nirapod.org · Research · Approach · News · People · Security
Nirapod Labs is a research lab for high-assurance cryptography. We implement the standards that secure the internet and formally verify the code: right results, no secret-dependent timing and no memory errors, for every possible input. A machine checks every proof, so anyone can run the checker and get the same answer.
- Functional correctness. The code computes exactly the result the standard defines.
- Constant-time execution. Secret values never shape the path the code takes or the memory it touches.
- Speculative constant-time. Secrets stay hidden even while the processor speculates past a wrong guess.
- Memory safety. Every read and write stays inside memory the code may use.
Each one holds for every possible input. How we prove them
- nirapod-sha2: formally verified SHA-2 and HMAC in Jasmin, proved correct and memory safe in EasyCrypt, and checked constant-time and speculative constant-time.
- pasta-glv-proofs: machine-checked Lean 4 proofs of the wNAF recoding and Straus ladder in the GLV scalar multiplication of
pasta_curves.
Questions about our work, or want to work with us: hello@nirapod.org. Found a flaw in our code, proofs or website: security@nirapod.org, and see our disclosure policy.