Re: Security Features offered by Wicket

[Sebastiaan van Erk]

Wicket does nothing to protect from CSRF attacks, and it is trivially vulnerable. Sure it's a lot more difficult with the standard ?wicket:interface type URLs than it would be with more predictable URLs, but you can still quite easily guess the URLs, and futhermore, to improve your chances of success you can simply include many images in the attacking page with different values for the URLs, i.e.,

img src="http://thesiteiwannahack.com/?wicket:interface=:11:formToHack::IFormSubmitListener::&myparam1=val1";

Regards,
Sebastiaan

Nino Saturnino Martinez Vazquez Wael wrote:

and then for page id 11, 12, 13, 14, for 1 to 100 for all I care, all in one page. Furthermore, most people actually LIKE predictable urls and go to great length to mount pages and make them bookmarkable. There's even a StatelessForm component, which is entirely vulnerable to CSRF. Thus, I'd say that even without a quickstart, it's obvious that Wicket does not offer any CSRF protection out of the box, and that if you want this kind of protection, you will have to do it yourself (which is probably not really difficult; though I think many people are not aware of these kind of attack vectors and don't even think about it, which is why it would be nice if Wicket *could* do it out of the box). I believe that answers the original question, that CSRF protection is *NOT* a security feature offered by Wicket.

While that is true.. It's also true that wicket devs favor stuff proven with a quickstart, because it becomes easier to make a fix for something you can see in code..

regards Nino

Ned Collyer wrote:

So as I've written once before a quickstart should be the way to go or just use one of the existing applications, phone book or blog tutorial etc. And make a hack at that...

Nick, I think you would be quite surprised at the level of auditing something has to pass to be used in a financial system, especially a bank. (unless u

have some dodgy bank)

If something is theoretically possible, then thats as good as "proven".

Gotta remember that hackers are a lot smarter in many instances than the
people who wrote the software to keep them out.


Nick Heudecker wrote:

Arthur,

Only what you can *prove* matters, not what you think.  Have you created
an
example application with a CSRF attack?

smime.p7s
Description: S/MIME Cryptographic Signature