We provide the first machine-checked proof of privacy-related properties (including ballot privacy) for an electronic voting protocol in the computational model. We target the popular Helios family of voting protocols, for which we identify appropriate levels of abstractions to allow the simplification and convenient reuse of proof steps across many variations of the voting scheme. The resulting framework enables machine-checked security proofs for several hundred variants of Helios and should serve as a stepping stone for the analysis of further variations of the scheme.
In addition, we highlight some of the lessons learned regarding the gap between pen-and-paper and machine-checked proofs, and report on the experience with formalizing the security of protocols at this scale.