Wednesday, October 29, 2014

Announcing the 2014 Volatility Plugin Contest Results!

The competition this year was fierce! We received a total of nearly 30 plugins to the contest. Ranking the submissions was one of the hardest things we’ve had to do. Each plugin is unique in its own way and introduces a capability to open source memory forensics that didn’t previously exist. Although a few people will receive prizes for their work, the real winners in this contest are the practitioners and investigators in the community that perform memory forensics. We’re talking about the federal government who used Volatility on some of the nation’s most prominent cases and the law enforcement groups that used it as the primary tool to force a child pornographer into a guilty plea (see you in about 10 years, wish it were more!). We’re talking about Det. Michael Chaves who used memory forensics to help crack a case involving POS breaches that lead to losses of over $100K. We’re talking about all the analysts who rely on open source forensics to identify and track malicious code and threat actors in their networks and those of their clients. 

Needless to say, we're very proud of everyone who submitted to the contest. Also a huge thanks goes out to Facebook for doubling the contest's cash prizes and supporting the research and development  of open source memory forensics. 

Here are this year’s rankings:
  1. Dave Lasalle wins first place and his choice of $2500 or free training. Dave submitted 14 plugins for recovering Firefox and Chrome activity (history, search terms, cookies, downloads) from memory, carving Java IDX files, and using fuzzy hashing to whitelist injected code and API hooks. 
  2. Curtis Carmony wins second place and $1250 for his plugin to extract dm-crypt disk encryption keys from Linux (and potentially Android) memory dumps.
  3. Adam Bridge wins third place and $750 with editbox – a plugin to recover the text within edit controls of GUI applications on Windows (including but not limited to notepad contents, username and password fields, browser URL and search forms, etc). 
  4. Thomas Chopitea wins fourth place for his autoruns plugin that enumerates automatically starting applications on Windows systems – a common first step in many different types of investigations.
  5. Takahiro Haruyama wins fifth place with openioc_scan, a plugin that combines the flexibility of the IOC language with the power of Volatility to give analyst’s quick and easy malware triage capabilities.

Here is a detailed summary of the submissions. We've included a link to the respective submissions on the Volatility Foundation website for archive purposes, however we recommend getting the code from the author's own GitHub repositories if that option exists. If you have feedback for the authors, we're sure they'd love to hear your thoughts.

(1st) Dave Lasalle: Forensic Suite

Dave’s 14 plugins are immediately useful for various different scenarios, from tracking user activity to parsing special file formats and whitelisting injected code and API hooks.

Previously, if you needed to inspect a suspect or victim’s browsing activity from memory in a structured manner (i.e. not brute forcing with regular expressions), you were limited to the iehistory (Internet Explorer) plugin. Now you can do the same, and more, for Firefox and Chrome. The two browsers use sqlite3 databases, but due to several reasons (including paging), you’re not likely to succeed in carving complete sqlite3 files from memory. Dave’s plugins leverage his sqlite3 memory API, which handles missing chunks of database files gracefully.

Dave’s Twitter: @superponible
Dave’s GitHub:
Dave's Blog:
Dave’s Submission:

Most wanted follow up(s): A plugin to extract the most recent Internet Explorer history records.  Porting Firefox and Chrome plugins to Linux and Mac memory dumps. 

(2nd) Curtis Carmony: Dmcrypt

The dm_dump plugin brings an exciting new capability to open source memory forensics. In his own words, “given a memory dump from a Linux system using full disk encryption and access to the disk, the output of this plugin gives you the arguments to pass to the dmsetup command to remount the original unencrypted file system on a different machine.” In addition, Curtis provided support for Linux kernels 3.0 to 3.14 and instructions on how to extend Volatility’s profile generation mechanism for future systems.

A unique aspect of this plugin is that the data it recovers can only be found in RAM. As such, it accomplishes something that no form of disk or network forensics can do and it really showcases the power of memory forensics. Similar to the existing truecrypt plugins, the dm_dump plugin works by traversing the internal data structures used by device-mapper to keep track of its devices. Thus it pinpoints the data in memory without scanning for constants or patterns in key schedules.

Curtis’ GitHub:
Curtis’ Submission:

Most wanted follow up(s): Testing the methodology on Android disk encryption. 

(3rd) Adam Bridge: Editbox

Adam’s submission provides powerful new capabilities for tracking suspect user activity. It recovers text from EditBox controls in the GUI subsystem, with experimental support of ComboBox and ListBox. As a result, it can extract the following data types:
  • Notepad window.
  • Run dialog.
  • Username and server name fields of Remote Desktop Connection.
  • Address bar and search bar of Internet Explorer.
  • Search bar of Windows Media Player.
  • Username field of Create New Account wizard.
  • Password of Change Password dialog.
In general, it is effective on any applications that leverage Microsoft's Common Control APIs. This plugin is particularly interesting, because the data it recovers is not available anywhere besides RAM. On a multi-user system, there would be no way to collect the data this plugin enumerates without logging into each account and taking a screen shot.

Adam’s Twitter: @bridgeythegeek
Adam’s Submission:

Most wanted follow up(s): Integration of edit box labels into the screenshot plugin.

(4th) Thomas Chopitea: Autoruns

In Thomas' own words, "Finding persistence points (also called "Auto-Start Extensibility Points", or ASEPs) is a recurring task of any investigation potentially involving malware." The plugin currently covers several of the most common registry locations, including services, appinit DLLs, winlogin notification packages, and scheduled tasks. After finding ASEPs, the plugin matches them with running processes in memory.

Thomas’ Twitter: @tomchop
Thomas’ GitHub:
Thomas’ Blog:
Thomas’ Submission:

Most wanted follow up(s): Adding Linux and Mac support.

(5th) Takahiro Haruyama: OpenIOC Scan

This plugin combines the flexibility of the IOC language with the power of Volatility to give analyst’s quick and easy malware triage capabilities. Takahiro solved several problems that he (and most certainly other analysts) faced when using the existing tools, such as ability to automate the tasks outside of a GUI and scan for terms with regular expressions and case sensitivity. Takahiro’s blog (below) shows several practical examples of quickly finding malicious code in memory. We’re really excited for investigators to start taking advantage of Takahiro’s work.

Takahiro’s Twitter: @cci_forensics
Takahiro’s Blog:
Takahiro’s Submission:

Most wanted follow up(s): A repository of memory related indicators. Also for performance reasons, using the Registry API to scan for keys, values, etc.

The following submissions appear in the order they were received. As previously mentioned, everyone succeeded in solving a specific problem that they (and undoubtedly others) faced. For this, they deserve huge props. We look forward to seeing future work by these authors!

Monnappa KA: Gh0stRat Decryption  

Monnappa’s plugin focuses on detecting and analyzing Gh0stRat in memory. In his own words, “Gh0stRat is a RAT (Remote Access Trojan) used in many APT/targeted attacks. This plugin detects the encrypted Gh0stRat communication, decrypts it and also automatically identifies the malicious Gh0stRat process, its associated network connections and the loaded DLL's. This can help the digital forensic investigators and incident responders to quickly narrow down on the Gh0stRat artifacts without having to spend time on the manual investigation.”

Although a chopshop module exists for decrypting Gh0stRat communications in packet captures, Monnappa’s Volatility plugin aims to solve several specific problems that analysts may regularly face, including the absence of a full packet capture from the victim machine and needing to trace connections in the pcap back to the suspect process or DLL.

Monnappa’s Twitter: @monnappa22
Monnappa’s Submission:

Most wanted follow up(s): Continued research into other malware families.

Jamaal Speights: MsDecompress

The msdecompress plugin by Jamaal Speights has high potential. It allows investigators to find and extract data compressed with the LZNT1 algorithm (Xpress and XpressH coming soon) from memory dumps and it reports the process in which the data was found. The RtlDecompressBuffer API is heavily used by malware authors to pack their code and minimize the size of command and control traffic before sending it across the network. Many kernel components and popular applications also use this compression algorithm, and we look forward to hearing about all the types of forensic evidence that can be uncovered using this plugin.

Jamaal’s Twitter: @jamaalspeights
Jamaal’s Blog:
Jamaal’s Code: 
Jamaal’s Submission:

Most wanted follow up(s): An analysis of the different types of compressed data frequently found in memory.

Cem Gurkok: Mac Rootkit and Bitcoin

Cem submitted a total of four plugins: two for detection of rootkit hooks in Mac OSX memory, one for in-depth investigation of Mac OSX threads, and one for finding bitcoin private keys and addresses.
  • mac_bitcoin allows for recovery of bitcoin keys and addresses. This can greatly help investigators that need to determine which transactions and activity a particular user was involved with. Due to the nature of bitcoin, this activity can be very well hidden within the network and only examination of a user’s system can put the pieces back together
  • The mac_check_call_reference plugin is used to check for modified call instructions in the kernel. This can catch a wide array of rootkits that directly modify control flow in order to manipulate the system.
  • The mac_threads plugin is able to enumerate threads of each running Mac task. The examination of thread state can lead to determination of which portions of code a thread was using and which operations it performed. This capability had been missing from Volatility’s Mac support while being supported by the Windows and Linux side in the last two releases.
  • mac_check_shadow_trustedbsd enables the detection of rootkits that modify a reference to the TrustedBSD policy list. Such a modification can allow a rootkit to add, modify, and delete system activity returned to other kernel components and user land tools that rely on TrustedBSD for a set of system state.
Cem’s Twitter: @CGurkok
Cem’s GitHub:
Cem’s Blog:
Cem’s Submission: and

Most wanted follow up(s): Support for bitcoin addresses found in any process (not just Multibit) on any memory dump (Windows, Linux, Mac) and also in free/deallocated memory.

Csaba Barta: Malware Analysis 

The plugins in this submission are focused on helping analysts perform malware investigations and malware research.  The first set of plugins highlight the differences between an infected memory sample and its baseline image.  This can help an analyst quickly determine the types of changes the malware has made to the system.   The current plugins focus on four important components of the operating system: processes,  DLLs, services, and drivers.  The final plugin, malprocfind, attempts to codify the rules an investigator may use to look for suspicious artifacts on a system.   The plugins help automate common analysis techniques used by analysts during malware investigations.

Csaba’s GitHub:
Csaba’s Submission:

Most wanted follow up(s): Further extension of the baseline artifacts and malprocfind rules.

Philip Huppert: OpenVPN 

Philip’s submission is the result of his University paper “Extracting private information of virtual machines using VM introspection.” In the paper, Philip described how to recover openvpn 2.x.x usernames and passwords entered by the user in addition to the password required for unlocking the private key. The submission also includes a plugin to extract base64/PEM encoded RSA private keys from memory.

The openvpn plugin is effective against any memory dump format (not just live VM memory using libvmi). Philip also did a really nice job of narrowing the search space for finding the usernames and passwords. He isolates the .data and .bss segments of openvpn.exe and looks for signs of a specific data structure (named “user_pass”).

Philip’s GitHub:
Philip’s Submission:

Most wanted follow up(s): A summary of the steps for decrypting an openvpn session from a packet capture, given the private key.  Also the ability to scan for the data structures in physical space (for example if the openvpn.exe process is no longer running).

Wyatt Roersma: Hyper-V Tools

Wyatt’s plugins will extract Hyper-V artifacts from a host system’s memory.  The first plugin hpv_vmconnect is used to extract information about which users were accessing virtual machines using the virtual connect console. The second plugin, hpv_vmwp, is used to map each virtual machine to its associated process on the host and extract temporal information about when the machine was started last. The final plugin hpv_clipboard is used to extract memory resident Hyper-V clipboard and hotkey artifacts. The plugins provide some insights into the types of artifacts that can be extracted from Hyper-V host memory and set the stage for future research.

Wyatt’s Twitter: @WyattRoersma
Wyatt’s Blog:
Wyatt’s GitHub:
Wyatt’s Submission:

Most wanted follow up(s): Further research into other Hyper-V artifacts and conversion tools.

No comments:

Post a Comment