Solforge: uma abordagem baseada em modelos de linguagem de grande escala para o teste diferencial de compiladores Solidity

Abstract

ABSTRACT Solidity is currently the most widely used language for developing smart contracts on blockchain platforms such as Ethereum. Ensuring the correct execution of these contracts and the reliability of their compilers is essential, since errors can lead to financial losses and security breaches. However, traditional manual testing methods are not sufficient to expose all potential vulnerabilities in Solidity compilers. This work aims to develop a tool for the automatic generation of Solidity smart contracts that will serve as test cases for compiler analysis and validation. To enhance the efficiency and depth of the testing process, fuzzing strategies will be employed, enabling the exploration of a large number of execution paths and revealing unexpected compiler behaviors. This contributes directly to strengthening the robustness of the smart contract ecosystem.