[Openid-specs-ab] Key-binding and dpop scope

[Dick Hardt]

You can do whatever you want in your application Kosuke!

A standardized spec enables software written by different parties to
interoperate. From my recollection, your use case is a single party. You
can ignore the authentication request part if you want. You won't have
quite the same security characteristics if the client does not
include dpop_jkt as a parameter.

On Sat, Aug 30, 2025 at 2:09 PM Kosuke Koiwai <kkoiwai at gmail.com> wrote:

> Can we separate the spec to add cnf to id_token and how to ask for it?
>
> As I asked before, I want to add the key info to id_token, and the key is
> the same as the one used for attestation-based client auth.
>
> https://lists.openid.net/pipermail/openid-specs-ab/2025-August/010892.html
>
> I want to add the cnf claim at the discretion of IdP, so RP doesn’t have
> to add any to the scope.
>
> Kosuke
>
>
> On Sat, Aug 30, 2025 at 19:07 Dick Hardt via Openid-specs-ab <
> openid-specs-ab at lists.openid.net> wrote:
>
>> "bound_key" is crisper and says what is wanted in the token rather than
>> what is to be done
>>
>> On Fri, Aug 29, 2025 at 6:51 PM Dag Helge Østerhagen via Openid-specs-ab <
>> openid-specs-ab at lists.openid.net> wrote:
>>
>>> +1 for both "key_binding" and "cnf". Sigh.
>>>
>>> /dag
>>> ------------------------------
>>> *From:* Dick Hardt <dick.hardt at gmail.com>
>>> *Sent:* Friday, August 29, 2025 7:47:45 PM
>>> *To:* Artifact Binding/Connect Working Group <
>>> openid-specs-ab at lists.openid.net>
>>> *Cc:* Dag Helge Østerhagen <dag at udelt.no>; george at practicalidentity.com
>>> <george at practicalidentity.com>; Filip Skokan <panva.ip at gmail.com>
>>> *Subject:* Re: [Openid-specs-ab] Key-binding and dpop scope
>>>
>>> `key_binding` as scope name?
>>>
>>> On Fri, Aug 29, 2025 at 6:35 PM Dag Helge Østerhagen via Openid-specs-ab
>>> <openid-specs-ab at lists.openid.net> wrote:
>>>
>>> Well,  currently the dpop header is used to signal token binding (and
>>> inclusion of the cnf claim) for access and refresh tokens.   I don't see
>>> any other use cases in the (near) future.
>>>
>>> /dag
>>> ------------------------------
>>> *From:* george at practicalidentity.com <george at practicalidentity.com>
>>> *Sent:* Friday, August 29, 2025 7:01:54 PM
>>> *To:* Artifact Binding/Connect Working Group <
>>> openid-specs-ab at lists.openid.net>
>>> *Cc:* Dag Helge Østerhagen <dag at udelt.no>
>>> *Subject:* Re: [Openid-specs-ab] Key-binding and dpop scope
>>>
>>> My thought is that might depend on whether the ‘cnf’ scope is only
>>> applied to the id_token or whether cnf claims should be added to other
>>> issued tokens as well. Currently the proposed key-binding spec is specific
>>> to id_tokens.
>>>
>>> George Fletcher
>>> Identity Standards Architect
>>> Practical Identity LLC
>>>
>>>
>>>
>>> On Aug 29, 2025, at 12:56 PM, Dag Helge Østerhagen via Openid-specs-ab <
>>> openid-specs-ab at lists.openid.net> wrote:
>>>
>>> I like «id_token_cnf», but wouldn’t just «cnf» be more aligned with
>>> other oidc scopes?
>>>
>>> /dag
>>> ------------------------------
>>> *From:* Openid-specs-ab <openid-specs-ab-bounces at lists.openid.net> on
>>> behalf of george--- via Openid-specs-ab <
>>> openid-specs-ab at lists.openid.net>
>>> *Sent:* Friday, August 29, 2025 6:14:16 PM
>>> *To:* Dick Hardt <dick.hardt at hello.coop>
>>> *Cc:* george at practicalidentity.com <george at practicalidentity.com>;
>>> Artifact Binding/Connect Working Group <openid-specs-ab at lists.openid.net
>>> >
>>> *Subject:* Re: [Openid-specs-ab] Key-binding and dpop scope
>>>
>>> That makes sense to me; including ‘cnf’ in the scope name. Would we ever
>>> want to allow the “key binding” mechanism to use something other than DPoP?
>>> If so, and the express purpose is to provide key binding for the id_token,
>>> then I’d recommend something like ‘id_token_cnf’.  It’s specific, clear and
>>> doesn’t preclude methods other than DPoP to provide the necessary data for
>>> the cnf claim.
>>>
>>> George Fletcher
>>> Identity Standards Architect
>>> Practical Identity LLC
>>>
>>>
>>>
>>> On Aug 29, 2025, at 11:00 AM, Dick Hardt <dick.hardt at hello.coop> wrote:
>>>
>>> I have no strong views on the scope name. Open to other ideas /
>>> suggestions / opinions!
>>>
>>> Perhaps `cnf` to align with the claim?
>>> ᐧ
>>>
>>> On Fri, Aug 29, 2025 at 3:57 PM <george at practicalidentity.com> wrote:
>>>
>>> Hi,
>>>
>>> Would it make sense to change the scope name identified in the
>>> key-binding spec from something specific like ‘dpop’ to something more
>>> generic? e.g. ‘id_token_kb’ ? Or maybe just make clearer that the RP is
>>> looking for key bound tokens? e.g. ‘dpop_kb’? I just worry that ‘dpop’ by
>>> itself does not communicate the intended behavior.
>>>
>>> Thoughts?
>>>
>>> George Fletcher
>>> Identity Standards Architect
>>> Practical Identity LLC
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> Openid-specs-ab mailing list
>>> Openid-specs-ab at lists.openid.net
>>> https://lists.openid.net/mailman/listinfo/openid-specs-ab
>>>
>>>
>>> _______________________________________________
>>> Openid-specs-ab mailing list
>>> Openid-specs-ab at lists.openid.net
>>> https://lists.openid.net/mailman/listinfo/openid-specs-ab
>>>
>>> _______________________________________________
>>> Openid-specs-ab mailing list
>>> Openid-specs-ab at lists.openid.net
>>> https://lists.openid.net/mailman/listinfo/openid-specs-ab
>>>
>> _______________________________________________
>> Openid-specs-ab mailing list
>> Openid-specs-ab at lists.openid.net
>> https://lists.openid.net/mailman/listinfo/openid-specs-ab
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openid.net/pipermail/openid-specs-ab/attachments/20250830/e2935ca3/attachment-0001.htm>