This artifact accompanies the paper titled “FlashSyn: Flash Loan Attack Synthesis via Counter Example Driven Approximation,” and offers a reusable Python implementation of FlashSyn, with scripts to replicate the main experiments. The artifact can be found on GitHub or Zenodo.

The artifact’s purpose is to enable exploration of our novel framework for automating the synthesis of flash loan attack vectors in decentralized finance (DeFi) protocols. Our artifact will generate flash loan attacks targeting 16 benchmarks in the paper. Our artifact will first rerun the experiments and reproduce experimental data. Then a separate script will be provided to extract tables in our paper from the experimental data.

The artifact assumes proficiency in Python (required) and familiarity with smart contract testing frameworks like Foundry (optional) for effective utilization. We provide a Docker image for Linux AMD, which was tested on another Ubuntu desktop. We recommend reviewers use a Ubuntu desktop with at least 16GB RAM and 4 processors for evaluation.