I definitely see that there is room for a potential attack depending on the
internal implementation. Wouldn't that be only relevant internally to the
auth server though? Would adding a feature flag to the discovery document
or visibility into configuration help reduce the attack surface? Perhaps, I
would need to see the explicit vulnerability and how it would be
potentially mitigated. From my uninformed perspective it seems like trying
to protect users against a phishing attack that uses a malicious auth
server to emulate an honest one, though I could be missing something
critical.

Any additional information would be appreciated.

- Warren

Warren Parad

Founder, CTO
Secure your user data with IAM authorization as a service. Implement
Authress <https://authress.io/>.


On Mon, Aug 9, 2021 at 10:44 PM Kevat Shah <[email protected]> wrote:

> @[email protected] - Using forgot password and registration verification
> links or codes could be used as vectors of attack.
>
> One possible point of insecurity that I have seen is this: after a user
> verifies their code or link, they are often sent to another page where they
> enter more information (such as additional user information or new
> passwords). The user isn't supposed to be authenticated at this step, and
> yet the application must know who the user is.
>
> How do we confirm the user's identity without authenticating them for
> these steps?
>
> This is the driving factor (the "why") for this ask.
>
> @[email protected] - I agree. Which channels should we use to
> discuss this issue?
>
> - Kevat
>
> On Mon, Aug 9, 2021, 4:17 PM Warren Parad <[email protected]> wrote:
>
>> I think it would be prudent to potentially ask *why?* What problem is
>> necessary to be solved by discussing/standardizing these particular
>> features? There could be, but without understanding, knowing how best to
>> tackle it is a challenging conversation without the right context.
>>
>> Warren Parad
>>
>> Founder, CTO
>> Secure your user data with IAM authorization as a service. Implement
>> Authress <https://authress.io/>.
>>
>>
>> On Mon, Aug 9, 2021 at 10:15 PM Kevat Shah <[email protected]> wrote:
>>
>>> How would that work? Would we need to work with W3C to ensure conformity
>>> of standards?
>>>
>>> On Mon, Aug 9, 2021, 4:11 PM <[email protected]> wrote:
>>>
>>>> Although the IETF has been involved in Best Commercial Practices (BCP)
>>>> (see https://www.ietf.org/rfc/bcp-index.txt )  which I think was the
>>>> subject of Kevat’s original email.
>>>>
>>>>
>>>>
>>>> So perhaps this is a subject matter that could co-exist in both the
>>>> IETF and W3C?
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> *From:* OAuth <[email protected]> *On Behalf Of *Tim Cappalli
>>>> *Sent:* Monday, August 9, 2021 4:06 PM
>>>> *To:* [email protected]
>>>> *Cc:* [email protected]
>>>> *Subject:* Re: [OAUTH-WG] Specifications for Identity Providers
>>>>
>>>>
>>>>
>>>> I don't think there is explicit ownership, but generally password and
>>>> magic link type "stuff" happens in W3C.
>>>>
>>>>
>>>>
>>>> There are existing work efforts around standardizing password reset
>>>> endpoint discovery, password complexity schemas, etc.
>>>> ------------------------------
>>>>
>>>> *From:* Kevat Shah <[email protected]>
>>>> *Sent:* Monday, August 9, 2021 16:03
>>>> *To:* Tim Cappalli <[email protected]>
>>>> *Cc:* [email protected] <[email protected]>
>>>> *Subject:* Re: [OAUTH-WG] Specifications for Identity Providers
>>>>
>>>>
>>>>
>>>> You don't often get email from [email protected]. Learn why this is
>>>> important <http://aka.ms/LearnAboutSenderIdentification>
>>>>
>>>> That's a good point. Is it fair to assume that W3C owns the standards
>>>> for most (if not all) standards related to Identity Providers? Or does it
>>>> make sense for IETF to start setting these standards in cases where W3C
>>>> standards don't exist?
>>>>
>>>>
>>>>
>>>> - Kevat
>>>>
>>>> On Mon, Aug 9, 2021, 2:56 PM Tim Cappalli <[email protected]>
>>>> wrote:
>>>>
>>>> I believe this topic would be more W3C scope, not IETF.
>>>>
>>>>
>>>>
>>>> tim
>>>> ------------------------------
>>>>
>>>> *From:* OAuth <[email protected]> on behalf of Kevat Shah <
>>>> [email protected]>
>>>> *Sent:* Sunday, August 8, 2021 16:37
>>>> *To:* [email protected] <[email protected]>
>>>> *Subject:* [OAUTH-WG] Specifications for Identity Providers
>>>>
>>>>
>>>>
>>>> Some people who received this message don't often get email from
>>>> [email protected]. Learn why this is important
>>>> <http://aka.ms/LearnAboutSenderIdentification>
>>>>
>>>> I propose that we expand upon specific functionality provided by
>>>> Identity Providers (IdPs) and tasks handled by them.
>>>>
>>>>
>>>>
>>>> To start with, there should be clear specifications for various
>>>> functionalities that IdPs provide such as:
>>>>
>>>>
>>>>
>>>> - Email verification on registration
>>>>
>>>> - Specifications regarding "forgot password" functionality
>>>>
>>>> - Specifications regarding "resest password" functionality for users
>>>> that are logged in
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> These specifications only pertain to Identity Providers, and allow an
>>>> industry-wide set of rules that each Identity Provider must follow. The
>>>> purpose of doing so would be to standardize various frequently used and
>>>> implemented flows that are secure and widely reusable.
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> Some problems that would be addressed by these specifications would be:
>>>>
>>>>
>>>>
>>>> - How to securely implement functionality where a user is sent a link
>>>> to verify their email address
>>>>
>>>>
>>>>
>>>> - How to securely implement functionality where a user is sent a
>>>> verification code to verify their email address
>>>>
>>>>
>>>>
>>>> - How to securely implement functionality where a user is sent a link
>>>> to reset their password
>>>>
>>>>
>>>>
>>>> - How to securely implement functionality where a user is sent a
>>>> verification code to reset their password
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>> OAuth mailing list
>>> [email protected]
>>> https://www.ietf.org/mailman/listinfo/oauth
>>>
>>
_______________________________________________
OAuth mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/oauth

Reply via email to