Wednesday 18 November 2026, 22:00 UTC
Notetakers: Thomas McCathy-Howe
Presentation:
https://datatracker.ietf.org/meeting/127/materials/slides-127-wimse-sessa-welcome-and-chair-updates-03
AI use guidance
Included Hackathon report out.
Draft: https://datatracker.ietf.org/doc/draft-ietf-wimse-wpt/
Presentation:
https://datatracker.ietf.org/meeting/127/materials/slides-127-wimse-sessa-workload-proof-token-01
Notes:
Brian Campbell: -02 status.
audFlemming Andreasen: inconsistent with the architecture and the HTTP
message signatures approach. Change made for logging is now
justified differently. Nothing says a service can't do authn and
authz at once. Needs more discussion.
Justin Richer: contention evident; take to the list.
Jonathan Rosenberg: agent acting for a user (Expedia agent calling
United for Joe) needs both the user's access token and agent
authentication; single header blocks that.
Jeff Lombardo: mTLS still lets you pass an access token; this
removes that. WWW-Authenticate doesn't require the credential to be
only the Authorization header.
Chair: run it to ground on the list and GitHub.
Draft: https://datatracker.ietf.org/doc/draft-ietf-wimse-arch/
Presentation:
https://datatracker.ietf.org/meeting/127/materials/slides-127-wimse-sessa-wimse-architecture-00
Notes:
Draft: https://datatracker.ietf.org/doc/draft-ietf-wimse-aims/
Presentation:
https://datatracker.ietf.org/meeting/127/materials/slides-127-wimse-aims-02
Notes:
Brian Campbell (presenting for the authors):
W2W auth proves who's calling on every request, works with SPIFFE
OAuth provides delegation in multiple modes (with uers or not)
PxP authorization: what can a request do
Transaction tokens inside the trust boundary
Fleming: does using access token on the outside preclude you from using
access token on the inside? Do you start with Tx Token?
Brian: definitely not saying that
IDJAG allows identities to cross domains
Shared signals: subscribe to udpates that may affect a posture of a
call
Goal is to localte underserved areas in other work
New work is starting/progressing in other groups based on discussion of
AIMS framework
Yaron: "the name sucks"
Brian: please help identify where improvements can be made; list of gaps
is more contentious
Dapeng: could provide a PR to apply DTR
Naveed: What about historically verifiable identifiers?Does it cover
organization? PRs are available
Brian: we want to facilitate interop; want to make the document more
advisory, which in turn mean nothing would get down to verifying
specific actions and accountability
Dan Druta: thanks for building a rocket ship from a lego box. If the aim
is to reuse standards, how would you do things differently if you're
starting from requirements?
Brian: Fair question, I don't know how I'd do it differently because
I've been focused on doing it with what we have now. It might be easier
greenfield, but very few people are greenfield.
Draft:
https://datatracker.ietf.org/doc/draft-carleton-workload-authz-grant/
Presentation:
https://datatracker.ietf.org/meeting/127/materials/slides-127-wimse-workload-authorization-grant-wag-01
Problem it's trying to solve: platform hosts workloads, wants to cross a
trust boundary without registering each workload
WAG is only standardizing what crosses the trust boundary
not defining new oauth - reusing it for workloads
Workload Identity Federation is already a pattern
Valuable to get rid of static credentials (API keys)
Shouldn't be a big move to swith to WAG if they already have something
similar
Relation to AIMS: defines the pattern
Relation to SPIFFE: allows SPIFFE, doesn't require SPIFFE
"nickname" shouldn't be a policy-bearing claim, could change
Acting for users:
IDJAG on behalf of user
"Who started this", when the user is the primary principal but
workload is making the action
Probably out of scope:
Questions:
Breno:
Paul: good question, main piece of having them separate is the type,
helps branch when you're processing it
Atul: there's no binding between things like txn tokens and spiffe,
we're repeating that between IDJAG and WAG
Paul: actor profile; use pieces of WAG in IDJAG or vice versa, depending
on who primary is
Jeff: Make sure you carry over all the latest IDJAG
Aaron (editor of IDJAG): Should we use IDJAG?
Nick Steele: this is for workload identity, IDJAG is useful for other
context (delegation)
Karl: OAuth already has software acting for itslef, it's already a
challenge in OAuth
Arndt: If you federate SPIFFE, you have to reflect the lifecycle on
another provider
Draft:
https://datatracker.ietf.org/doc/draft-heigh-wimse-workload-identity-metadata/
Presentation:
https://datatracker.ietf.org/meeting/127/materials/slides-127-wimse-sessa-workload-identity-metadata-00
Workloads have lots of properties, would be useful to present them when
you're presenting the identity
SPIFFE-land encodes properties in the identifier
Proposing two ways to deliver it
In WIT
Workload Metadata API Endpoint
Benefits:
Downside: requires a network roundtrip
Open questions
Fleming: interesting, raises a bunch of questions
Aaron: nitpicky, first thought was "this is a static file", maybe should
be called something different because it's an API
Jeff: soudns more like an introspection than metadata; worried about the
lookups
Omri Gazitt: WRT last question (SCIM structured objects vs array of
strings), we ran into this with the AuthZEN Profile for Authorization
Claims in JWT Access Tokens. We went with array of strings
https://openid.github.io/authzen/authzen-oauth-authorization-claims-1_0.html#name-rendering-the-claim
Justin: what about schema?
Yaroslav: will define schema, not registry
Danny Zollner: there's value here, maybe better handled with SCIM
Yaroslav: SCIM is for more static things, this will be more agile/quick,
not a push
Danny: SCIM could be complementary
Arndt: could be helpful for authorization, esp with service identifiers
not instance
Yaroslav: this is all signed by the issuer of the trust domain
Mirja: if we're taking metadata out of band, where in the process do you
need which metadata?
Karl: lot of the similar claims are in WAG, are there multiple
endpoints?
Yaroslav: we need to collaborate
Dan Druta: how much is under SBOM?
Shlomo Heigh: To clarify if my 10 seconds was too quick: The original
use case driving this idea was indeed service ids. We want to authorize
on service ids but not lose the ability to identify the individual
workload.
Draft:
https://datatracker.ietf.org/doc/draft-reece-wimse-cross-org-delegation/
Presentation:
https://datatracker.ietf.org/meeting/127/materials/slides-127-wimse-sessa-hackathon-report-02
Covers items in WIMSE and OAuth
Test vectors for requirements published as a separate artifact
=======================================================================
Thursday 19 November 2026, 22:00 UTC
Notetakers:
Presentation:
https://datatracker.ietf.org/meeting/127/materials/slides-127-wimse-sessb-welcome-and-chair-info-00
Draft:
https://datatracker.ietf.org/doc/draft-ietf-oauth-identity-chaining/
Presentation:
https://datatracker.ietf.org/meeting/127/materials/slides-127-wimse-sessb-identity-chaining-02
Presentation:
https://datatracker.ietf.org/meeting/127/materials/slides-127-wimse-sessb-oauthwimsespiffe-boundaries-00
Presentation:
Draft:
https://datatracker.ietf.org/doc/draft-gazitt-oauth-authzen-issuance/
Draft:
https://datatracker.ietf.org/doc/draft-gazitt-oauth-authzen-token-exchange/
Draft:
https://datatracker.ietf.org/doc/draft-gazitt-oauth-authzen-claims/
Presentation:
https://datatracker.ietf.org/meeting/127/materials/slides-127-wimse-sessb-openid-authzen-oauth-00
Draft:
https://datatracker.ietf.org/doc/draft-ietf-oauth-spiffe-client-auth/
Presentation:
Draft:
https://datatracker.ietf.org/doc/draft-fletcher-transaction-token-chaining-profile/
Presentation:
Draft:
https://datatracker.ietf.org/doc/draft-araut-oauth-transaction-tokens-for-agents/
Draft:
https://datatracker.ietf.org/doc/draft-araut-oauth-transactiontokens-bcp/
Presentation:
https://datatracker.ietf.org/meeting/127/materials/slides-127-wimse-sessb-transaction-tokens-for-agents-txn-tokens-bcp-00
Draft:
https://datatracker.ietf.org/doc/draft-tulshi-oauth-transactional-access-tokens/
Presentation:
Draft:
https://datatracker.ietf.org/doc/draft-mcguinness-oauth-actor-profile/
Presentation: