Automated Detection of Under-Constrained Circuits in Zero-Knowledge Proofs

Explore a groundbreaking technique for detecting vulnerabilities in zero-knowledge proof circuits presented at PLDI 2023. Dive into the world of arithmetic circuits and domain-specific languages used in constructing zero-knowledge proofs. Learn about the critical issue of underconstrained circuits and their potential for exploitation in blockchain applications. Discover how the proposed method combines SMT solving with uniqueness inference to effectively reason about circuit constraints. Examine the implementation of this approach in the QED2 tool and its impressive results in analyzing Circom circuits. Gain insights into the detection of previously unknown vulnerabilities in widely-used circuits and the implications for cryptocurrency security. Understand the importance of semantic reasoning over finite field equations in verifying the uniqueness of output signals and preventing malicious exploitation of zero-knowledge proof systems.

[PLDI'23] Automated Detection of Under-Constrained Circuits in Zero-Knowledge Proofs