Fuzzing Zero-Knowledge Infrastructure
11-15, 10:30–11:00 (Asia/Bangkok), Stage 3

Zero-knowledge (ZK) infrastructure is highly complex and highly critical for the correct operation of L2 chains; that is, a single bug can result in massive financial and reputational damage. To find such potential million-dollar bugs before they are exploited, we have developed a novel fuzzing technique that can find logic flaws that impact liveness or safety of ZK infrastructure. Our fuzzer has already found 16 such issues in four ZK systems, namely Circom, Corset, Gnark, and Noir.

Valentin is a researcher at Consensys Diligence. He has years of experience in developing tools and techniques that help us in producing more reliable and correct software. He received a Ph.D. in Computer Science from ETH Zürich and has been a lead developer for several software analysis tools. At Consensys Diligence, he focuses on various R&D efforts, including the Harvey smart-contract fuzzer, the Scribble specification language, and several specialized fuzzers for critical infrastructure.