Attempting to make a centralized binary reverse engineering framework for
python. Initially, this makes heavy use of frida
in the backend, but should
be expandable and has already gone beyond Frida in some ways.
pip3 install https://github.com/bannsec/revenge/archive/master.zip
# Or
pip3 install revenge
The goal is for this to be mostly platform independent. Since the backend is python and Frida, it should support Windows, Mac, Linux and Android. YMMV.
Check out RTD for the documentation: http://revenge.readthedocs.org/en/latest/
Specifically watching Windows Messages handling
# Automatically discover Windows message handling locations and show event messages as they are handled.
revenge -I notepad.exe windows_messages notepad.exe
# Only show information about windows message WM_CHAR and WM_KEYDOWN from notepad.exe
revenge -I notepad.exe -rw windows_messages notepad.exe -wm WM_CHAR WM_KEYDOWN
Use Frida stalk to trace through things
# Only look at traces from notepad's Windows Message handler function
revenge stalk notepad.exe --include-function notepad.exe:0x3a50 -I notepad.exe
Find things in memory.
# Find where your string 'hello world' is in notepad (will check for char and wchar versions)
revenge find notepad.exe --string "Hello world"
{'0x55d78c422250': 'StringUTF8', '0x55d78c453820': 'StringUTF8'}
Drop into an interactive shell from the command line
$ revenge ipython ls -f /bin/ls
Spawning file ... [ DONE ]
Attaching to the session ... [ DONE ]
Enumerating modules ... [ DONE ]
Python 3.6.7 (default, Oct 22 2018, 11:32:17)
Type 'copyright', 'credits' or 'license' for more information
IPython 7.5.0 -- An enhanced Interactive Python. Type '?' for help.
In [1]: process
Out[1]: <revenge.process.Process at 0x7fa036bc14e0>
Replacing functions dynamically during execution
# Replace function located at offset 0x64a in a.out binary, returning value 0x123
revenge stalk ./a.out --resume -rf "a.out:0x64a?0x123"
# Disable alarm and ptrace functions
revenge stalk test2 -f ./test2 --resume -rf ":alarm?1" ":ptrace?1"
There have been some changes recently with Frida and it appears they are no longer dual building their pip installable wheels. For the time being, you can build your own version of frida if you want dual architecture support.
Building instructions here: https://frida.re/docs/building/ Be sure to install the latest node: https://www.digitalocean.com/community/tutorials/how-to-install-node-js-on-ubuntu-22-04
Once you have it built, you can force the python output into your path:
export PYTHONPATH=/home/user/frida/build/frida-linux-x86_64/lib/python3.12/site-packages