Advancements in Smart Contract Security and Development

The field of smart contract research is moving towards leveraging large language models (LLMs) and innovative detection methods to improve security and development. Recent studies have focused on applying LLMs to detect vulnerabilities in smart contracts beyond the Ethereum Virtual Machine (EVM) ecosystem, such as in Solana and Algorand. Additionally, there is a growing interest in developing automated systems for detecting logic-level usage violations of reusable components in smart contracts. Furthermore, researchers are exploring the use of LLMs to generate executable and validated proof-of-concept (PoC) tests for smart contract bug reports. The development of novel EVM bytecode representation methods, such as the Stable-Semantic Graph (SSG), is also gaining attention for accurate similarity detection and vulnerability identification. Noteworthy papers include: The paper on SCRUTINEER, which presents a system for detecting logic-level usage violations of smart contract reusable components with high precision and recall. The SmartPoC paper, which introduces an automated framework for generating executable and validated PoC tests from textual audit reports, achieving high success rates on benchmark datasets.

Sources

Prompt Engineering vs. Fine-Tuning for LLM-Based Vulnerability Detection in Solana and Algorand Smart Contracts

SCRUTINEER: Detecting Logic-Level Usage Violations of Reusable Components in Smart Contracts

SmartPoC: Generating Executable and Validated PoCs for Smart Contract Bug Reports

Esim: EVM Bytecode Similarity Detection Based on Stable-Semantic Graph

Beyond Code Similarity: Benchmarking the Plausibility, Efficiency, and Complexity of LLM-Generated Smart Contracts

Built with on top of