NDSS 2025 – MTZK: Testing And Exploring Bugs In Zero-Knowledge (ZK) Compilers

Why It Matters

ZK compilers are foundational to privacy‑focused blockchain solutions; undetected bugs can enable invalid proofs that jeopardize financial assets and trust in decentralized systems.

Key Takeaways

• MTZK uses metamorphic relations to test ZK compiler correctness
• Tested four major ZK compilers, found 21 bugs
• Developers patched 15 vulnerabilities quickly after disclosure
• Bugs enable malicious invalid proofs, risking crypto assets
• Framework provides scalable, automated testing for future compiler releases

Pulse Analysis

Zero‑knowledge proofs have become a cornerstone of privacy‑preserving protocols, from confidential transactions to identity verification. To make these proofs accessible, developers rely on domain‑specific languages and specialized ZK compilers that translate high‑level code into arithmetic circuits. While the convenience is undeniable, the correctness of the compilation step has received little systematic scrutiny, leaving a hidden attack surface that could be exploited by adversaries seeking to forge proofs and siphon value.

The MTZK framework tackles this gap by leveraging metamorphic testing—a technique that creates related input variants based on predefined relations and checks for consistent outputs. By mutating ZK DSL programs and observing whether the resulting circuits maintain logical equivalence, MTZK automatically flags incorrect compilations. In its evaluation, the researchers applied the method to four widely used industrial ZK compilers, uncovering 21 bugs, many of which allowed malformed proofs to pass verification. The rapid response from developers, who patched 15 issues, underscores both the severity of the flaws and the practical value of an automated testing pipeline.

For the broader blockchain ecosystem, MTZK’s findings serve as a wake‑up call. As decentralized finance and privacy‑centric applications scale, the integrity of underlying cryptographic tooling becomes a competitive differentiator. Integrating metamorphic testing into the development lifecycle can preempt costly exploits, bolster regulator confidence, and accelerate adoption of ZK technologies. Future work may extend MTZK to emerging proof systems and incorporate continuous integration hooks, ensuring that compiler updates remain secure by design.