Privacy-Preserving Integrity Evidence for Student-Society Voting-Adjacent Workflows: A Phase C Pilot of Project Simurgh at Macquarie University

This record is a supplementary preprint to the main Project Simurgh architecture paper, “Project Simurgh: Privacy-Preserving Device Integrity Proofs for Capture-Resistant High-Stakes Sessions”.

It reports a bounded Phase C case study of Project Simurgh in a student-society voting-adjacent workflow at Macquarie University. The study evaluates whether a consent-based shadow-mode system can collect privacy-preserving session-integrity evidence alongside a real student-society voting event without recording ballot choices or affecting the official result.

Project Simurgh enforced ballot-choice exclusion through explicit submit-call JSON construction, client-side ballot-field exclusion safeguards, and server-side forbidden-field rejection. After data collection ended, a collection-closure flag returned HTTP 410 Gone on all write routes. The system used an HMAC-SHA-256 audit chain to record session lifecycle events while avoiding ballot choices, screen recordings, webcam/audio data, typed content, pasted content, direct identifiers, and personal device identifiers.

The Phase C pilot recorded 31 consented sessions, of which 30 submitted sessions formed the primary analysis dataset and one withdrawn session was excluded from analysis. The report builder emitted ballot_choice_recorded_by_simurgh: false as an implementation-enforced invariant, and the closeout gate suite verified this invariant after the 30 submitted sessions. At closeout, 359/359 automated tests, 8/8 smoke gates, 10/10 security-audit gates, and 5/5 closure gates passed.

This work is voting-adjacent and supplementary. It does not implement ballot cryptography, voter eligibility verification, coercion resistance, tally protection, fraud detection, or public-election certification. The contribution is a bounded empirical case study showing how the Project Simurgh architecture can be applied to a small student-governance setting.

Supplementary preprint v1.0; main Project Simurgh baseline: v0.4.18;