The majority of deployed asymmetric cryptography implementations (RSA, DH, ECDH/ECDSA with GF(p) curves) need to perform calculations on integers that are larger than a single machine word. Just like every software package, implementations of multi-precision integer arithmetic sometimes have bugs. This talk investigates the implications of these bugs and shows how they can be used by attackers to exploit asymmetric cryptographic primitives. Isolating bug patterns and understanding exploitation requirements allows us to develop strategies for automated bug hunting.