A Symbolic Execution-Based Approach for Smart Contract Vulnerability Detection

Abstract

Before deploying smart contracts to Ethereum, a crucial step is to review the contract code for potential security vulnerability. Symbolic execution is currently a prevalent method for detecting vulnerability in smart contracts, but it lacks robust support for arbitrary modifications by owners. Therefore, this paper, leveraging symbolic execution technology, investigates detection methods specifically addressing this type of vulnerability, and presents concrete implementation and experimental validation. In the initial phase, this paper conducts an in-depth study of vulnerable contracts by debugging the source code and Ethereum Virtual Machine (EVM) opcode instructions. The analysis encompasses opcode instructions and the contract’s global state, summarizing vulnerability characteristics, extracting crucial opcode instructions. Subsequently, based on symbolic execution technology, the paper proposes corresponding detection methods. Real-world smart contracts are employed in this study, categorized into a dataset of vulnerable contracts susceptible to attacks and a dataset of normal contracts. Experimental evaluations are conducted to assess the effectiveness and accuracy of the system’s detection capabilities. The results indicate that the system implemented in this paper achieves the intended design goals and enhances the efficiency of vulnerability detection.