Question regarding CVE applicability of missing HttpOnly flag

["Vincent Danen" <vdanen () redhat com>]

I have a question for the list (and MITRE I guess) regarding a few CVEs I've seen regarding "missing HttpOnly flag in 
the Set-Cookie header".  I just had a discussion with someone and we had differing points of view so I thought it would 
be interesting to see what others thought.

Given that a few CVEs have been assigned for this (see http://cve.mitre.org/cgi-bin/cvekey.cgi?keyword=httponly 
although not all of them are applicable to the scenario as some are not doing correct things when the flag is set).

My point of view is that lack of HttpOnly is not a security flaw, but would fall under the category of "hardening".  My 
underlying thought is that HttpOnly is a good protection mechanism for when a server/web application has an XSS flaw 
that could be already exploited.  Even the MITRE descriptions say "makes it easier for remote attackers" (which doesn't 
sound like a flaw).  If it were "makes it _possible_ for remote attackers" I wouldn't have a problem with it at all.

Setting the flag is a good thing to do, just like running SELinux (or AppArmor), running a virus scanner, and having a 
firewall yet we don't necessarily have CVEs for the absence of those things.

I'm wondering what the justification is for calling this a flaw, when there are other similar scenarios that we do not.


-- 
Vincent Danen / Red Hat Product Security

Attachment: signature.asc
Description: OpenPGP digital signature