This is a first stab at what a categorical theory of cryptography might look like. It essentially gives a generalized definition of a cryptosystem in categorical terms in terms of a category of interactive computations.

One interesting thing is that the definition given is a common generalization both of cryptosystems and error-correcting codes.

The document is pretty incomplete and doesn't cover some other ideas I had about how to formalize interactive proof systems and one-way functions in this setting.