Mimikatz implementation in pure Python. -offline minidump parsing currently-
Runs on all OS's which support python>=3.6
Install it via pip or by cloning it from github.
The installer will create a pypykatz executable in the python's Script directory. You can run it from there, should be in your PATH.
pip3 install pypykatz
pip3 install minidump minikerberos asn1crypto
Clone this repo
git clone https://github.com/skelsec/pypykatz.git cd pypykatz
python3 setup.py install
Dumping LIVE system LSA secrets
pypykatz live lsa
Parsing minidump file of the LSASS process
pypykatz minidump <minidump file>
Using pypykatz -detailed-
Foreword: there is an awesome help menu as well.
The command structure is the following
pypykatz <ouput options> <command> <subcommand (opt)>
-o filed will result in output being printed to
Increasing the number of
v increases the size of memory to be shown on the screen.
Warning! Too much data might result in cross-boundary read attempts! Parameter:
pypykatz.py -vv mindidump <minidumpfile>
Write output to file:
pypykatz.py -o <output_file> minidump <dumpfile>
Write output in JSON
Together with the
-o option it will write the output to a file, otherwise will print the output to
pypykatz.py --json -o <output file> minidump <dumpfile>
Stores the kerberos tickets in BOTH
.ccache formats to the directory given.
WARNING! An output directory is expected, as the
.kirbi format supports only ONE ticket/file so get prepared to be swimming in those files when dealing with multiple/large dump files.
pypykatz.py -k <output_dir> minidump <dumpfile>
Minidump command options
This parameter tells pypykatz to look for all
.dmp files in a given directory
pypykatz.py minidump <folder_with_dumpfiles> -d
Supplying this parameter will force pypykatz to recursively look for
Only works together with directory parsing.
pypykatz.py minidump <folder_with_folder_of_dumpfiles> -d -r
If you want to help me getting this project into a stable release you can send mindiumps of the lsass.exe process to the following link: https://pypykatz.ocloud.de/index.php/s/NTErmGJxA42irfj
IMPORTANT: please DO NOT send dumps of your own machine's lsass process!!! I will be able to see your secrets including hashes/passwords! Send dump files from machines like virtual test systems on which you don't mind that someone will see the credentials. (if you have a test domain system where kerberos is set up that would be the best)
Also I'd apprechiate if you wouldn't spam me...
Why do I need these dumps files?
In order to create mimikatz in Python one would have to create structure difinitions of a gazillion different structures (check the original code) without the help of the build-in parser that you'd naturally get from using a native compiler. Now, the problem is that even a single byte misalignemt will render the parsing of these structures run to an error. Problem is mostly revolving around 32 - 64 aligments, so 32 bit Windows version lsass dumps are apprechiated as well!
I need data I can verify the code on and administer necessary changes on the parsers until everything works fine.
Submitting issues on this github page wouldn't help at all without the actual file and github wouldn't like 40-300Mb file attachments.
First step is to have the minidump file parsing capability done in a platform independent way, so you can enjoy watching secrets in your favourite OS. Currently aiming for full sekurlsa::minidump functionality.
This project is still work in progress, there is no guarantee that anything will stay/look/feel the same from one second to another.
Most of my big python projects are aiming for maximum protability, meaning I only use 3rd party packages where absolutely necessary. As of this point three additional packages are used, and I intend to keep it this way.
David Buxton for the timestamp conversion script