Skip to content

Latest commit

 

History

History
28 lines (22 loc) · 1.77 KB

File metadata and controls

28 lines (22 loc) · 1.77 KB

Proof Requirements

Current state

  • src/abi/Types.idr — Academic workflow types
  • src/abi/Layout.idr — Memory layout
  • src/abi/Foreign.idr — FFI declarations
  • No dangerous patterns in ABI layer
  • Claims: "mathematical guarantees that student data never reaches AI systems in identifiable form", GDPR compliance, privacy-first

What needs proving

  • Anonymization irreversibility: Prove that SHA3-512 anonymization cannot be reversed within any computationally feasible bound (this is a standard cryptographic assumption — formalize the reduction)
  • Data flow isolation: Prove that student PII never flows to AI model inputs without passing through the anonymization layer
  • GDPR right-to-erasure: Prove that deletion requests remove all linked data (no orphaned PII in caches or logs)
  • Grading integrity: If automated grading is involved, prove grade calculations are deterministic and auditable
  • Consent tracking completeness: Prove that all data processing is gated by valid, unexpired consent

Recommended prover

  • Idris2 — For data flow and information flow control properties (dependent types can track taint)
  • Coq — For the cryptographic reduction argument (if formalizing the SHA3 assumption)

Priority

  • HIGH — The suite explicitly claims "mathematical guarantees" about student privacy. If these guarantees are not formally verified, the claim is dangerous — especially in an educational context with minors' data and GDPR obligations.

Template ABI Cleanup (2026-03-29)

Template ABI removed -- was creating false impression of formal verification. The removed files (Types.idr, Layout.idr, Foreign.idr) contained only RSR template scaffolding with unresolved {{PROJECT}}/{{AUTHOR}} placeholders and no domain-specific proofs.