Vulnerability and Transaction behavior based detection of Malicious\n Smart Contracts

Abstract

Smart Contracts (SCs) in Ethereum can automate tasks and provide different\nfunctionalities to a user. Such automation is enabled by the `Turing-complete'\nnature of the programming language (Solidity) in which SCs are written. This\nalso opens up different vulnerabilities and bugs in SCs that malicious actors\nexploit to carry out malicious or illegal activities on the cryptocurrency\nplatform. In this work, we study the correlation between malicious activities\nand the vulnerabilities present in SCs and find that some malicious activities\nare correlated with certain types of vulnerabilities. We then develop and study\nthe feasibility of a scoring mechanism that corresponds to the severity of the\nvulnerabilities present in SCs to determine if it is a relevant feature to\nidentify suspicious SCs. We analyze the utility of severity score towards\ndetection of suspicious SCs using unsupervised machine learning (ML) algorithms\nacross different temporal granularities and identify behavioral changes. In our\nexperiments with on-chain SCs, we were able to find a total of 1094 benign SCs\nacross different granularities which behave similar to malicious SCs, with the\ninclusion of the smart contract vulnerability scores in the feature set.\n