Full Disclosure mailing list archives

RE: Microsoft MCIWNDX.OCX ActiveX buffer overflow


From: "Drew Copley" <dcopley () eeye com>
Date: Fri, 15 Aug 2003 11:27:41 -0700



-----Original Message-----
From: Tri Huynh trihuynh () zeeup com 
Subject: [Full-disclosure] Microsoft MCIWNDX.OCX ActiveX 
buffer overflow 


Hi, List

I'm very happy with all the supportive feedbacks. The 
MCIWNDX.OCX is originally shipped with Visual Studio 5.0 (or 
VB 5.0) and it is a Microsoft-signed ActiveX 
(http://support.microsoft.com/default.aspx?scid=http://support
.microsoft
.com
:80/support/kb/articles/Q173/3/52.asp&NoWebContent=1)
However, while most of the ActiveXs in Visual Studio 5.0 are 
updated and patched in VS 6.0, MCIWNDX.OCX is not patched, 
and a new version of the ActiveX called MCI32.ocx is 
introduced. Unfortunately,  MCIWNDX.OCX is still shipped with 
Visual Studio 6.0 CD and it is placed  in 
"\Common\Tools\Vb\Unsupprt\Mciwindx" folder of the Visual 
Studio 6.0 Enterprise Edition Disk 3; and the ActiveX is also 
installed by default in the Enterprise Edition. Since it is 
installed by default and registered with a CLSID, it is a 
timing bomb and should be removed.

As most of us already notice, an unsecure local ActiveX can 
be exploited by making a website that tries to inject 
shellcode to take control of the client machine. Although the 
ActiveX is patched locally, the hacker can still use the 
CODEBASE/CLSID properties to instruct the browser to download 
a Vulnerable Microsoft-signed ActiveX and exploit it.

I haven't seen any widespread of ActiveX attacks conducted by 
hackers yet, but however, since the way IE handling ActiveX 
in default is unsecure, we will probably see many hacking 
attemps using this weakness.

Regards,

Tri Huynh
SentryUnion


PS : I just recognize that several places in my recent post I 
mistakenly wrote the ActiveX name MCWNDX.OCX instead of 
MCIWNDX.OCX. :-) Sorry for all the confusing. Now, I figure 
out that coffee and girls do destroy human mnemonic :-)

Ah, yes... Looking back on it, I see it now. That is what I get for
looking at the copy of the report. MCIWNDX.OCX was mentioned once, at
the first, and this was cut out of subsequent replies apparently by
mistake.

The issue, btw, sounds like a heap based overflow, as opposed to a stack
based overflow. Generally, these issues can be exploitable, but it can
tend to be very difficult to exploit them universally. 

I hope that Microsoft sees this. Visual Studio 6 is still used by an
enormous amount of people, and I am not sure if uninstalling it will
even remove the activex... And regardless, as Thor pointed out and as
Guninski originally noted, these activex are signed. You will have to
click "Okay" still to install them, however, though it will say,
'Approved by Microsoft'.

Unfortunately, Microsoft may not see this post, and they seem to have
ignored the original post you made to security@microsoft. When they do
not reply, that generally means they could not find the bug, they could
not duplicate it, and they do not know you. Rather rude and
unprofessional. Maybe they can fix this problem.

If they do not see this post, then there will be this effective zero day
out there. 95% of the world uses Internet Explorer according to the
latest stats. With around 400 million internet users and these being -
generally - the wealthiest of the world... This is quite an open door.

Now, back to my fantasies about being a hitman instead of a security
researcher. ;)









----- Original Message ----- 
From: "Drew Copley" <dcopley () eeye com>
To: <jasonc () science org>; "'Thor Larholm'" <thor () pivx com>; 
"'Tri Huynh'" <trihuynh () zeeup com>; <bugtraq () securityfocus com>
Cc: <full-disclosure () lists netsys com>
Sent: Wednesday, August 13, 2003 3:48 PM
Subject: RE: [Full-disclosure] Microsoft MCWNDX.OCX ActiveX 
buffer overflow




-----Original Message-----
From: Jason Coombs [mailto:jasonc () science org]
Sent: Wednesday, August 13, 2003 12:36 PM
To: Thor Larholm; Tri Huynh; bugtraq () securityfocus com
Subject: RE: [Full-disclosure] Microsoft MCWNDX.OCX 
ActiveX buffer 
overflow


What about pointing the OBJECT tag codebase to a known, 
or probable, 
location on the victim's own hard drive?

It apparently is not on people's systems, is the point. If it is not
the
multimedia control and there is such an activex, then thor 
is correct, 
and it can simply be pointed at remotely.


ActiveX never implemented any type of "same origin 
policy" the way 
JavaScript does, so a local codebase reference should work as a 
technique to silently activate any Microsoft-signed 
ActiveX control.

Partly true, though I can't run files using activex on your system 
locally, there are various checks now in place.


But I could be mistaken, this is commentary from memory not 
experimental result.




I'd much rather spend my time conducting security audits of Linux 
and trying to help those companies threatened by SCO's copyright 
claims defend themselves in court.


I would rather be home, watching television, or playing a 
video game. 
Actually, it would be nice to be surfing now. From a purely
fantastical
viewpoint, I suppose bounty hunting would be a bit funner, 
or perhaps 
being a professional hitman.

Now, back to complete seriousness.

Jason Coombs
jasonc () science org

-----Original Message-----
From: full-disclosure-admin () lists netsys com
[mailto:full-disclosure-admin () lists netsys com]On Behalf Of Thor 
Larholm
Sent: Wednesday, August 13, 2003 8:22 AM
To: Tri Huynh; bugtraq () securityfocus com
Cc: full-disclosure () lists netsys com
Subject: Re: [Full-disclosure] Microsoft MCWNDX.OCX 
ActiveX buffer 
overflow


The MCWNDX.OCX binary is digitally signed by Microsoft, 
and as such 
you can plant it on the users machine just by pointing 
the codebase 
attribute of your OBJECT tag to an archived copy of the 
file on your 
own server.

This also applies to other outdated ActiveX controls, even when a 
newer
(patched)  version exists and is installed on the users 
machine you 
can still re-introduce the old, buggy version since it is 
digitally 
signed by Microsoft.


Regards
Thor Larholm
PivX Solutions, LLC - Senior Security Researcher




_______________________________________________
Full-Disclosure - We believe in it.
Charter: http://lists.netsys.com/full-disclosure-charter.html







_______________________________________________
Full-Disclosure - We believe in it.
Charter: http://lists.netsys.com/full-disclosure-charter.html


Current thread: