[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <mpro.maavbm0atu7hc02hi.taviso@cmpxchg8b.com>
Date: Thu, 13 Sep 2012 19:59:46 +0200
From: Tavis Ormandy <taviso@...xchg8b.com>
To: oss-security@...ts.openwall.com
Subject: Re: note on gnome shell extensions
Vincent Danen <vdanen@...hat.com> wrote:
> * [2012-09-13 18:03:33 +0200] Marcus Meissner wrote:
>
> > On Thu, Sep 13, 2012 at 05:39:57PM +0200, Tavis Ormandy wrote:
> > > On Mon, Sep 10, 2012 at 02:48:38PM -0600, Vincent Danen wrote:
> > > > * [2012-09-08 18:14:10 -0600] Kurt Seifried wrote: SUSE has some
> > > > interesting info in their bug:
> >> >
> > > > https://bugzilla.novell.com/show_bug.cgi?id=779473#c4
> >> >
> > > > By the sounds of it, this should be harmless. Vincent Untz says
> > > > that the browser plugin doesn't actually install the extensions,
> > > > it's passed to another process via a dbus call to gnome-shell, which
> > > > sends the uuid of the extension to the extensions.gnome.org web site
> > > > in order to download the extension.
> >> >
> > > > See:
> >> >
> > > > http://git.gnome.org/browse/gnome-shell/tree/js/ui/shellDBus.js#n305
> > > >
http://git.gnome.org/browse/gnome-shell/tree/js/ui/extensionDownloader.js#n27
> >> >
> > > > which is:
> >> >
> > > > let message = Soup.form_request_new_from_hash('GET',
> > > > REPOSITORY_URL_INFO, params);
> >> >
> > > > And REPOSITORY_URL_INFO is hardcoded earlier:
> >> >
> > > > const REPOSITORY_URL_BASE = 'https://extensions.gnome.org'; const
> > > > REPOSITORY_URL_DOWNLOAD = REPOSITORY_URL_BASE +
> > > > '/download-extension/%s.shell-extension.zip'; const
> > > > REPOSITORY_URL_INFO = REPOSITORY_URL_BASE + '/extension-info/';
> > > > const REPOSITORY_URL_UPDATE = REPOSITORY_URL_BASE +
> > > > '/update-info/';
> >> >
> > > > I don't think this is something that can be exploited, based on the
> > > > above.
> >>
> > > Not sure I follow the logic, can't I just upload something malicious
> > > to extensions.gnome.org and then force you to download it? I mean, I
> > > can try it if you're not convinced it's possible.
> >
> > There are supposed to be reviewers before it gets activated, but exactly
> > this concern Sebastian also voiced.
> >
> > > They surely do not have a magical technique for determining if my code
> > > is or can become malicious.
> >
> > Exactly.
>
> Yeah, this is definitely a possibility, but could happen regardless of
> this with some social engineering (hey, download my cool foo extension!)
> and have something malicious up there. This is pretty much the same
> thing, just making it easier.
Well, no. This is like saying it's pointless to patch vulnerabilities,
because I can just make you download malware. You can't just make me
download malware, because I know how to make trust decisions.
You could make me download a malicious gnome extension, because you can do
so without interaction or my consent.
> It's not much different than having a malicious app in the
> iTunes/Android/Whatever app store. The flaw there isn't so much in the
> app store, but the app. Wouldn't the same thought apply here?
>
I've uploaded my malicious android app, how do I make you install it?
I can create http://foo.com/malware.rpm, that's clearly not a vulnerability
and working as designed. But if I can force you to download and install it
without you having the opportunity to make a trust decision, that clearly is
a vulnerability.
Do you agree that I can upload something malicious to extensions.gnome.org?
Do you agree that I can make you install it without consent, interaction, or
the opportunity to make a trust decision?
If so, then I don't understand the objection :-)
Tavis.
--
-------------------------------------
taviso@...xchg8b.com | pgp encrypted mail preferred
-------------------------------------------------------
Powered by blists - more mailing lists
Please check out the
Open Source Software Security Wiki, which is counterpart to this
mailing list.
Confused about mailing lists and their use?
Read about mailing lists on Wikipedia
and check out these
guidelines on proper formatting of your messages.
