This repo provides essential security rules for Cursor.
⚠️ Why do you need these rules?By default, Cursor can generate unsafe code—such as exposing secrets or running dangerous commands. These rules act as guardrails to help you avoid risky patterns and enforce best practices automatically.
These rules aim to enforce safe coding practices, control sensitive operations, and reduce risk in AI-assisted development.
✨ Simply add these rules to your .cursor/rules directory (or your main directory) and you'll instantly be safer.
- 🔒 Secure Development Principles
- 🤖 Secure MCP Usage
- 🐍 Python Security Best Practices
- 🕵️♂️ No Secrets in Frontend
- 🚫 No Unsafe System Commands
- ✅ Enforce safe coding practices
- 🛑 Prevent accidental exposure of secrets
- 👮♂️ Control sensitive operations
- 🤝 Foster a security-first development culture
Are you a security researcher or developer passionate about AI safety? If you have ideas to improve these rules or want to add new ones, we encourage you to contribute! Your expertise can help make AI-assisted development safer for everyone 📝
We are Matan Kotick and Amit Ziv, security researchers specializing in the field of AI Agents.
Contact us if you have any questions about this interesting topic!