BPMTK
Didier Stevens
<!-- end header -->
Quickpost info
Quickpost info
Monday 2 February 2009
CommNet at TechEd Barcedlona 2008
It was surprising to see the CommNet desktops at our disposal at TechEd Barcelona 2008. This time, you were not required anymore to perform a Windows logon to the machine with your attendee account. A generic, limited user account was already logged-on. Every attendee had to use this account.
This is a bad idea. Even a limited user account can be compromised with spyware, as I’ve shown with my Basic Process Manipulation Tool Kit.
cmd.exe was disabled, but this policy is still easy to bypass:
Tuesday 9 December 2008
Updates: bpmtk and Hakin9; PDF and Metasploit
Hakin9 has published my bpmtk article. The article mentions bpmtk version 0.1.4.0; however, this new version has no new features. But it comes with extra PoC code, like a LUA-mode keylogger and “rootkit”. New blogposts will explain this new PoC code.
And upcoming bpmtk version 0.1.5.0 contains a new feature to inject shellcode. Just have to update the documentation.
On the PDF front: I’ve produced my first Ruby code 
. I worked together with MC from Metasploit to optimize the PDF generation code in this util.printf exploit module. It uses some obfuscation techniques I described 8 months ago.
Thursday 23 October 2008
Excel Exercises in Style
I developed another variant of my “Excel macro injects embedded DLL” script.
In stead of creating and loading a temporary DLL from VBScript, I inject and execute shellcode directly from the VBA application.
Some HIPS would prevent my previous script from running, because it loaded an unapproved DLL. But my new version doesn’t load a DLL.
Of course, writing shellcode is more difficult than developing a PE executable.
Thursday 3 July 2008
bpmtk: A New Version With bpmtk.dll Included
Here is a new version of the Basic Process Manipulation Tool Kit (bpmtk).
Some noteworthy changes:
- bpmtk.dll has been added
- for ASCII: and UNICODE:, now you can specify a string with spaces by enclosing it in double quotes (ASCII:”My Name”)
- write and search-and-write use VirtualProtectEx to change the virtual page protection when a write fails
So now you can also load the bpmtk as a DLL in a process and it will execute its configuration. The configuration is embedded in the DLL as an ASCII string. To change the configuration, you’ve 2 options:
- edit the source code and recompile the DLL
- binary edit bpmtk.dll and insert your config between the strings #BPMTK_CONFIG_BEGIN\r\n and #BPMTK_CONFIG_END\r\n. Terminate all lines of your config with CRLF
Wednesday 25 June 2008
bpmtk: Bypassing SRP with DLL Restrictions
In my last bpmtk post, I argued that although whitelisting DLLs (supplementary to whitelisting EXEs) prevents my Excel macro from loading the embedded DLL, it would be far too difficult to build and maintain such a whitelist of DLLs. However, Cd-MaN commented with a technique to add DLL restrictions to the Software Restriction Policies without building an exhaustive DLL whitelist, the details are in his excellent blogpost (step 3).
In a nutshell, Cd-MaN configures SRP to restrict allowed DLLs to those found in the Windows and Programs Files directories (a restricted user is not allowed to write to these directories).
To bypass this SRP configuration with my Excel macro, I had to update it so that it would perform the process manipulation directly, in stead of doing this manipulation from within the embedded DLL. And here his how I did it:
The trick is to call WriteProcessMemory directly from within the script to disable SRP. Because I didn’t want to recode my search-and-write function from C to VBscript, I hardcoded the addresses to write to (this will only work for the specified version of advapi32.dll). The effect of these 2 WriteProcessMemory calls is to patch advapi32.dll inside the Excel process, thereby disabling SRP so that the embedded DLL is allowed to load (of course, now that SRP is disabled for Excel, I can also just start another program).
A new version of my bpmtk utility (with the DLL version of bpmtk.exe) will be posted soon.
Monday 9 June 2008
Quickpost: Embedding an Executable in a VBscript
My latest bpmtk post got some people to ask me for the VBscript. I’ll do better, I’m posting the Python program I wrote to generate the script. You can download it here.
You have to provide it 2 arguments: the name of the executable to embed and the name of the VBscript to generate, like this:
file2vbscript cmd.exe cmd.vbs
This will generate a VBscript that will write cmd.exe to the current directory and execute it (create a new process). If you want to load a DLL in stead of executing an EXE, use the -l option:
file2vbscript -l mydll.dll mydll.vbs
And to use it in an Office application (Office VBA doesn’t take long subs), use the option -o:
file2vbscript -ol mydll.dll mydll.vbs
This will split the embedded file over several subs, to accommodate for the size limitation of Office VBscripts.
Quickpost info
Thursday 5 June 2008
bpmtk: How About SRP Whitelists?
After having showed you how my Basic Process Manipulation Tool Kit can be used to bypass Software Restriction Policies, I wanted to follow this with a post showing how SRP whitelisting can prevent this. However, while preparing this new post, I got an idea how I could bypass SRP whitelists (under certain conditions), but I’ve no idea how to prevent this. I finally decided to post this without a solution, maybe you’ll come up with one.
With a SRP whitelist, starting a program is denied by default:
As an administrator, you’ve to explicitly specify the programs that are allowed to be executed by your users (if there are many programs, maintaining this whitelist becomes time consuming). Because of this whitelist, tools like gpdisable or bpmtk can’t be executed to disable SRP. However, if I can execute these tools without starting a new process, SRP will not block them …
Applications with embedded scripting can also be used to manipulate processes. For example, the scripting features of Microsoft Office allow you to call the system APIs I’ve been using in my bpmtk. It’s often not easy (even impossible) to convert a C program to VBscript, but I’ve a workaround.
First, we adapt our C program from an EXE to a DLL (entrypoint DllMain in stead of main), because VBscript can load a DLL.
We’ll use Excel’s scripting features. I’ve created an Excel spreadsheet that embeds a DLL that can be executed with a mouse-click:
The MyDLL dialog is displayed by the embedded DLL.
The DoIt button starts this Sub:
DoIt will create a temporary file (in the user’s temporary file folder), write the embedded DLL to it (DumpFile), and then load the DLL (LoadLibrary).
Generating the temporary filename:
Writing the embedded DLL to the temporary file:
Each DumpFileX sub writes bytes to the temporary file (the DLL is embedded in these subs by including the hex dump in strings). It’s necessary to split this over several subs, because of the sub size limitation.
Once the DLL is stored in the temporary file, we call LoadLibrary to load our library in the Excel process. And this executes our code inside the Excel process. Because of this, SRP will not deny it, and our code can disable SRP.
Creating temporary files and loading libraries is normal behavior for programs, SRP will not block this. Even most HIPS will not block this, because loading a library is not the same as injecting a DLL (injecting a DLL is loading a library inside another process). The only thing that might be considered abnormal by the HIPS, is that a temporary file is mapped into memory, but there are also legitimate programs that do this.
SRP has an option to whitelist DLLs, but then you’re facing the huge task of identifying and specifying all DLLs your programs use!
If you implement a SRP whitelist because you absolutely want to control the programs executed by your users, take some time to reflect on your users and the scripting capabilities of your whitelisted applications. And if you really have to prevent the technique I show here, you’ll have to find another solution than SRP whitelists. Unfortunately, I’ve not found one yet… If you’ve an idea, post a comment (banning applications with embedded scripting or disabling scripting is not an option).
Tuesday 3 June 2008
Quickpost: bpmtk Config File Embedding
After a rather long detour in PDF file format land, let’s pick up where we left the bpmtk.
My Basic Process Manipulation Tool Kit requires a configuration file with instructions to manipulate processes, like this one to start cmd.exe in a restricted environment:
start cmd.exe search-and-write module:. unicode:DisableCMD hex:41
Save this configuration in a text file, for example start-cmd.txt. And then start bpmtk with this file:
bpmtk start-cmd.txt
You can also embed this configuration file inside the bpmtk executable, like this:
bpmtk start-cmd.txt bpmtk-cmd.exe
This will create a copy of bpmtk.exe, called bpmtk-cmd.exe, with start-cmd.txt embedded as a resource (called BPMTK).
When you execute bpmtk-cmd.exe (without any arguments), the embedded script will be executed. Use this
trick if you often have to execute the same command, or if you have to execute bpmtk in an environment where you cannot provide an argument.
Quickpost info
Wednesday 28 May 2008
I Still Use Foxit Reader
Foxit Reader has been my default PDF reader for more than a year now, as an alternative to the Adobe Acrobat Reader that stalled too often when starting up.
While playing with the PDF file format, I created several PDF files that uncovered potential security issues with Foxit Reader.
A PDF file with an OpenAction triggering an URI action causes Adobe Acrobat to prompt the user for approval, before accessing the URI:
But Foxit Reader opens Internet Explorer and visits the site without confirmation prompt. I submitted a feature request to Foxit Software for this.
Another example is a JavaScript inside a PDF file that switches the reader to full screen mode. Adobe Acrobat Reader will warn you for spoofing attacks and ask for your permission to switch to full screen, while Foxit Reader does this immediately.
Of course, these warnings will only help a user that is aware of the potential risks. But in a corporate environment, you can also set the appropriate registry keys to block all these actions by default.
It was also trivial to assemble some simple malformed PDF files that cause problems for Foxit Reader, but not for Adobe Reader. I submitted these files to Foxit Software.
Adobe Acrobat Reader allows you to disable JavaScript. Until recently, Foxit Reader required a JavaScript plugin for JavaScript support. Omitting the plugin was a simple way to disable JavaScript. But since version 2.2, JavaScript is embedded in the main executable and there is no configuration switch to disable it. Many Foxit Reader users have requested this feature.
If you absolutely want to disable JavaScript in Foxit Reader 2.3, there’s a quick and dirty trick. Search for the ASCII string JavaScript (preceded and terminated by byte 00) in the Foxit Reader executable (you should find only one occurrence), and replace it with javascript, for example. Actually, this patch will not disable the JavaScript interpreter for Foxit Reader, but it will prevent Foxit Reader from recognizing the /JavaScript name in a PDF document, effectively making it to ignore JavaScript instructions (names are case-sensitive).
You can make this patch permanently by editing the Foxit Reader executable with an hex editor, or do it temporarily by patching in memory with my bpmtk utility. The command to achieve this is:
search-and-write module:. hex:004A61766153637269707400 hex:006A
Of course, this is not a serious risk analysis of Foxit Reader. I started to use Foxit Reader as a solution to the Adobe Acrobat Reader performance problems, not for security reasons. And now that I’ve delved into the PDF file format, I did some random tests with Foxit Reader and Adobe Acrobat Reader. This gave me the impression that Adobe has more experience with security risks and vulnerabilities, than Foxit Software, and that this experience is reflected in the design of their products.
I’ll still be using Foxit Reader as my main PDF reader, and I’ll still analyze suspect PDF files in a controlled environment.
Wednesday 19 March 2008
bpmtk: Spying on IE
I was asked if it’s possible to intercept IE’s HTTPS requests. It is, it’s not difficult, and you don’t need admin rights to do it on your own processes. In other words, a malware doesn’t even need admin rights to spy on your IE process, if said malware is also running under your user account.
We need to hook the API calls to WinINet functions, like HTTPOpenRequest. We can do this by patching the Delayed Import Address Table of executables calling WinINet functions. In our case, to spy on IE, we need to patch the DIAT of urlmon.dll. One simple way to hook these API calls, is to develop a DLL that will patch the DIAT, diverting the calls to our own functions. Our functions will just call the original functions while intercepting the data.
Here is an example for HTTPOpenRequest:
HookHTTPOpenRequestA is our hook function for HTTPOpenRequest. It will just output the flags, verb and objectname parameters to the debugger, and then call the original HTTPOpenRequest function with unmodified arguments (which we saved in variable OriginalHTTPOpenRequestA). BTW, if the declaration and use of OriginalHTTPOpenRequestA looks confusing to you, read the explanation of function pointers in C.
Patching the DIAT is easy, use the PatchDIAT function that I provide with my Basic Process Manipulation Tool Kit (it’s in iat.c).
PatchDIAT needs the name of the executable we want to patch (urlmon.dll), the name of the API to patch (wininet.dll), the name of the function to patch (HttpOpenRequestA), the address of our hooking function (HookHttpOpenRequestA) and a variable to store the address of the original function (OriginalHttpOpenRequestA). PatchDIAT returns S_OK when patching was successful.
We package everything in a DLL, while hooking some other functions, like InternetReadFile (to intercept actual data), and then inject this DLL in IE with my toolkit:
I’ve stored a test file on my server: https://DidierStevens.com/files/temp/test.txt. When you browse to this test file with the patched IE, you’ll see this in Sysinternal’s DebugView:
Lines 0 to 4 indicate the patching IE was successful.
Line 5 shows IE opening a connection to didierstevens.com on port 443 (that’s 1BB in hexadecimal).
Line 6 shows the preparation of an HTTPS GET request to file /files/temp/test.txt. Flags 00C00000 indicate HTTPS and keep-alive.
Line 7 shows that the call to InternetReadFile was successful and read 25 bytes (0×19).
Line 8 shows the actual data retrieved by IE: This is just a text file.
The next lines indicate we unloaded our DLL with success (thus undoing the patch).
As you can see, we can intercept data before it is encrypted by the HTTPS connection (/files/temp/test.txt) and after it is decrypted (This is just a text file.). This works because we patch the executable before it calls API functions that handle the encryption/decryption, so we get access to the unencrypted data.
I kept my demo DLL very simple to show you the basic principles. A complete spying program would have to hook more functions and tie all the data together to present it in a user friendly way.
It’s also simple to adapt my IE spying DLL to tamper with the data. For example, it could redirect IE to another web site by changing the lpszServerName argument before it calls the original InternetConnect function.
Older Posts »
<!-- begin footer -->
<!-- begin sidebar -->
