A vc-compatible zero-knowledge framework for selective academic grade verification

Abstract

Academic transcript verification suffers from an "all-or-nothing" disclosure problem: students must reveal complete records even when verifiers only need specific threshold confirmations. This thesis proposes, implements, and evaluates a privacy-preserving system that enables threshold verification (e.g., "Mathematics ≥ 80") without disclosing exact grades. The system integrates W3C DID/VC standards with Groth16 zero-knowledge proofs. Salted Merkle tree commitments enable O(log N) selective disclosure, while a 3-step verification protocol cryptographically binds Issuer-signed payloads to ZKP public inputs. Evaluation demonstrates practical feasibility: proof generation completes in approximately 0.3 seconds, verification in approximately 10ms, with proof packages of approximately 1 KB. Security analysis confirms computational soundness, zero-knowledge properties, and resistance to key-substitution and replay attacks. Key contributions include a verification binding model ensuring VC-ZKP consistency, self-certifying did:key resolution mitigating key substitution attacks, and an end-to-end prototype that uses W3C VC data model for credential issuance while defining a research-specific verification package for privacy-preserving presentation.