Real-time crypto scam detection powered by advanced EVM simulation
Sentinel Security protects you from crypto scams by analyzing transactions BEFORE they execute. Get instant warnings about honeypots, time-locked scams, and malicious contracts.
- 🎣 Honeypot Contracts: Tokens you can buy but never sell
- ⏰ Time-Bomb Scams: Contracts that work now but fail later
- 👑 Privilege Abuse: Owners can trade, but users cannot
- 💸 Hidden Fees: Excessive or variable transaction costs
- 🔄 Behavioral Drift: Contracts that change behavior over time
- 🎭 Proxy Deception: Malicious logic hidden in implementations
We have comprehensive documentation covering every aspect of the system:
- 🏛️ System Architecture: High-level design, components, and diagrams.
- 🔬 Analysis Pipeline: Deep dive into the 8 layers of detection logic.
- ⚙️ Setup & Deployment: Installation guide for Backend, ML, and Snap.
- 🔌 API Reference: Endpoints for the Backend and ML service.
- 👤 User Guide: How to use the Snap and interpret results.
- ❓ FAQ: Answers to common and technical questions.
Tests your transaction at multiple timestamps to detect delayed honeypots:
Current Block: ✅ Success
+1 Hour: ✅ Success
+7 Days: ❌ REVERTED → 🚨 TIME-BOMB DETECTED
Simulates the SAME transaction from different perspectives:
Random User: ❌ REVERTED
Contract Owner: ✅ SUCCESS → 🚨 HONEYPOT CONFIRMED
Traces execution to explain WHY transactions fail:
"❌ The contract checked who you are (CALLER) and compared it
against a stored whitelist (SLOAD). Since you're not on the
list, it blocked your transaction."
Monitors contracts over time:
"⚠️ Risk increased +40 since last scan 3 days ago"
"New flag: drain() function appeared after proxy upgrade"
graph TD
User["User / MetaMask"] -->|"Transaction Request"| Snap["MetaMask Snap"]
Snap -->|"Analyze (POST /rpc)"| Backend["Backend API"]
subgraph Backend System
Backend -->|Coordinate| Pipeline["Analysis Pipeline"]
Pipeline -->|"Fork Chain"| EVM["Global EVM Executor"]
Pipeline -->|"Fetch Risk"| ML["ML Service"]
Pipeline -->|"Store/Fetch"| Redis[("(Redis Cache)")]
end
ML -->|"Calibrated Score"| Pipeline
Pipeline -->|Verdict| Snap
Technology Stack:
- Backend: Node.js, TypeScript, Express
- EVM: @ethereumjs/evm (raw EVM simulation)
- ML Engine: Python, XGBoost, Scikit-learn
- Database: Redis (scan history)
- Ethereum Mainnet (Chain ID: 1)
- Polygon (Chain ID: 137)
- Optimism (Chain ID: 10)
- Arbitrum (Chain ID: 42161)
- Base (Chain ID: 8453)
- BSC (Chain ID: 56)
- Sepolia Testnet (Chain ID: 11155111)
We welcome contributions! Please:
- Fork the repository
- Create a feature branch
- Submit a pull request
MIT License - see LICENSE file for details
This project was developed for the KHacks hackathon to protect users from crypto scams.