Authors, In my opinion it is better to keep the document focused on pure ML-KEM and possibly use the same document outline to create a separate draft for PQ/T hybrid registration. From the COSE implementation side, there is not a strong relationship between ML-KEM and hybrids (I will use a crypto library with different APIs for the different families and they will be tested independently). From an operational perspective, they serve different purposes and there is a strong value in having ML-KEM available soon with a separate future option for hybrid use.
Brian S. On Wed, Nov 18, 2026 at 3:33 AM Hannes Tschofenig <hannes.tschofenig= [email protected]> wrote: > Hi Mark! > > Thanks for the review comments. Here is a corresponding PR: > https://github.com/cose-wg/COSE_PQC_KEM/pull/28 > > A potential challenge with including support for hybrid KEMs is that it > could significantly delay the completion of the COSE-PQC-KEM document, as > it would create a dependency on draft-irtf-cfrg-concrete-hybrid-kems. > Hence, I am also cautious of adding support for hybrid KEMs. > > Ciao Hannes > > Am 17.11.2026 um 13:07 schrieb tirumal reddy: > > Hi Mark, > > Thanks for the review. > > 1. The draft has targeted pure PQC from day one, for deployments migrating > to PQC. Hybrids are out of scope. > 2. Good point, ML-KEM is IND-CCA2 on its own; the KDF plays no role in > that. > 3. I will update it to AKP only. > > Both fixes will be in -01. > > Cheers, > Tiru > > On Tue, 17 Nov 2026 at 06:51, Mark Atwood <[email protected]> wrote: > >> Hi all, >> >> Three things before a WGLC on pqc-kem: >> >> 1. Please register hybrids here too (the CFRG concrete hybrid KEMs, e.g. >> MLKEM768-X25519). LAKE builds on this draft and is ML-KEM-only until COSE >> has hybrid identifiers to point to. >> >> 2. Section 3 says the FO transform "is performed using the KDF". ML-KEM >> is already IND-CCA2 on its own (FIPS 203, 3.2). That paragraph needs fixing. >> >> 3. Section 8 (AKP) and Section 10.2 (Elliptic Curves registry) describe >> two different key models. Pick one first. >> >> Mark Atwood >> _______________________________________________ >> COSE mailing list -- [email protected] >> To unsubscribe send an email to [email protected] >> > > _______________________________________________ > COSE mailing list -- [email protected] > To unsubscribe send an email to [email protected] > > _______________________________________________ > COSE mailing list -- [email protected] > To unsubscribe send an email to [email protected] >
_______________________________________________ COSE mailing list -- [email protected] To unsubscribe send an email to [email protected]
