- From: Marcos Caceres <marcosscaceres@gmail.com>
- Date: Tue, 4 Dec 2007 14:17:01 +1000
- To: public-appformats@w3.org
I guess one thing we don't need to worry about at the moment is concerning ourselves with the widget.system() API, as we currently don't spec it ( should we?:) ).... And I'm not sure what we can do with regards to eval() as I gather that is a problem for the web at large.... Kind regards, Marcos On Dec 4, 2007 9:07 AM, Thomas Roessler <tlr@w3.org> wrote: > > FYI, an incidental observation of security holes introduced by some > real-life widgets. I was particularly amazed to see that the two > available free widgets for that particular functionality were making > the same (rather severe) mistake. > > Regards, > -- > Thomas Roessler, W3C <tlr@w3.org> > > > > > > ----- Forwarded message from Thomas Roessler <tlr@w3.org> ----- > > From: Thomas Roessler <tlr@w3.org> > To: bugtraq@securityfocus.com > Cc: brett@webfroot.co.nz, mail@ben-ward.co.uk > Date: Tue, 4 Dec 2007 00:04:57 +0100 > Subject: [MacOS X] Insecure eval() in Twitgit and Twitterlex dashboard > widgets > > Twitgit [1] and Twitterlex [2] are two MacOS X Dashboard widgets > (developed in JavaScript) that can be used to display twitter.com > updates. > > Both regularly retrieve data using the Twitter JSON API and parse > whatever is returned with eval(). Both relax the dashboard's > JavaScript sandbox to enable the widget.system() call, which indeed > amounts to the equivalent of system(3); i.e., if an attacker can > take over the widget, the attacker can take over the user's account > (and, quite often, the system). > > The data are retrieved through plain HTTP. Therefore, these widgets > are vulnerable to at least: > > - cross-site-scripting attacks through Twitter > - subversion of Twitter and, in the case of Twitterlex, also > subversion of a server used for update notifications > - man-in-the-middle attacks against local networks > > (Also, deliberately malicious behavior by either Twitter or the > author of at least Twitterlex is a risk from a security perspective; > if one was to assume malice, then Twitterlex could be classified as > a nifty backdoor.) > > What makes this case particularly interesting is that this is a case > in which -- along with the development platform, JavaScrit -- the > borders between Web and local vulnerabilities get increasingly blurry. > > It would probably be an interesting exercise to go through some more > dashboard widgets and grep for eval. I'd bet quite a bit that > there's much more out there. > > 1. http://inner.geek.nz/projects/twitterlex/ > 2. http://ben-ward.co.uk/widgets/twitgit/ > > Regards, > -- > Thomas Roessler, W3C <tlr@w3.org> > > > > ----- End forwarded message ----- > > -- Marcos Caceres http://datadriven.com.au
Received on Tuesday, 4 December 2007 04:17:16 UTC