Skip to content

Latest commit

 

History

History
196 lines (174 loc) · 7.52 KB

README.md

File metadata and controls

196 lines (174 loc) · 7.52 KB
Raw

Spwn

This repository started as a translation of pwninit. It has been created because I love the utilities provided by pwninit, but I'm too lazy to learn Rust and I wanted to customize it, so I rewrote it in python (and added some a lot more features).

Features

  • Auto detect files (binary, libc, loader)
  • Download loader from libc version (if missing)
  • Unstrip the libc with pwn.libcdb.unstrip_libc
  • Set binary and loader executable
  • Set runpath and interpreter for the debug binary
  • Generate a basic script from a template
  • Interactively generate functions to interact with the binary
  • Print basic info about the files:
    • file
    • checksec
    • libc version
    • potentially vulnerable functions
    • cryptographic constants
    • seccomp rules
  • Launch decompiler
  • Run cwe_checker
  • Launch custom user-provided commands
  • Launch custom user-provided python scripts

Usage

spwn [-h] [-i] [-so] [-io] [-nd] [--config] [{inter,i,ionly,io,nd,nodecomp,config} ...] [template]

options:
  -h, --help       show this help message and exit
  -i, --inter      Interactively create interaction functions
  -so, --sonly     Create the interaction script without analyzing the binary
  -io, --ionly     Create the interaction functions, without doing any analysis
  -nd, --nodecomp  Don't open the decompiler
  --config         Setup configs and quit

If the files have weird names (such as the libc name not starting with "libc"), the autodetection will fail and fall in the manual selection, the best fix for this is to rename the files.

To understand how the interactions creation works, I suggest to just try it out. It should be pretty straight forward, but if you want to pwn as fast as possible, you cannot waste any time :)

Installation

Non python tools:

sudo apt update
sudo apt install patchelf elfutils ruby-dev
# Or the equivalent for you package manager
sudo gem install seccomp-tools  # Might not need `sudo`

To install cwe_checker follow the instructions in their repository.

Main package:

pip install spwn

You might need to add ~/.local/bin/ to your $PATH

Files layout

Files in the current working directory don't get modified. If there is any library, the debug_dir directory gets created, all the ELFs get copied there and the binary in that folder is modified to set the run path and the interpreter. Everything is set relative to the current working directory, thus, even if the files you want to work with are in ./{debug_dir}/, you should run everything from the current working directory. If there is only one binary, no debug_dir gets created, because there is no need to modify it.

Customization

This tool is written because I wanted to customize pwninit as much as possible. If you want to customize your own spwn you can:

  • Clone this repo
  • Modify whatever you want
  • In the repository's root directory: pip install -U .

or directly modify the files in: ~/.local/lib/python3.{version}/site-packages/spwn

Note that the default configurations and templates, gets written only if they are not already present (or updated if some fields are missing), so, if you want to customize those, you have to modify the files as specified in the configurations section.

Configurations

You can configure some stuffs in the config.json file. Configuration gets written in ~/.config/spwn/ in the first run of spwn or by calling spwn --setup. In the same directory you will also find template.py, the template of the script generated by spwn, which you can modify to your liking.

This is the default config.json file:

{
    "debug_dir": "debug_dir",
    "script_file": "a.py",
    "pwn_process": "r",
    "tab": "\t",
    "template_file": "~/.config/spwn/template.py",
    "custom_template_prefix": "template_",
    "suppress_warnings": false,
    "yara_rules": "~/.config/spwn/findcrypt3.rules",
    "preanalysis_commands": [],
    "postanalysis_commands": [],
    "preanalysis_scripts": [],
    "postanalysis_scripts": [],
    "idafree_command": "",
    "decompiler_command": ""
}

Multiple templates

You can have multiple templates and select which one to use from command line. You have to place your templates in the same directory of the base template (template_file), and name it {custom_template_prefix}{name}.py. To use it, you just have to specify name in the command line (spwn {name}).

The whole file will be treated as a format string, so, be careful to put double curly brackets if they don't have to be treated as format specifiers (my_set = {{1, 2, 3}}). The actual format specifiers (which you have to place in single curly brackets) are: {binary}, {libc}, {debug_dir} and {interactions}.

Suppress warnings

If set to true, don't show warning messages for non installed non-vital dependencies.

Custom commands

The pre and post analysis commands, are in the form [command, timeout]. command is a list of strings and should contain the "{binary}" or "{debug_binary}" string in order to be formatted with the correct executable path. You should use debug_binary only in the post analysis and if your command will run the binary. If you set timeout to false, the program gets run with subprocess.Popen, thus the analysis will go on while running it and the process will go on after spwn will have terminated. This might be used, for example, to run the ROP-gadgets search in the background. If you want to run the program without a timeout (discouraged) you can set it to null. A couple of examples are:

["one_gadget {binary}", 1]
["ropr -njs {debug_binary} > gadgets", false]

Custom script

You can even run whole python scripts, all you have to do is to specify their path in the preanalysis_scripts or postanalysis_scripts. If you just provide the file name, it will be searched in the config directory. The scripts must contain a main function that takes one parameter: files. This parameter is a FileManager object and its structure is as follows:

class FileManager:
    # Three `Binary` objects
    self.binary
    self.libc    # Can be None
    self.loader  # Can be None
    # libc and loader have their own type that are a subclasses of `Binary`
    self.other_binaries  # list of relative paths

class Binary:
    self.name  # relative path to the original binary
    self.debug_name  # relative path to the debug binary, if there is none it is equal to `self.name`
    self.pwnfile  # `pwn.ELF` object

For example:

def main(files):
    print(f"The binary is {files.binary.name}")

Decompiler

For the decompilers commands, the syntax is the same of the pre and post analysis commands. I created a special config, rather than putting it in a pre analysis command, because I use IDA freeware and it can decompile only x86/x86_64 binaries, so I have to use another decompiler for other architectures (I have created this feature before the custom scripts thing, but since the decompiler is something that you will almost always launch, I left it to make it easier to use). If you want to use always the same decompiler, leave idafree_command empty and if you don't want to launch any decompiler, just leave both configs empty. If you wish to modify the conditions to select the decompiler, you can either modify the open_decompiler function in analyzer.py or create a custom script.

If you have any question or feature request, feel free to ask here.