Taras,

On Wed, Feb 29, 2012 at 12:47 PM, Taras <ox...@oxdef.info> wrote:
> Andres, please read inline
>
>
>>> Every request inside user's area of web
>>> application with params will be detected as vulnerable. I' have decided
>>> to improve it (with making some research) and want to discuss with you
>>> interesting topic - automation detection of CSRF flaws. My current view
>>> of CSRF-detection process is:
>>>
>>> 1. Does request has cookie (e.g. session cookie)?
>>> 2. Does request has CSRF token in query string of POST payload?
>>> 3. If request has such token does web app check it?
>>> 4. Same with header based protection (Referer/Origin)
>>>
>>> Interesting points are:
>>>
>>> * What can we consider as CSRF token? Currently I calculate entropy of
>>> it (as for password strength) plus some fast string and dictionary checks
>>> * Strict mode on/off - do we need to audit GET requests? Not always...
>>> * How to determine that CSRF token is checked? I am thinking about
>>> checking of response codes and diffs of responses (like blind SQL
>>> injection testing)
>>>
>>> All experiments will be in separate branch and, hope, in nearest future
>>> w3af will detect CSRF flaws much better:)
>>>
>>> Any comments here?
>>
>>
>
>>     Tons of comments , since I've worked in this for Nexpose :) First
>> of all, here's a comment on how AppScan does it [0] , which is
>> basically a "swap the two sessions" technique. Here [1] I wrote on how
>> we did it for Nexpose:
>
>
>>Posted by Andres Riancho in Nexpose on Jan 30, 2012 5:30:59 PM
>
> There is about a month between your post and my letter in case that I didn't
> read it on Nexpose blog! :)
>
>
>> In order to identify the vulnerability, Nexpose is going to follow
>> these main steps:
>> * Crawl the web application without credentials, store non-authenticated
>> forms
>> * Login to the web application with the user provided credentials
>> * Crawl the web application keeping session state and store authenticated
>> forms
>> * Clear session cookies
>> * Login to the web application with the user provided credentials
>> * Crawl the web application keeping session state and store authenticated
>> forms
>> * For each form that can only be accessed after authenticating,
>> analyze if it has an anti-CSRF token. The basic way of doing this is
>> to check if it has a hidden parameter that considerably changed its
>> value (using the two form instances gathered during steps 3 and 6)
>
>> * Report forms that don't have anti-CSRF tokens as vulnerable to CSRF
>
> And you forget to add important part:
>>
>> Another approach includes performing HTTP requests to verify that
>> the anti-CSRF token is properly verified by the application, but extensive
>> testing showed that the accuracy of that heuristic was slightly lower
>> than the one of the implemented algorithm and speed was reduced due
>> to the need of more HTTP requests.

This can be done as an "optional" step for the previously defined
algorithm, something like a "double check" for the CSRF.

I'm not saying that w3af should implement this algorithm, I'll let you
guys decide which one is better !

>
>>     Those steps "sound easy" but there are many details that are not
>> stated there, details that of course I can't disclose. You'll start to
>> figure them out by yourself if you decide to implement an approach
>> similar to [0] or [1].
>
> I totally understand that it is not so easy as somebody may think.
> But it same case it is really interesting to research and implement in w3af.

Sure, it is!

>
>>     Implementing CSRF detection in w3af is somehow difficult since it
>> won't be easy to implement it as an audit plugin because it needs
>> (usually) more than one session, or a lot of information that is not
>> usually available to audit plugins. It looks like we'll need a new
>> scan phase.
>
> I see that Nexpose and AppScan uses 2 phases approach to detect such flaws.
> And it generates 2xtime more requests to web application in discovery
> phrase. And we of course can implement it in same way in w3af. But in same
> time my approach is simpler and doesn't need one more session. I try to make
> more work without sending requests. And I also think that scanner must not
> only try to find CSRF token but also check that web app verify this token it
> before any action. What do you think about way which I have suggested?
>
>
>> [0] http://www-01.ibm.com/support/docview.wss?uid=swg21298693
>> [1]
>> https://community.rapid7.com/community/nexpose/blog/2012/01/30/nexpose-reaches-owasp-top10-coverage
>>
>> Regards,
>>
>>> [0]
>>>
>>> http://w3af.svn.sourceforge.net/viewvc/w3af/trunk/plugins/audit/xsrf.py?view=markup
>>> [1] http://w3af.svn.sourceforge.net/viewvc/w3af/branches/csrf/
>>>
>>> --
>>> Taras
>>> http://oxdef.info
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Keep Your Developer Skills Current with LearnDevNow!
>>> The most comprehensive online learning library for Microsoft developers
>>> is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
>>> Metro Style Apps, more. Free future releases when you subscribe now!
>>> http://p.sf.net/sfu/learndevnow-d2d
>>> _______________________________________________
>>> W3af-develop mailing list
>>> W3af-develop@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/w3af-develop
>>
>>
>>
>>
>
>
> --
> Taras
> http://oxdef.info



-- 
Andrés Riancho
Director of Web Security at Rapid7 LLC
Founder at Bonsai Information Security
Project Leader at w3af

------------------------------------------------------------------------------
Virtualization & Cloud Management Using Capacity Planning
Cloud computing makes use of virtualization - but cloud computing 
also focuses on allowing computing to be delivered as a service.
http://www.accelacomm.com/jaw/sfnl/114/51521223/
_______________________________________________
W3af-develop mailing list
W3af-develop@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/w3af-develop

Reply via email to