WASIF: In-depth detection of vulnerabilities in Wasm smart contracts via information flows and function invocation sequences

Abstract

With the widespread adoption of WebAssembly (Wasm) smart contracts in popular blockchain platforms such as EOSIO, vulnerability attacks on Wasm smart contracts have become a serious problem. To protect the legitimate interests of blockchain users, it is necessary to detect vulnerabilities in Wasm smart contracts. However, detection faces a great challenge in that the source code of Wasm smart contracts is rarely released publicly. Although many approaches have made great progress in vulnerability detection, they still suffer from inefficiently generating function invocation sequences to track inter-function dependencies, ineffectively tracking sensitive information flows, and a considerable number of False Positives (FPs). To address these issues, we present a new concolic fuzzing approach for detecting vulnerabilities in Wasm smart contracts via information flows and function invocation sequences, namely, WASIF. Also, we implement the open-source prototype of the WASIF and conduct extensive experiments to evaluate it. The experimental results show that WASIF effectively and efficiently detects vulnerabilities in Wasm smart contracts and outperforms the state-of-the-art concolic fuzzer WASAI on most metrics.