Add information about key agreement about tswg-did-method-webs-specification

[peacekeeper]

I don't know how key agreement works in KERI, can you point me to something?

I know that Ed25519 keys can be converted to X25519 for key agreement, and we could add such converted keys to the DID document. Some DID methods do that "automatically", and some don't. There are some use cases and ecosystems where this is not acceptable.

[swcurran]

There definitely needs to be a way to declare both the Verification and KeyAgreement, and have the controller explicitly do the derivation of (for example) Ed25519 to X25519 vs. the DID Method doing it for you.

[2byrds]

Sam and Phil provided information on key agreement:
KERI supports more than Ed25519 (ECDSA for instance would use the same curve, no translation). So specifying translation is problematic, NIST algorithms use the same keys as the signing keys. The right strategy would be similar to how we are handling endpoints, domain migration info, etc. Normally in KERI we would use BADA-RUN for key agreement (just like we do for endpoints). But we should support two choices; BADA-RUN the assertion for the key agreement encryption key or anchor the key agreement info to the KEL via an ACDC credential.

[2byrds]

I included this wording in #64
Happy to adjust in a future PR.