[COSE] Signed object format review: PPS Pulse envelope vs COSE/JOSE

[Hossein Hezami <[e-mail]>]

Hi,

I would appreciate feedback on the signed-object format used in an
individual Internet-Draft:

Title: PulseProof Sentinel Protocol
Draft: draft-hezami-pulseproof-sentinel
Datatracker:
https://datatracker.ietf.org/doc/draft-hezami-pulseproof-sentinel/
Repository: https://github.com/pps-protocol/pulseproof-sentinel
Documentation: https://pps-protocol.github.io/pulseproof-sentinel/docs/

PPS defines a signed authentication artifact called a Pulse. The Pulse
envelope is currently a CBOR map containing:

- the signed message bytes
- a signature map
- optional rotation certificates

The mandatory profile uses Ed25519 signatures over deterministic
CBOR-encoded
payloads.

The protocol is intended for offline OTP-style authentication, constrained
terminals, transaction signing, silent duress signaling, and offline
multi-device approval.

I would appreciate feedback on the following questions:

1. Would COSE_Sign1 or COSE_Sign be a better fit for the PPS Pulse envelope?
2. Are there advantages to using COSE protected/unprotected headers instead
   of the current custom CBOR map?
3. Is the algorithm identifier approach compatible with COSE algorithm
   registration practices?
4. Would using COSE improve interoperability with existing implementations?
5. Are there COSE extensions or profiles that would be relevant for
   time-bound authentication proofs?

The draft currently defines a custom CBOR envelope because it needs to carry
protocol-specific fields such as epoch, counter, rotation sequence, duress
state indicators, threshold signatures, policy hash, and context hash.

I am open to changing the format if COSE provides a more interoperable and
well-reviewed approach.

Thanks,
Hossein Hezami
Independent