Karkinos
A large library database to assist in binary exploitation on Linux. This tool can help identify unknown libraries by providing the location of known symbols, it can help locate the name of packages that contain a given library and it can find the debugging version of a library you are working with.
Once the library is identified you can dump useful information such as symbols (both exported symbols and special useful calculated locations that are useful for exploitation), gadgets for ROP chains or one shot (AKA magic gadgets or one gadgets) and their constraints.
The usage is inspired by the excellent libc-database.
It supports many architectures including:
- x86 (i386 / amd64)
- arm (arm / arm64)
- mips (mips / mips64)
- ppc (ppc / ppc64)
- sparc (sparc / sparc64)
- sh4
- hppa
- m68k
- riscv
There are many libraries indexed over many distributions spanning many years. The libraries indexed are:
- glibc
- libstdc++
- glibc-ld
- libgcc
- musl
Install
The preferred method is to just download and extract the latest release. Karkinos will manage updates by itself from that point on. Alternatively, you can clone this repository.
Once you have Karkinos through either method, you can then access the tool by running:
$ ./kark.py --help
On the first run it will attempt to extract the databases from the compressed files,
you will need to have the xz and cat binaries installed and availible through the PATH env
var.
Usage
usage: kark.py [-h] [--libdb {glibc,libstdc++}] [--distro DISTRO]
[--arch {x86,amd64,i386,arm,arm64,mips,mips64,ppc,ppc64,sparc,sparc64,m68k,hppa,sh4}]
[--endian {little,big}]
{find,dump,info,update} ...
description:
karkinos is a library database to assist with exploitation by helping to
identify libraries from known offsets or to dump useful offsets from those
identified libraries. Each database indexes symbols, gadgets and where
possible one shot gadgets (AKA magic gadgets or one gadgets).
architectures indexed:
- x86 (amd64, i386)
- arm (arm, arm64)
- mips (mips, mips64)
- ppc (ppc, ppc64)
- sparc (sparc, sparc64)
- m68k
- hppa
- sh4
- riscv
libraries indexed:
- glibc
- libstdc++
- glibc-ld
- libgcc
- musl
commands:
- find find a library by symbol offsets, file, build id or file hash
- dump dump symbols/gadgets for a given library
- info print some information about a specific library
- update check for updates to the database
- version display version information and exit
positional arguments:
{find,dump,info,update}
command to execute
args arguments for specific command, see examples
optional arguments:
-h, --help show this help message and exit
--libdb {glibc,libstdc++}
the library database to use
--distro DISTRO the linux distribution to filter in symbol search
--arch {x86,amd64,i386,arm,arm64,mips,mips64,ppc,ppc64,sparc,sparc64,m68k,hppa,sh4}
architecture to filter in symbol search
--endian {little,big}
endianess to filter in symbol search
examples:
./kark.py find fgets b20 puts 9c0 fwrite 8a0
./kark.py find 50390b2ae8aaa73c47745040f54e602f
./kark.py find b417c0ba7cc5cf06d1d1bed6652cedb9253c60d0
./kark.py find /lib/x86_64-linux-gnu/libc.so.6
./kark.py --arch arm --endian big find system 440
./kark.py --distro ubuntu fgets b20 puts 9c0
./kark.py dump centos_glibc-2.12-1.107.el6_4.2.x86_64
./kark.py dump opensuse_glibc-2.19-16.9.1.i686 fgets system str_bin_sh
./kark.py info ubuntu_libc6-udeb_2.27-3ubuntu1_amd64
./kark.py update
Screenshots
TODO
- Make the database faster (queries are far from optimised)
- Make the database smaller (schema is not optimal)
- Clean the code up (was hobbled together very quickly)
- More gadgets, more one shot gadgets
- Make usable as a library (remove output, reorganise code)