[OAUTH-WG] WGLC review of draft-ietf-oauth-security-topics-13

[Pedram Hosseyni <[e-mail]>]

Dear all,

I have a few comments about the leakage of access tokens and the 
underlying assumptions:

Section 2, A5 should be clarified:

"a resource server can be compromised by an attacker": Is the assumption 
that the attacker cannot get access to the resources stored at the 
compromised RS (or only parts of it)? Otherwise, the attacker does not 
need to get an AT anymore. A more differentiated view on RS 
compromisation is already given in Section 4.8.2, but this is not 
reflected in A5.

Also, A5 states that "an access token may be sent to an 
attacker-controlled resource server due to a misconfiguration," which 
does not require the compromisation of an honest RS.

Perhaps one could be more generic here and just say that the AT leaks to 
the attacker. However, a misconfigured RS endpoint not only gets the 
request to this endpoint (containing an AT) but can also respond and 
provide the RO with wrong resources. At the same time, if, say, the RO 
thinks that she is connected to her cloud storage, the attacker would 
get access to all uploaded data.

Is this really what A5 should express, or is the primary focus on the 
leakage of the AT?

Also, for this or the next version of this document, the Cuckoo's Token 
attack (see Section IV-A of http://arxiv.org/abs/1901.11520/ ), should 
be addressed. We also discussed this issue extensively at the last OSW 
in Stuttgart.

Typo: Section 3.5: MTLS -> mTLS

Best regards
Pedram Hosseyni

-- 
Pedram Hosseyni, M.Sc.
Room V38 2.438
Institute of Information Security - SEC
Universität Stuttgart
Universitätsstraße 38
D-70569 Stuttgart
Germany
Phone: [phone]
https://sec.uni-stuttgart.de