Fault Injection Attacks Against zkSTARKs

Alexander Dalton; Markus Schofnegger; Daniel Page

Paper 2026/835

Fault Injection Attacks Against zkSTARKs

Alexander Dalton

, Independent

Markus Schofnegger, [[alloc] init]

Daniel Page, University of Bristol

Abstract

Fault injection attack targetting schemes with Zero-Knowledge (ZK ) properties have been relatively absent in the wider literature. One of the few examples has recently shown a ZK signature scheme to be vulnerable to fault injection attacks. In this paper we detail candidate fault injection attacks against Zero-Knowledge Scalable Transparent Argument of Knowledge (zkSTARK) provers, designed to violate the construction’s zero knowledge capabilities. zkSTARK proving systems are complex, with a huge amount of diversity in implementation specifics. We match the variety within the STARK implementation ecosystem, proposing a variety of ffault injection attacks against different algorithmic primitives. To the best of our knowledge this marks the first exploration of the fault injection surface of zkSTARKs, and of the wider class of general purpose ZK proving systems.

Note: * Expanded experiments to feature more fine-grained instruction level simulation of the attacks * Added attacks against recursive and DIF varients of Cooley--Tukey to complement existing DIT attack

Metadata

Available format(s): PDF
Category: Attacks and cryptanalysis
Publication info: Preprint.
Keywords: Fault Injection Zero Knowledge zkSTARK
Contact author(s): research @ alexdalt co uk
History: 2026-07-16: revised; 2026-04-28: received; See all versions
Short URL: https://ia.cr/2026/835
License: CC BY

BibTeX

@misc{cryptoeprint:2026/835,
      author = {Alexander Dalton and Markus Schofnegger and Daniel Page},
      title = {Fault Injection Attacks Against {zkSTARKs}},
      howpublished = {Cryptology {ePrint} Archive, Paper 2026/835},
      year = {2026},
      url = {https://eprint.iacr.org/2026/835}
}