add verifier public key to the session transcript

[awoie]

Depends on #448 (on the removal of client_id from SessionTranscript).

This PR includes the thumbprint of the public verification key of the verifier in the SessionTranscript.

Fixes #400

[c2bo]

Type should either be bstr if we use the hash directly, or we need to clarify that the hash is base6url encoded

[awoie]

Type should either be bstr if we use the hash directly, or we need to clarify that the hash is base6url encoded

changed to bstr.

[paulbastian]

I think this is a substantial important changes, that needs further elaboration on the motivation:

• add a note under this section that this is a security measurement for unsigned requests
• please add security consideration that the RP should check that jwk_thumbprint matches its key and what to do when it does not match

[awoie]

I think this is a substantial important changes, that needs further elaboration on the motivation:

• add a note under this section that this is a security measurement for unsigned requests

I can add a note regarding the purpose of the same section since it only applies to mdocs.

• please add security consideration that the RP should check that jwk_thumbprint matches its key and what to do when it does not match

The RP does not have to check that the jwk_thumbprint has to match since the SessionTranscript is computed by both parties, verifier and wallet, individually. If the Verifier uses a different key than the wallet, the deviceSigned won't verify since the signature/mac is invalid.

[awoie]

@paulbastian please review again. I added the purpose of this to the respective section.

[paulbastian]

I believe a dedicated section in the security considerations to give guidance for Verifier in these cases would be appropriate, but that shouldn't block this from being merged.

[Sakurann]

I believe a dedicated section in the security considerations to give guidance for Verifier in these cases would be appropriate, but that shouldn't block this from being merged.

I agree.. we should be more diligent in adding security considerations when we do PRs that apply cc @danielfett

[awoie]

I believe a dedicated section in the security considerations to give guidance for Verifier in these cases would be appropriate, but that shouldn't block this from being merged.

I agree.. we should be more diligent in adding security considerations when we do PRs that apply cc @danielfett

We have already a note on that in the PR as @paulbastian requested. IMO, a security considerations section would only make sense if it applies to all credential formats but we decided to do this for mdocs only to begin with. If @paulbastian could unblock since we added a note?