BeTrueCore Makes Vote-Buying Unverifiable and Sybil Attacks Unprofitable

Vote-buying collapses when the commodity is unverifiable. That’s the core engineering insight behind BeTrueCore, a modular governance stack that turns exploits into self-destructing economic traps.

The system uses mid-session choice mutability and continuous MACI key rotation to guarantee receipt-freeness. A participant can show an external buyer any intermediate action as proof of compliance, but the buyer cannot mathematically verify the final time-locked choice. Because the internal currency, Voting Weight of a Unit (VWU), is strictly inalienable, the transaction lacks an economic subject. Put simply: you can’t sell what a buyer can’t confirm.

Sybil Attacks Require Building a Real Community

Mass synthetic identities are filtered across five non-contiguous layers. L0 checks keystroke dynamics and behavioral biometrics. L1 binds each identity to a unique ZK-nullifier chain to prevent double-voting. L5 runs a read-only AI Sentinel that monitors for long-horizon synchronized activity. The practical consequence: to simulate a community, an adversary must actually build one. The cost shifts from capital expenditure to cognitive impossibility.

Seven Actions Per Day, No More, No Less

Each participant is restricted to a maximum of seven cryptographic actions per 24-hour session, requiring less than 60 minutes of total interaction. Variable participation frequency is allowed—weekly or monthly sessions incur no penalty on accumulated rank or badges. Inactivity simply flattens the dynamic VWU trajectory, preventing offline accounts from hoarding governance alpha. An adversary trying coordinated capture cannot rely on high-throughput scripts or passive account hoarding; every action demands distinct human cognitive engagement across irregular intervals.

Coordinator Compromise Can’t Corrupt the Data

The MACI coordinator must publish a mathematically verifiable state transition ZK-proof simultaneously with the result. A compromised coordinator can only affect availability, not integrity. The AI Sentinel layer runs strictly read-only, so any infrastructural breach grants observation rights but zero execution power. The architecture is designed so that the only way to shift system state is to simulate genuine human judgment—at scale and over time.

BeTrueCore is now in Phase 2 of its roadmap, working on circuit integration and MACI optimization ahead of an MVP launch on an EVM testnet. The open-source repository and preprint paper (The Notary Under Attack: An Adversarial Model for Cryptographic Collective Intelligence, DOI 10.5281/zenodo.21111544) invite feedback from ZK cryptographers and Solidity developers who want to test whether receipt-freeness holds under a stronger adversary model—say, one who observes all intermediate choices in real time.

Source: Designing Infrastructure Where Exploits Destroy Themselves
Domain: ethresear.ch