Replit Prompts: Build Secure Apps Fast
Templates for Replit Agent that generate production-ready code
Replit Agent builds complete apps from natural language prompts. These templates include security requirements that prevent hardcoded secrets, SQL injection, and missing auth. Copy the prompt, customize for your use case, and let Replit build - then review before deploying.
Why Replit prompts matter for security
Replit Agent transforms natural language into working applications. It handles file structure, dependencies, and deployment configuration automatically. But like all AI tools, it optimizes for functionality - not security.
A prompt like "build me a user dashboard" generates working code. But that code might hardcode API keys, skip input validation, or use string concatenation in database queries. The fix: include security requirements in your prompt.
These templates address the vulnerabilities most common in vibe coded apps: hardcoded secrets, missing authentication, and SQL injection. Use them as starting points, then review the generated code.
Replit prompt templates (copy-paste ready)
Each template includes security requirements tailored to Replit's environment. Copy, customize the bracketed sections, and paste into Replit Agent.
Secure Full-Stack App
Complete app with auth and database
Build a full-stack web app for [PURPOSE]:
FEATURES:
- User authentication (signup, login, logout)
- Protected dashboard showing user data
- CRUD operations for [RESOURCE]
- Responsive design with Tailwind
SECURITY REQUIREMENTS:
- Use Replit Secrets for all API keys and credentials
- Hash passwords before storing (use bcrypt)
- Validate all user input on the server with Zod
- Use parameterized queries for all database operations
- Check authentication on every protected route
- Use httpOnly cookies for sessions
STACK:
- Next.js or Express backend
- Replit PostgreSQL for database
- Tailwind CSS for styling
Generate secure code following these patterns.API with Auth Middleware
REST API with consistent auth checks
Create a REST API with authentication:
ENDPOINTS:
- POST /api/auth/register - Create account
- POST /api/auth/login - Get session
- GET /api/me - Get current user (protected)
- GET /api/[resource] - List resources (protected)
- POST /api/[resource] - Create resource (protected)
- PUT /api/[resource]/:id - Update resource (protected)
- DELETE /api/[resource]/:id - Delete resource (protected)
SECURITY:
- Create auth middleware that runs on ALL protected routes
- Verify session token validity
- Check resource ownership before update/delete (prevent IDOR)
- Use parameterized queries with Replit PostgreSQL
- Validate request body with Zod schemas
- Rate limit auth endpoints (5 per minute)
- Store secrets in Replit Secrets, never in code
Return proper HTTP status codes (401 unauthorized, 403 forbidden, 404 not found).Secure Form Handler
Form with validation and CSRF protection
Build a contact form with full security:
FORM FIELDS:
- Name (required, 2-100 characters)
- Email (required, valid email format)
- Message (required, 10-2000 characters)
- Honeypot field (hidden, must be empty)
SECURITY FEATURES:
- Client-side validation for UX feedback
- Server-side validation (never trust client)
- Sanitize all input before storing
- Rate limit: 3 submissions per minute per IP
- CSRF token validation
- No sensitive data in error messages
STACK:
- React with form handling
- Express API route
- Zod for validation
- Store submissions in Replit Database
Include both validation layers.File Upload Service
Secure file uploads with validation
Create a file upload feature with security:
ALLOWED FILES:
- Images: jpg, png, gif, webp (max 5MB)
- Documents: pdf (max 10MB)
SECURITY REQUIREMENTS:
- Validate file type by magic bytes, not just extension
- Reject files larger than limits on both client and server
- Generate random filenames (UUID), never use user input
- Store outside public directory or use signed URLs
- Scan uploaded content for common attack patterns
- No path traversal - sanitize all file paths
- Log all upload attempts
IMPLEMENTATION:
- Multer or similar for handling multipart
- Store in Replit Object Storage or external S3
- Return signed URLs for download
Validate properly, not just by extension.Real-time Chat App
WebSocket chat with message validation
Build a real-time chat application:
FEATURES:
- User authentication required to chat
- Multiple chat rooms
- Real-time message delivery
- Message history (last 100 messages)
SECURITY:
- Authenticate WebSocket connections
- Validate message content (max 1000 chars)
- Sanitize messages before broadcasting (prevent XSS)
- Rate limit messages (10 per minute per user)
- Store messages with parameterized queries
- Don't expose user data in room lists
STACK:
- Socket.io or ws for WebSocket
- Express for auth endpoints
- Replit PostgreSQL for storage
- React frontend
Validate all messages server-side.Replit-specific security tips
Replit has unique features for security. Use these patterns in your prompts and code.
Use Replit Secrets
Click the lock icon in sidebar. Add API keys there, access via process.env.SECRET_NAME
const apiKey = process.env.STRIPE_SECRET_KEYEnable Always On
For production apps, enable "Always On" in deployment settings to prevent sleep
// No code needed - configure in Replit dashboardUse Replit Database
Built-in PostgreSQL is secure by default. Connect via DATABASE_URL secret
const pool = new Pool({ connectionString: process.env.DATABASE_URL })Configure .replit properly
Set run command without exposing secrets
run = "npm start" # Good
# run = "API_KEY=xxx npm start" # Bad - exposes secretPost-generation security checklist
After Replit Agent generates your app, verify these security points before deploying.
All API keys should be in Replit Secrets, not in code files
sk_, api_key, secret, password, tokenEnsure .replit file does not expose secrets in run command
run =, entrypointLook for string concatenation in SQL queries
${, + userId, template literal in queryEvery protected route should use auth middleware
authenticate, requireAuth, sessionUser input should be validated with Zod or similar
z.object, schema.parse, validateCORS should not be * in production
cors, origin: '*', Access-ControlReplit vs Bolt.new vs Lovable
Each AI app builder has different strengths. Choose based on your workflow.
| Feature | Replit | Bolt.new | Lovable |
|---|---|---|---|
| Best for | Full development + hosting | Quick app generation | UI-focused apps |
| IDE included | Yes (full browser IDE) | No (generates code) | Minimal editor |
| Built-in hosting | Yes (free + paid tiers) | Yes (Netlify) | Yes (Netlify) |
| Database support | PostgreSQL built-in | External only | Supabase focus |
| Collaboration | Real-time multiplayer | Export to GitHub | GitHub sync |
| AI model | Replit Agent + Ghostwriter | Claude | Claude |
Use Replit when: You want a full IDE, need collaboration, or want built-in PostgreSQL. Replit is the most complete environment for iterating on AI-generated code.
Frequently asked questions
What prompts work best with Replit Agent?
Detailed, structured prompts work best. Describe your app functionality, tech stack preferences, and security requirements upfront. Replit Agent builds iteratively, so include success criteria it can verify. Prompts like "build a todo app with auth" work, but adding "use parameterized queries" and "validate all input" prevents security gaps.
How do I make Replit generate secure code?
Include security requirements directly in your prompt. Replit Agent follows instructions, so phrases like "never hardcode secrets," "use environment variables," and "validate input with Zod" guide it toward secure patterns. Always review the generated .replit and secrets configuration before deploying.
Can I use Replit for production apps?
Yes, Replit supports production deployments with their Deployments feature. However, review security before going live: check for hardcoded secrets, verify auth middleware is on all routes, and ensure database queries use parameterized queries. The speed is real - just add a security review step.
What is Replit Agent vs Replit Ghostwriter?
Replit Agent is the newer AI that builds complete apps from prompts, similar to Bolt.new or Lovable. Ghostwriter is the inline code completion assistant that works as you type. Use Agent for new projects and Ghostwriter for editing existing code. Both benefit from clear, security-focused instructions.
How do I use secrets in Replit?
Use the Secrets tab in Replit (lock icon in the sidebar). Add your API keys and credentials there, then access them via process.env.SECRET_NAME in your code. Never hardcode secrets in files. Prompt Replit with "use Replit Secrets for all API keys" to ensure proper handling.
What is the difference between Replit and Bolt.new?
Replit is a full development environment with hosting, databases, and collaboration features. Bolt.new is focused on rapid app generation. Use Replit when you need an IDE, want to iterate extensively, or need built-in hosting. Use Bolt.new for quick prototypes you will export elsewhere.
Does Replit Agent support databases?
Yes. Replit has built-in PostgreSQL and key-value storage. You can also connect external databases. When prompting, specify "use Replit PostgreSQL" or "connect to Supabase" depending on your needs. Always include "use parameterized queries" in your prompt to prevent SQL injection.
Scan Your Replit Projects
Prompts help, but they don't catch everything. VibeShip Scanner automatically detects hardcoded secrets, SQL injection, and missing auth in your Replit-generated code.
Scan Your Code FreeRelated guides
External resources
- Replit - Official site
- Replit Documentation
- Replit Agent Docs
- OWASP Top 10 - Security reference