W3C home > Mailing lists > Public > public-web-security@w3.org > June 2011

Re: Req for feedback? Add attribute to elements to defeat clickjacking

From: Giorgio Maone <g.maone@informaction.com>
Date: Tue, 07 Jun 2011 18:31:59 +0200
Message-ID: <4DEE527F.5050102@informaction.com>
To: sird@rckc.at
CC: public-web-security@w3.org
> Let's say we add an attribute (or CSS property) that will instruct the
> browser that when hovered, they should appear over everything that
> might be over them, and make it's opacity opaque.

What about smart phones and tablets ("tap" devices) where hovering never
happens? Not exactly a rare use case these days...

-- G

Eduardo Vela wrote, On 07/06/2011 18.20:
> Hi!
> 
> One of the most complicated problems we have now a days is that there
> are some widgets which are specifically designed to be iframed (ads,
> games, like buttons, etc..) are difficult to protect against
> clickjacking.
> 
> So, there's this idea that might solve some of those problems.
> 
> Let's say we add an attribute (or CSS property) that will instruct the
> browser that when hovered, they should appear over everything that
> might be over them, and make it's opacity opaque.
> 
> So something like:
> 
> <style>
> #buyButton:hover{
>     visibility: forced;/* or something else, I don't know.. */
> }
> </style>
> <button id="buyButton">Click here to purchase server for $500.00.</button>
> 
> 
> Will make the buy button visible no matter if there are things over it
> (in other windows).
> 
> There are some restrictions that should be made for this to be effective:
> 1. This property should only work on hover, or active elements (eg.
> with the mouse over, or selected by <tab>).
> 2. If this is inside an iframe, then it shouldn't obscure anything
> outside of the frame window (so, if there's a 1x1 frame, the button
> will still be invisible).
> 3. If part of the element is outside of the window (eg.
> top:-200;left:-200), it shouldn't obscure chrome UI (like the
> addressbar or so).
> 
> To mitigate against attacks abusing (2) or (3) if the "forced
> visibility" fails, then mouse click events to that object could be
> disabled, but even if nothing is done it would be ok, since the frame
> can detect it's size and position (so it would know that the click
> might be invalid).
> 
> I was wondering if you can find any other (abuse) cases for this
> feature, or if you have any thoughts. Or if there might be other
> (better) solutions for this problem.
> 
> Greetings!!
> 
> -- Eduardo
> 
Received on Tuesday, 7 June 2011 16:34:35 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:26:19 UTC