v3 consent to attribute release : model terms-of-use as attribute ?

[Tom Zeller]

>> If so, I'm wondering how to handle that attribute specially
>>during attribute resolution so it is available to the consent flow but
>>not actually released.
>
> As an independent question, the answer to that is that you could do a
> couple of things. One is that you could include it during standard
> resolution, but just not attach any SAML encoders to it, which means it
> doesn't matter whether it's formally "released" or not, it would never
> show up outside.

Omitting the encoders from the attribute definition is what I started
doing, but wasn't sure if that was a good idea.

> Alternatively, it's a simple matter to run a dedicated attribute
> resolution step for a specific attribute (I've done that in some of the
> flows) and just ignore filtering. Basically you end up with an
> AttributeContext in some dedicated spot that isn't used to actually
> generate attributes in the assertion.

Like my response above, I figured it would be okay to resolve a single
attribute in the TOU flow, but again wasn't sure if that would be
confusing or conflating down-the-road.

Thanks.