Show HN: Cryptographic proofs that algorithms stay fair over time [pdf]

This technical brief introduces a novel dual-baseline verification framework utilizing Zero-Knowledge Proofs (SNARKs) to enhance algorithmic fairness. It tackles a critical gap in traditional auditing by not only verifying that declared rules were followed (Symbolic Consistency) but also ensuring that rule application remains consistent across time and stakeholders (Temporal Consistency). The framework prevents selective rule enforcement and post-hoc manipulation.

The current implementation leverages Circom + SnarkJS with a roadmap for integrating Succinct's recursive proving system (SP1) for enhanced scalability. This system is actively deployed for collaborative research, enabling verifiable claims on areas like long-term bias detection, stakeholder equity, and behavioral consistency in algorithmic and social systems.

Current Deployment: Full working system with dual research applications:

Technical Documentation: https://docs.skyless.network — Comprehensive framework documentation with dual-lens analysis (governance verification + behavioral social trust) Live Prototype: https://www.meetskyla.com — Interactive demonstration of cryptographic verification of symbolic state transitions Open Source Architecture: https://github.com/skylessdev/skyla/blob/main/ARCHITECTURE.m... The current implementation runs production-ready proof engines using Circom + SnarkJS with full API functionality. The system supports chained verification via previousProofHash, with symbolic reasoning, recursive state logic, and identity mathematics actively deployed for collaborative research.