Hybrid Post-Quantum primitive codecs

[ddebernardy]

The IETF working groups tied to COSE and HPKE are set to expand the use of hybrid PQ primitives that don't yet have a multicodec. Specifically:

https://datatracker.ietf.org/doc/draft-ietf-cose-hpke-pq-pqt/

• ML-KEM-768 + P-256
• ML-KEM-768 + X25519
• ML-KEM-1048 + P-384

https://datatracker.ietf.org/doc/draft-ietf-jose-pq-composite-sigs/

• ML-DSA-44 + P-256
• ML-DSA-65 + P-256
• ML-DSA-87 + P-384
• ML-DSA-44 + Ed25519
• ML-DSA-65 + Ed25519
• ML-DSA-87 + Ed448

It would be sweet if they got a codec, so those could be used with COSE and UCAN without needing to split hybrid keys as two separate did:keys (which has complex ramifications like split UCANs) or using custom codecs (which can break interoperability).

If I understood well, multikey is supposed to help:

https://github.com/cryptidtech/multi-key#supported-key-formats

but per the spec:

https://github.com/cryptidtech/provenance-specifications/blob/main/specifications/multi-key.md

"Comment : A Multi-Key object contains a string that describes the key. This is application specific."

hence my presumed need for a specific codec. (Or perhaps I'm just misunderstanding and I can use that with a did:key?)

Thanks in advance.