Tagline: Secure your Sui assets from backdoors.
Backdoor is an AI-powered Web3 security assistant built specifically for the Sui ecosystem. It empowers users to proactively scan for smart contract vulnerabilities and risky token approvals before interacting with decentralized applications. If a breach has already occurred, Backdoor helps users analyze compromised wallets to pinpoint how the exploit happened and trace stolen assets across the network, turning reactive loss into proactive security.
Since this project is built entirely using Google AI Studio's powerful prototyping tools, you can interact with the live application directly here: 👉 Launch Backdoor on Google AI Studio
- Proactive Vulnerability Scanning: Scan Sui smart contracts and token approvals for hidden risks before interacting with them.
- Drain Analysis: Upload or paste historical data from a compromised wallet to break down exactly how the exploit occurred.
- Fund Tracing Assistant: Receive guidance on tracking stolen assets and identifying malicious addresses on the Sui blockchain.
- Sui-Centric Intelligence: Tailored specifically around Sui network primitives, object ownership models, and Move language security.
system_prompt.txt: The core AI architecture, system rules, and security logic running behind the app.docs/sample_test_cases.md: Pre-made test prompts for hackathon judges to copy-paste and try out.
- Click the Live Application Link above.
- Open our Sample Test Cases Guide to find copy-and-paste examples to test the AI's response accuracy!
AI Collaboration & Acknowledgments This project was proudly developed in partnership with Google's AI ecosystem:
- Google AI Studio: Used as the primary IDE, sandbox environment, and rapid prototyping platform.
- Gemini 1.5 Pro / Flash: Acted as the core security co-pilot, assisting with Web3 Move language threat analysis and system prompt engineering.