FFGDetector: Vulnerability Detection in Cross-Contract Feature Flow Graph Using GCN
Zhifei Liu, Jian Jiao · 2024
With the rapid development of blockchain technology, smart contracts have been widely used in various decentralized applications. However, the security issues of smart contracts, especially cross-contract vulnerabilities, have become important risks affecting the security of the blockchain ecosystem and are challenging to detect. To address this challenge, we propose FFGDetector, an innovative cross-contract vulnerability detection tool for smart contract based on improved GCN (Graph Convolutional Neural Network), which can automatically analyze the call relationship and data flow between smart contracts, generate and merge contract feature flow graphs. We have validated the effectiveness and accuracy of FFGDetector through extensive experiments on reentrancy, timestamp dependency, txorigin and delegatecall vulnerabilities and the results indicate that our method is more effective.