[Oisf-users] Suricata Cookie Header Questions

[Anoop Saldanha]

On Fri, Apr 15, 2016 at 3:08 AM, Duane Howard <duane.security at gmail.com> wrote:
> Also, somewhat related:
> Are these parsed irrespective of direction? For example, if a bonkers server
> sent "Cookie:" or a bad client set "Set-Cookie:" would those both still map
> to the cookie buffer?
>

Correct.

> On Thu, Apr 14, 2016 at 2:36 PM, Duane Howard <duane.security at gmail.com>
> wrote:
>>
>> Thanks Victor,
>>
>> The example response on the wiki page doesn't make it clear that
>> 'Set-Cookie:' also applies here. In fact it includes 'Set-Cookie:' in the
>> http_header, http_raw_header definition.
>>
>> On Thu, Apr 14, 2016 at 1:53 PM, Victor Julien <lists at inliniac.net> wrote:
>>>
>>> On 14-04-16 22:16, Duane Howard wrote:
>>> > According
>>> > to:
>>> > https://redmine.openinfosecfoundation.org/projects/suricata/wiki/HTTP-keywords
>>> > """Although cookies are sent in an HTTP header, you can not match on
>>> > them with the http_header keyword. Cookies are matched with their own
>>> > keyword, namely http_cookie."""
>>> >
>>> > I had a couple of questions about this:
>>> >
>>> >  1. Is this documentation still accurate?
>>>
>>> Yes.
>>>
>>> Note that it also applies to Set-Cookie.
>>>
>>> >  2. Does this mean that Suricata will effectively strip the 'Cookie:
>>> >     foo=120398' out from the http_header buffer? Or will it still
>>> >     contain 'Cookie:' for example? [0]
>>>
>>> It is stripped (or actually not included in reconstruction):
>>>
>>> https://github.com/inliniac/suricata/blob/master/src/detect-engine-hhd.c#L165
>>>
>>> >  3. If above is true, is the correct way to perform relative matches of
>>> >     'Cookie: foo=12398' to other header fields to use a plain content
>>> >     match, and not the http_header buffer?
>>>
>>> Use http_raw_header. It does include the cookie headers.
>>>
>>> >
>>> > If my understanding of this is correct I think there's at least one ET
>>> > rules that are probably not working as intended, since I see content
>>> > negations of 'Cookie|3A|'; http_header; as well as Cookie value checks
>>> > against http_header. Example rule from ET Open[1]
>>> >
>>> > [0] Does this:
>>> > GET /foo.js HTTP/1.1
>>> > Host: no.evil.com <http://no.evil.com>
>>> > Cookie: UID=17220493a47a2391519d61c144a531063; UIDR=1649534063; CP3=4
>>> > Connection: keep-alive
>>> > Cache-Control: max-age=0
>>> > Accept: */*
>>> > User-Agent: Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML,
>>> > like Gecko) Chrome/49.0.2623.110 Safari/537.36
>>> > Referer: http://www.referme.com
>>> > Accept-Encoding: gzip, deflate, sdch
>>> > Accept-Language: en-US,en;q=0.8
>>> >
>>> > yield this for http_header?
>>> > Host: no.evil.com <http://no.evil.com>
>>> > Connection: keep-alive
>>> > Cache-Control: max-age=0
>>> > Accept: */*
>>> > User-Agent: Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML,
>>> > like Gecko) Chrome/49.0.2623.110 Safari/537.36
>>> > Referer: http://www.referme.com
>>> > Accept-Encoding: gzip, deflate, sdch
>>> > Accept-Language: en-US,en;q=0.8
>>> >
>>> > Example of ET Rule that's probably broken:
>>> > [1] alert http $EXTERNAL_NET any -> $HOME_NET any (msg:"ET
>>> > CURRENT_EVENTS DRIVEBY EgyPack Exploit Kit Cookie Set";
>>> > flow:established,from_server; *content:"Cookie|3a| visited=TRUE";
>>> > http_header;* *content:"Cookie|3a| mutex="; http_header;*
>>> > reference:url,www.kahusecurity.com/2011/new-exploit-kit-egypack/
>>> > <http://www.kahusecurity.com/2011/new-exploit-kit-egypack/>;
>>> >
>>> > reference:url,www.vbulletin.com/forum/forum/vbulletin-3-8/vbulletin-3-8-questions-problems-and-troubleshooting/346989-vbulletin-footer-sql-injection-hack
>>> >
>>> > <http://www.vbulletin.com/forum/forum/vbulletin-3-8/vbulletin-3-8-questions-problems-and-troubleshooting/346989-vbulletin-footer-sql-injection-hack>;
>>> >
>>> > reference:url,blog.webroot.com/2013/03/29/a-peek-inside-the-egypack-web-malware-exploitation-kit/
>>> >
>>> > <http://blog.webroot.com/2013/03/29/a-peek-inside-the-egypack-web-malware-exploitation-kit/>;
>>> > classtype:bad-unknown; sid:2014407; rev:3;)
>>> >
>>>
>>> Looks like this is indeed broken.
>>>



-- 
-------------------------------
Anoop Saldanha
http://www.poona.me
-------------------------------