Succinct’s SP1 Faces Scrutiny Over Security Vulnerability Disclosure
Succinct’s SP1 ZKVM faces scrutiny after LambdaClass revealed a critical security vulnerability in its proof generation, stemming from two distinct flaws. This exploit was disclosed in collaboration with 3Mi Labs and Aligned.
Key points of the vulnerability:
- Missing Verification Step allowed manipulation of proof components due to lack of proper accuracy checks.
- Incomplete Proof Flag wasn’t consistently enforced, creating a potential loophole.
- Polynomial Evaluation Issue in Plonky3 failed to fully verify calculations before confirming valid proofs.
While Succinct addressed the vulnerabilities quickly, concerns about transparency in security practices for zero-knowledge virtual machines (ZKVMs) have emerged. Notable integrations of SP1 technology include:
- Mantle Network transitioning to a zero-knowledge validity rollup.
- AggLayer using SP1 for cross-chain interoperability security.
- Taiko adopting SP1 as a ZK prover for layer-2 execution.
- New project Soon developing an SVM rollup framework with ZK fault proofs powered by SP1.
LambdaClass emphasized that the implications of the flaw require further assessment. The interplay between the two issues means fixing one may not suffice. Developer Fede expressed concerns over Succinct’s urgency in communication regarding the issue.
Avail’s Anurag Arjun acknowledged the responsible action taken by Succinct but highlighted the need for better disclosure practices. He confirmed that Avail’s implementation was not exposed to risk due to reliance on Succinct’s permissioned proprietary prover.
Succinct's updated version 4, Turbo, resolves the identified vulnerabilities. The case illustrates that even well-audited code can contain bugs and underscores the importance of ongoing improvements in security.
The situation raises questions around balancing security, transparency, and user protection in the evolving landscape of ZKVM systems.