Thanks, Aaron and Karl for your answers.
I wasn't aware of the ongoing discussion concerning discovery, thanks for
pointing it out. It all makes more sense now.

Regards,
Judith

On Fri, Dec 5, 2025, 17:43 Karl McGuinness <[email protected]> wrote:

> A key advantage of having a new API resource is that the resource *could*
> be protected and the response contextual to the requesting client.  A
> challenge that exists today with Authorization Server Metadata is that it
> doesn't support per-client responses which creates complications with
> change management and discovery that would benefit from per-client metadata
> values .  Additionally scopes_supported already doesn't support resource
> based scopes for use with resource indicator requests without forcing
> implementers to construct resource+scope URIs as scope names (e.g
> https://app.example/resource-name/scope:read).  We basically need a
> audience+resource+scope tuple to fully qualify a grant for a Resource
> Authorization Server than can have N resources each with Y scopes which is
> what is being proposed in the PR.
>
> -Karl
>
> On Fri, Dec 5, 2025 at 8:20 AM Aaron Parecki <aaron=
> [email protected]> wrote:
>
>> For all practical purposes, it is not likely that an enterprise IdP will
>> want to publish a list of all possible Resource Authorization Servers and
>> scopes that are configured. We do not expect the IdP's Authorization Server
>> Metadata to include data from other trust domains.
>>
>> There is an open issue to discuss how a Client can discover the possible
>> Resources it can obtain an ID-JAG for from the IdP:
>> https://github.com/oauth-wg/oauth-identity-assertion-authz-grant/issues/52
>>
>> The current thinking is that this would be better accomplished by
>> defining a new API resource at the IdP, analogous to "userinfo", which
>> returns a list of resources supported based on the context of the user
>> presented in the request. This could then also include the list of scopes
>> per resource.
>>
>> Aaron
>>
>>
>> On Fri, Dec 5, 2025 at 12:10 AM Judith Kahrer <judith.kahrer=
>> [email protected]> wrote:
>>
>>> Hello,
>>>
>>> while studying draft 01 of the Identity Assertion JWT Authorization
>>> Grant, I started wondering how the fact that the IdP Authorization Server
>>> handles scopes for a different trust domain impacts the ecosystem like the
>>> authorization server metadata.
>>> To my understanding the IdP Authorization Server supports scopes for a
>>> different - possibly multiple - trust domains. Doesn't that affect the
>>> meaning of the scopes_supported property in the authorization server
>>> metadata? Shouldn't the IdP Authorization Server in its authorization
>>> server metadata also include which trust domains and what scopes for those
>>> trust domains it supports? Currently, section 6 "Authorization Server (IdP)
>>> Metadata" only specifies the urn:ietf:params:oauth:token-type:id-jag as
>>> a value for identity_chaining_requested_token_types_supported. I get
>>> the feeling that this is not enough. What do you think?
>>>
>>> Best regards,
>>> Judith Kahrer
>>> _______________________________________________
>>> OAuth mailing list -- [email protected]
>>> To unsubscribe send an email to [email protected]
>>>
>> _______________________________________________
>> OAuth mailing list -- [email protected]
>> To unsubscribe send an email to [email protected]
>>
>
_______________________________________________
OAuth mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to