RE: Defeating Citi-Bank Virtual Keyboard Protection

[Michal Zalewski <lcamtuf () dione ids pl>]

On Sat, 6 Aug 2005, Debasis Mohanty wrote:

> Read the description section again, perhaps you have missed out the
> following -
> .     The Virtual Keyboard is dynamic
> .     The sequence in which the numbers appears will change every time,
> the page is refreshed
>
> Hence, desiging something the way that you have proposed is not going to
> workout here.

Again, I might be wrong (I am not a Citibank customer), but I understand
that, when you visit the logon page, you're presented with an on-screen
keypad with keys in randomized and possibly constantly changing (dynamic)
order, and must enter your PIN or other authentication data by clicking
appropriate on-screen keys using your mouse.

What I proposed (and I'm sure I'm not innovative here) went along the
lines of hooking up and intercepting the mouse click button, and then, at
the exact moment of mouse click, capturing the position of the mouse
pointer, and a bitmap of its nearest surroundings - ideally, before the
event is delivered to the browser window. That should work regardless of
the method used to shuffle displayed keys, is very much workable on
Windows and under X11, and shouldn't be particularly resource or
bandwidth consuming.

This is a generalised way of snooping virtual keyboards and similar
on-screen mouse-driven input interfaces.

Cheers,
/mz
http://lcamtuf.coredump.cx/
_______________________________________________
Full-Disclosure - We believe in it.
Charter: http://lists.grok.org.uk/full-disclosure-charter.html
Hosted and sponsored by Secunia - http://secunia.com/