RE: Spi's products worth a try? Or any suggestions for developers' tool?

["Peine,Holger" <Holger.Peine () iese fraunhofer de>]

> -----Original Message-----
> From: App Master [mailto:appmasterzero () hotmail com] 
> Sent: Montag, 7. November 2005 22:05
> To: araheja () techquotes com
> Cc: webappsec () securityfocus com
> Subject: Re: Spi's products worth a try? Or any suggestions 
> for developers' tool?
> [...]
>  You see, a lot of security 
> products are just 
> like machine guns that fire strings at an application and 
> then grep the HTML 
> for another response string.  This is the reason that after 
> you run them it 
> takes so long to verify if the results are correct or not, 
> because its 
> mostly pure signature matching -- stateless -- of raw HTML and server 
> response codes, without any visibility as to what is occuring 
> in the browser 
> (at the application level), or if the application is causally 
> or statefully 
> affected by injected values.
>
> Hailstorm does it differently, using what you might think of 
> as active 
> payloads. It monitors what each injected payload does and 
> then monitors 
> browser memory (it uses a baked-in version of Mozilla) to 
> trap when code or 
> events execute in the application space as a result of its 
> actions. This is 
> a world of difference between other black-box tools. 

I'm not really convinced (yet) by this argument. While I 
generally agree that there should be room for improvement
in security analysis by paying more attention to the application
state, I don't see how the above statements support this. 
I see only a weak connection between the general statement 
about observing state and the second statement about observing
browser behavior instead of HTTP traffic, and I don't see which
observations could be derived from browser behavior that could
not equally be derived from the HTTP data (after all, a browser's
behavior is determined by its input data, leaving aside some
vendor-specifc idiosyncrasies which are on topic here). 
For example, I can decide from parsing the HTML whether a
certain XSS-Javascript would be executed or not; what's
the added value of monitoring the Javascript interpreter
in the browser?

So, while I have a gut feeling that there is an interesting
point hidden in your argument, could you please elaborate a
bit (including an example) to bring out that point?

Regards,
Holger Peine

-- 
Dr. Holger Peine, Security and Safety
Fraunhofer IESE, Fraunhofer-Platz 1, 67663 Kaiserslautern, Germany
Phone +49-631-6800-2134, Fax -1299 (shared)
www.iese.fraunhofer.de/Staff/peine -- PGP key on request or via
http://pgp.mit.edu