A shared library utilizing the ELF standard's .init section and mprotect() to on-the-fly patch instructions into a binary.
The code in options.hpp is setup for the x86_64 version of ls on Linux. (Specifically ls version 8.25)
Currently, all xeno() does is patch a mov rax, <address of test> and after that, a jmp rax call.
test() is internally defined in xenogrip.cpp, all it does at the moment is call print() and prints a message, letting you know the process of patching succeeded. "It worked!"
The current target for INITIAL_WRITE_POINTER is the entry point to main().
INITIAL_WRITE_POINTER should be set to where ever in memory you wish to patch. A common place to set this would be the entry point to a function.
Finding where you want to write to can be done many ways, but using Radare2 and analyzing the binary to find the offset is how I did it.
(Using the lovely example code user Wisk wrote in the question I asked on the Reverse Engineering Stack Exchange, I was able to simplify this stuff MASSIVELY.)
Fire up CMake's GUI, or use the command line and do the usual configure than generate. Run make (or whatever you're using) in the build directory.
Running it is simple. For the example (while in your build directory), run:
LD_PRELOAD=./libxenogrip.so ls
As long as your using the 64 bit version of ls and version 8.25, you should see the message "It worked!" instead of a listing of files in your directory.
You'll also see some information about things as they happen prefixed with [xeno].