Thank you very much for some direction. (And thank you for the Cookbook too.)
I'm trying to move on to the next level. You are way out there in front, I'm
trying to catch up. <g>
I'm working on chapter 16 in the Cookbook (injection). I'm new to these structures
and how to 'read' them. Once I get how all of this comes together, things will be
clearer.
Thanks again for you help; and also to vol-users, this is the only contact I have with
folks using Volatility or have a clue with what I am talking about. <g>
Mike Lambert
Date: Wed, 29 Feb 2012 08:29:39 -0500
Subject: Re: [Vol-users] A little direction help please
From: michael.hale(a)gmail.com
To: dragonforen(a)hotmail.com
CC: vol-users(a)volatilityfoundation.org
Mike,
On Sun, Feb 26, 2012 at 7:28 PM, Mike Lambert <dragonforen(a)hotmail.com> wrote:
I am conducting an examination of a test system
infected with a sample from
the internet. See attached virustotal scan if you are interested what it is.
fawi.exe is the dropped infection result of sample ssl.exe
I am having trouble answering the question, "Where is the malware?". This is
not the usual rundll or infected program in memory, it is injected. I think
it is in the kernel, but this is a new area for me. Any direction on how I
answer that question is appreciated. (One thought I had is: that the
malware is hooked into the SSDT that zone alarm's vsdatant.sys hooks and
since vsmon.exe probably uses vsdatant.sys, that is why the malware looks
like it is in vsmon.exe - but it is just hooked into the kernel.)
Don't overlook your biggest clues. The virustotal scan indicates PWS,
Zbot, Injector, etc. Its probably a zeus variant. The IP that
Explorer.exe is connected to is a known zeus CC:
https://zeustracker.abuse.ch/monitor.php?host=analyticdns.org.
Zeus doesn't use any kernel hooks, but there are at least 3 or 4
analysis by different folks showing how they detected zeus (google for
"zeus volatility malware" or just run through the malware commands
http://code.google.com/p/volatility/wiki/CommandReference#Malware_and_Rootk…
with a Volatility 2.0 build).
Here is where I am at and how I got there:
The infection started executing ssl.exe, here is some prefetch info:
SSL.EXE-028F0541.pf Saturday, February 25, 2012 23:08:41 Z (UTC)
Nothing in PSSCAN, no ssl.exe anywhere in fact.
It dropped:
\DOCUMENTS AND SETTINGS\MIKE\APPLICATION DATA\YPYCUB\FAWI.EXE and executed
it:
FAWI.EXE-398259A4.pf Saturday, February 25, 2012 23:08:41 Z (UTC)
PSSCAN shows it ran for 3 secs
Offset Name PID PPID PDB Time
created Time exited
0x0420ca70 fawi.exe 3812 2468 0x0cd44000 2012-02-25
23:08:37 2012-02-25 23:08:40
Seeing this:
fawi.exe pid: 3812
Unable to read PEB for task.
did not surprise me since it had exited.
I could easily see how it is in NTUSER.DAT and starts with the run key:
HKCU\Software\Microsoft\Windows\CurrentVersion\Run
{AED24BF3-7C21-B878-2172-0CFD5C3474CA}
"C:\Documents and Settings\Mike\Application Data\Ypycub\fawi.exe"
Glows Anger Hebrew
Orb Networks
8.3.0.0
c:\documents and settings\mike\application data\ypycub\fawi.exe
f4cc71ca8e5522519b9c6e8d350d876b (MD5)
Since fawi.exe has terminated, my only clue is explorer.exe communications
seen by cports.exe
2/25/2012 5:08:39 PM Added Explorer.EXE TCP
192.168.1.44:1185 94.63.147.98:80
(An interesting note: Win32dd and FDPRO were both unable to collect a memory
image. I hibernated the system and converted hiberfil.sys to a file I could
analyze with Volatility.)
Explorer.exe seemed the first suspect as it was the only process sending
network i/o out
0x043d7360 explorer.exe 1756 1732 0x107ff000 2012-02-25 11:30:26
I pulled the dlls, files and reg keys for PID 1756, along with a memdmp.
Nothing interesting in the dlls files or keys. I examined the dmp and found
an unusual string in it along with ssl.exe and fawi.exe and this
"Glows Anger Hebrew" A search of the internet shows one reference in a
similiar infection in February. I found it on the internet just 2 days after
the first sample was sent to virustotal.
I searched the 458 MB memory image for "Glows Anger Hebrew" and got 5 hits.
Here is my text file for the strings function
357229672:Glows
280642408:Glows
257105340:Glows
113457472:Glows
357230696:Glows
Results from strings is:
357230696 [kernel:df864868 ] Glows
357229672 [kernel:df864468 ] Glows
257105340 [kernel:e1ec1dbc ] Glows
113457472 [1456:2ac0940 ] Glows
280642408 [1456:45b8368 ] Glows
PID 1456 is:
vsmon.exe pid: 1456
Command line : C:\WINDOWS\system32\ZoneLabs\vsmon.exe -service
PSSCAN is
0x043e6020 vsmon.exe 1456 848 0x102da000 2012-02-25 11:30:25
I dumped PID 1456 and searched the dmp. I also grabbed a couple of other
processes (because 3 were kernel references). See search results in the
attached JPG (cut and paste test did not work out well).
So, strings said that the malware is injected in vsmon.exe. Is that correct?
Now I guess my next step is to identify the thread? Or is this malware
'everywhere' in that it is in the kernel, and everyone mapps to it?
I'd view page permissions on 2ac0xxx and 45b8xxx in pid 1456 (vadinfo
command) and see if they're executable. If they're read-only they're
probably in vsmon.exe because vsmon.exe, being part of ZoneAlarm,
opened the malicious file to scan it.
MHL
> I need some help on next steps. Am I moving in the right direction?
>
> BTW, who else is doing this type of work? and are you interested in
> colaborating? I am in Houston, TX.
>
> Thanks for your help!
>
> Mike Lambert
> dragonforen(a)hotmail.com
> michael-lambert.us/index.htm
>
>
>
>
> _______________________________________________
> Vol-users mailing list
> Vol-users(a)volatilityfoundation.org
>
http://lists.volatilityfoundation.org/mailman/listinfo/vol-users
>