Microsoft Uses AI Agents to Verify Rust Cryptography Code
Microsoft Research · rss · 2026-07-14
Microsoft Research published an article detailing how to scale formal verification of production-grade cryptographic algorithms using Rust, Lean, Aeneas, and AI agents. - **Background**: Cryptographic code demands extremely high security, which traditional testing cannot fully guarantee. Microsoft rewrote algorithms in SymCrypt (the crypto library used by Windows and Azure) using safe Rust and performed formal verification using the Lean framework. - **Method**: The Aeneas tool is used to translate Rust code into pure Lean models to prove compliance with standard specifications. AI agents were introduced to write independently verifiable proofs, thereby scaling automation. - **Progress**: A code branch containing complete proofs for SHA-3 and ML-KEM (post-quantum cryptography) has been open-sourced, with ongoing expansions to more algorithms like AES-GCM. - **Advantage**: Software engineers can continue writing high-performance Rust code while verification engineers work on the generated Lean models without interfering with each other.
More from coding & agent
- Codex turns out 123 screensavers in one playful batch — intellectronica · 2026-07-21
- Grok Build adds `grok doctor`, resumable sessions and remote image paste — mark_k · 2026-07-21
- Autoresearch proposes packaging ML runs as studies with questions, analysis, and code diffs — morgymcg · 2026-07-21
- CHAP defines approvals, handoffs, and audit logs for human-agent workflows — DeliveryTechnical199 · 2026-07-21
- The author says Codex reached 20x and is now debugging spec decoding on a hybrid parallel setup — TheZachMueller · 2026-07-21
- Axcess adds an MCP connector for WCAG accessibility checks that scanners miss — modelcontextprotocol · 2026-07-21