Signature drafts, code points, and process fights
Post-quantum signature work dominated IETF lists, from a second-preimage gap in OpenPGP to overlapping TLS code-point requests and a scope question on hybrid unforgeability. Process arguments over an independent SSH submission and extension criticality in LAMPS ran alongside narrower ACME and EAP updates.
OpenPGP composite signatures have a second-preimage gap
Working group discussion on the OpenPGP list identified a second-preimage resistance gap in the composite post-quantum signature draft, described as a gap in RFC 9980. Authors propose raising minimum digest sizes and will rerun working group last call. The change matters for anyone implementing composite OpenPGP signatures, because digest size is part of the security argument the draft is trying to close.
CFRG considers a strongly unforgeable hybrid
The Crypto Forum Research Group is discussing whether to scope work on a strongly unforgeable, SUF-CMA-preserving post-quantum/traditional hybrid signature, with replies requested by 30 October. A poll and the thread show interest, with one dissent on migration speed. The decision would determine whether the research group treats preservation of classical unforgeability as in-scope hybrid work.
CNSA 2 SSH profile accused of skipping working group review
Daniel Bernstein questioned whether draft-becker-cnsa2-ssh-profile reached publication by evading SSH working group review through an independent submission, citing BCP 92 and NSA ties. The ISE and others defended the process. The dispute is about whether a CNSA 2 SSH algorithm profile can advance outside ordinary working group consensus.
TLS hybrid code-point request overlaps an adopted draft
TLS working group members flagged a SignatureScheme code-point request in draft-beck-tls-hybrid-codepoints, a non-working-group draft, as a duplicate that overlaps the adopted draft-ietf-tls-composite-mldsa. Nine messages from five participants treated the overlap as a process problem, not a routine assignment. Duplicate code points would leave implementers with two claims on the same hybrid ML-DSA identifiers.
LAMPS argues the one-signature binding must be critical
On the LAMPS list, participants debated whether the binding in draft-ietf-lamps-one-signature-certs-04 must be a critical extension so legacy validators cannot ignore it. The case for criticality is that a non-critical binding allows silent misuse by software that does not understand the extension. How that bit is set determines whether one-signature certificates fail closed in older verifiers.
Working group split on signedDocumentBinding criticality
A separate LAMPS thread focused on whether the signedDocumentBinding extension in the same one-signature certificates draft must be critical. Six participants over sixteen messages did not settle the point. The choice controls how strictly document binding is enforced when a certificate carries only one signature.
ACME weighs a minimal update that drops ES256
The ACME working group discussed shortening draft-ar-acme-pqc-tlsjws, after a new version notification for draft-ar-acme-pqc-tlsjws-01, to a minimal update of RFC 8555 that would remove ES256 as mandatory to implement. The narrower text would change the signature baseline without a broader protocol rewrite. Certificate authorities that speak ACME would need to plan for clients that no longer treat ES256 as required.
EAP-AKA post-quantum draft needs fragmentation
The EMU working group discussed fragmentation for post-quantum attributes and large SUCIs in draft-ietf-emu-pqc-eapaka-02, and noted that RFC 9048 needs an update. Existing message limits do not fit the larger post-quantum material. Mobile authentication that adopts these attributes will depend on that fragmentation path.