Enhancing Smart Contract Security with Explainable AI: A Framework for Re-entrancy Vulnerability Detection and Explanation

Abstract

Smart contracts, integral to decentralized applications, are unfortunately plagued by security vulnerabilities. Re-entrancy attacks pose a particularly insidious threat, allowing attackers to exploit subtle interactions within the contract’s code. While traditional detection methods exist, they often struggle with false positives and a lack of transparency, hindering the ability of developers to understand and fix the problems. This research introduces a novel framework that combines the power of Explainable AI (XAI) with a deep learning approach to address these shortcomings. Neural networks utilize the proposed BiLSTMs model to design a framework that detects re-entrancy patterns because they excel at analyzing complex dependencies across the length of smart contract code. The decision to explain models is enhanced through XAI techniques, which improve the entire process. The explanations reveal which code sections from the codebase contribute to vulnerability classification. Combining two detection methods intends to improve re-entrancy vulnerability detection performance while providing rapid remedy recommendations.