AI agents, DNS delegation, and TLS attestation

IETF lists focused on authentication for AI agents, refined DNS delegation controls, and the threat model for attested TLS. Parallel threads covered post-quantum SSH identifiers, an agent commerce audit layer, power-aware routing, and an IAB position on age restrictions.

WIMSE adopts AI agent authentication work

The WIMSE working group opened a call for adoption of the draft on AI agent authentication and authorization (draft-klrc-aiagent-auth). Discussion produced broad support along with technical comments on delegation, multi-hop security, identifiers, attestation, and non-HTTP transports. Standardized agent credentials matter as autonomous systems begin to act across organizational boundaries.

DNSOP considers DELEXT delegation type ranges

Authors of the DELEG and DELEXT work proposed four delegation type ranges that govern NS inclusion and referral behavior. The DNSOP list examined the resulting semantics and the implications for resolvers. Explicit type ranges would give operators finer control over how referrals and name-server sets are constructed and interpreted.

SEAT debates attested TLS threat model

The SEAT working group discussed the proper scope of the TLS 1.3 threat model for attested TLS, including whether WeakDH and WeakHash belong and how far formal results from ESORICS and AsiaCCS should be treated as authoritative. Agreement on the model will determine which security properties the group can claim. Implementers need a stable baseline before deploying attestation extensions.

SSH ML-DSA code-point request paused

An IANA request for SSH public-key algorithm names covering ML-DSA remains held pending a draft merge. A Go implementer asked the designated experts to reactivate the registration. Timely code points affect how quickly SSH deployments can adopt lattice-based signatures.

Cedulon audit layer bypasses reported

On the SCITT list, participants described two bypasses in the Cedulon reference implementation that allow off-book settlements to evade audit, one through an extract-versus-settlements mismatch and another through self-signed keys. The authors had invited the community to break the design. Closing such gaps early is required if agent-to-agent commerce is to produce reliable audit trails.

LSR power-group draft faces scope questions

The LSR working group ran an adoption call for draft-many-lsr-power-group and received cross-WG comments on its relationship to GREEN YANG models and a TEAS dependency. The draft would let operators express power-related groupings inside link-state routing. Clear scope boundaries across the involved working groups will decide whether the work remains focused.

IAB statement on age-based restrictions

The IAB published a statement addressing age-restriction mandates and online safety. The follow-on thread examined service scale, moderation limits, and the role of decentralization. Protocol architecture choices intersect with regulatory pressure on age verification, making the IAB position relevant to future standards work.