As part of the re-encryption correctness proof (which in PR#115 is always included together the result of the re-encryption), it is possible to take some optional "metadata" information into consideration. This "metadata" is intended to serve as additional information that binds the re-encryption request with its external context.

For example, in the case of NuCypher KMS, this metadata could include a timestamp (so that Ursula cannot respond with an old re-encryption and/or correctness proof), the policy ID (so Ursula cannot reuse work done for other policies, even for the same capsule, Alice & Bob), etc.

Do we want to leave this format arbitrary? This has implications wrt to (de)serialization of CorrectnessProofs and, consequently, of CFrags