype
Create an onboarding flow: Screen 1: [Welcome + value prop] Screen 2: [Key feature explanation] Screen 3: [Setup requirements] Screen 4: [Success state]
## Output Template
AI Prototype Prompt
Target Tool: [v0.dev / Lovable / Bolt.new]
Feature: [Feature Name]
Prompt (Copy-Paste Ready)
[FULL PROMPT HERE]
Instructions
- Go to [tool URL]
- Paste the prompt above
- Click "Generate" or "Build"
- Wait ~30-60 seconds for prototype
- Share the generated link
Expected Output
The prototype will include:
Iteration Tips
If the prototype needs changes:
- To adjust styling: Add "Use [color/font/spacing] instead"
- To add features: List new features clearly
- To remove clutter: Specify "Remove [feature X]"
Next Steps
- [ ] Share prototype link with team
- [ ] Gather feedback from stakeholders
- [ ] Iterate on design based on feedback
- [ ] Use as specification for engineering
## Pro Tips
1. **Start simple:** Prototype core flow first, add polish later
2. **Use real data:** Sample data should look production-ready
3. **Show edge cases:** Empty states, errors, loading
4. **Be specific:** "Large blue button" > "call-to-action button"
5. **Iterate quickly:** Generate → test → refine → repeat
6. **Share early:** Get feedback on prototype before engineering starts
7. **Document learnings:** Note what works/doesn't for future prototypes
## Common Mistakes to Avoid
❌ Too many features in first prompt (overwhelming)
✅ Core flow first, add features incrementally
❌ Vague descriptions ("nice interface")
✅ Specific details ("Card layout, 3 columns, shadcn components")
❌ No example data (generic placeholders)
✅ Realistic sample data (actual user names, realistic numbers)
❌ Forgetting edge cases (only happy path)
✅ Include empty, loading, error states
❌ No iteration plan (one-and-done)
✅ Expect 2-3 rounds of refinement
## When to Use Each Tool
**Use v0.dev when:**
- Need polished UI components
- Building React/Next.js apps
- Want shadcn/ui integration
- Frontend-only prototype
**Use Lovable when:**
- Need full-stack prototype
- Require authentication/database
- Building complete app MVP
- Want production-ready code
**Use Bolt.new when:**
- Need quick demo (< 10 min)
- Simple component/feature
- No account setup time
- Throwaway prototype
## Output Quality Self-Check
Before delivering the prototype prompt, verify:
- [ ] **Prompt is copy-paste ready** -- No placeholders like "[insert X]" remain; all values are filled in
- [ ] **Tool-specific best practices followed** -- shadcn/ui for v0, Supabase schema for Lovable, simplicity for Bolt
- [ ] **Core user flow is complete** -- The main happy path is fully described, not just individual screens
- [ ] **Edge states specified** -- Empty, loading, error, and success states are included
- [ ] **Sample data is realistic** -- Uses production-like names, numbers, and content (not "Lorem ipsum")
- [ ] **Multi-screen flows are cohesive** -- If multiple screens, navigation, state persistence, and progress indicators are specified
- [ ] **PRD requirements addressed** -- Cross-checked against PRD acceptance criteria (if PRD exists)
- [ ] **Design context included** -- Colors, fonts, and component styles match the product (if design system exists)
- [ ] **Instructions are clear** -- PM knows exactly where to paste, what to expect, and how to iterate
If any check fails, fix it before delivering. The prompt should work on the first try.
---
Remember: A 5-minute prototype can save 2 weeks of engineering waste. Invest in prototyping before building.
Du fichier au premier essai
Le téléchargement contient exactement le contenu affiché sur cette fiche. Cette copie peut différer de la dernière version du dépôt source.
Consulter la source et ses fichiers annexesCe téléchargement contient uniquement SKILL.md. Les scripts, références, modèles et dépendances éventuels ne sont pas inclus. Récupérez les fichiers nécessaires depuis la source en conservant leur arborescence avant de poursuivre.
1. Télécharger le fichier
Télécharger SKILL.mdDepuis la racine d’un projet de test, enregistrez le fichier sous .skill-downloads/generate-ai-prototypes-mmh8qrot/SKILL.md, ou utilisez la commande suivante. Ce dossier sert à examiner la skill avant son installation.
Commande pour macOS/Linux ou WSL : un shell POSIX et curl sont nécessaires. Sous Windows sans WSL, utilisez le bouton de téléchargement et créez le dossier manuellement.
mkdir -p '.skill-downloads/generate-ai-prototypes-mmh8qrot' && curl --fail --show-error --location 'https://skills-guides.com/api/skills/generate-ai-prototypes-mmh8qrot/download' --output '.skill-downloads/generate-ai-prototypes-mmh8qrot/SKILL.md.part' && mv '.skill-downloads/generate-ai-prototypes-mmh8qrot/SKILL.md.part' '.skill-downloads/generate-ai-prototypes-mmh8qrot/SKILL.md'2. Vérifier les prérequis
Lisez SKILL.md avant de l’utiliser. Si le fichier cite des scripts ou services externes, examinez-les et préparez les accès nécessaires dans votre environnement de test.
Aucune section de prérequis reconnue dans ce fichier. Cela ne signifie pas qu’il est autonome : les bibliothèques, outils et accès peuvent être cités ailleurs dans le contenu ou dans le dépôt source.
Les plateformes de la fiche sont des indications du catalogue. Les chemins ci-dessous suivent la documentation des outils ; ils ne prouvent pas que cette skill y a été exécutée. Aucune version minimale de ces outils n’est attestée pour cette fiche.
3. Charger la skill
Après vérification, copiez le dossier complet dans le projet avec l’une des commandes ci-dessous, puis ouvrez une nouvelle session de l’outil. La commande s’arrête si un dossier du même nom existe déjà : comparez les fichiers avant de le remplacer.
Claude Code
test -f '.skill-downloads/generate-ai-prototypes-mmh8qrot/SKILL.md' && test -s '.skill-downloads/generate-ai-prototypes-mmh8qrot/SKILL.md' && mkdir -p '.claude/skills' && test ! -e '.claude/skills/generate-ai-prototype' && cp -R '.skill-downloads/generate-ai-prototypes-mmh8qrot' '.claude/skills/generate-ai-prototype'OpenSkills
OpenSkills doit déjà être installé. La procédure locale install/read est contrôlée avec OpenSkills 1.5.0, qui exige Node.js ≥ 20.6.0. Aucun slug du catalogue n’est résolu automatiquement.
Installation officielle d’OpenSkillsopenskills install './.skill-downloads/generate-ai-prototypes-mmh8qrot'openskills read 'generate-ai-prototypes-mmh8qrot'La lecture doit afficher le contenu du fichier installé et son dossier de base. Elle ne lance ni la skill ni ses dépendances.
Pour un agent qui lit AGENTS.md, synchronisez ensuite l’index. Cette commande modifie AGENTS.md : relisez le diff avant de l’utiliser.
openskills syncMode manuel ou autre outil : joignez SKILL.md à une conversation, ou ouvrez-le dans un agent capable de lire les fichiers. Ajoutez les ressources annexes nécessaires. Il n’existe pas de dossier universel reconnu par tous les outils.
4. Faire un premier essai
Utilisez une entrée courte et fictive : un endpoint d’API, un petit document ou un brief d’interface. Dans une nouvelle conversation, joignez le fichier ou donnez son chemin à l’agent, puis adaptez cette demande :
Lis .skill-downloads/generate-ai-prototypes-mmh8qrot/SKILL.md et applique ses instructions à cet exemple : [décrire une entrée de test]. Indique les fichiers, dépendances ou accès qui manquent. Produis le résultat demandé par la skill sans inventer de données ni exécuter de script non examiné.Résultat attendu : une sortie correspondant aux instructions de la skill, ou une liste précise des prérequis manquants. Une simple confirmation de lecture ne valide pas le premier usage. Si une ressource externe est indisponible, ce parcours reste à tester.
Bonnes pratiques React Hook Form
Developpement
Maîtrisez React Hook Form : évitez les anti-patterns, gérez les abonnements, et traitez correctement reset, dirty state et entrées contrôlées.
Migration Emotion vers Mantine
Developpement
Migrer les styled-components Emotion vers des composants Mantine avec style props et modules CSS. Pour convertir .styled.tsx ou supprimer les imports @emotion.
Détecteur de dérive de périmètre
Developpement
Analyse les diffs git par rapport à une intention pour détecter la dérive de périmètre, fichiers non liés et PR trop larges, puis suggère de garder, diviser ou justifier.