[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20041009095815.GA2819@sivokote.iziade.m$>
From: guninski at guninski.com (Georgi Guninski)
Subject: Re: Yet another IE aperture
i didn't notice you have disclosed this (or a very similar to it bug).
besides me more than 5 people tested variations of the testcase and it worked
for all of them.
can you comment on this testcases:
http://www.guninski.com/where_do_you_want_billg_to_go_today_1_demo2.html
http://www.guninski.com/where_do_you_want_billg_to_go_today_1_demo.html
redirect1.pl is hosted on apache and is:
-----------------------
#!/usr/bin/perl
print "Location: http://georgi.df.ru/xml2.xml\r\n\r\n";
-----------------------
note: if the xml is not well formed, parseError returns at least one line of
it, not to mention other exploit scenarios.
--
georgi
On Sat, Oct 09, 2004 at 03:28:25AM +0200, GreyMagic Security wrote:
> >Georgi Guninski security advisory #71, 2004
> >http://www.guninski.com/where_do_you_want_billg_to_go_today_1.html
>
> .. snip ..
>
> >By opening html in IE it is possible to read at least well formed xml from
> >arbitrary servers. The info then may be transmitted.
>
> GreyMagic disclosed the EXACT same issue on August 2002, over two years ago.
> Microsoft, at the time, took over 6 months to resolve the issue (initially
> reported to them on Feb 2002) and eventually released a patch (MS02-047).
>
> See http://www.greymagic.com/security/advisories/gm009-ie/ for more details
> and a live PoC (it also shows a neat method to get partial content from
> documents that aren't well-formed xml).
>
> That said, all our tests of this issue currently throw an "Access denied"
> exception, as they properly should. However, these tests are performed in
> the Internet Zone. Your tests might have been performed in another zone that
> had "Access data sources across domains" set to "Enabled," which would
> enable this vulnerability by design.
Powered by blists - more mailing lists