pwncat is a post-exploitation platform
for Linux targets. It started out as a wrapper around basic bind and reverse shells and has grown from there. It streamlines common red team operations while staging code from your attacker machine, not the target.
pwncat used to only support Linux, but there has been a lot of work recently to support multiple platforms. Currently, there is alpha support for Windows targets. Please see the latest documentation for details on how to use pwncat with a Windows target.
pwncat intercepts the raw communication with a remote shell and allows the user to perform automated actions on the remote host including enumeration, implant installation and even privilege escalation.
After receiving a connection, pwncat will setup some common configurations for working with remote shells.
- Disable history in the remote shell
- Normalize shell prompt
- Locate useful binaries (using
- Attempt to spawn a pseudoterminal (pty) for a full interactive session
pwncat knows how to spawn pty’s with a few different methods and will cross-reference the methods with the executables previously enumerated. After spawning a pty, it will setup the controlling terminal in raw mode, so you can interact in a similar fashion to
pwncat will also synchronize the remote pty settings (such as rows, columns,
TERM environment variable) with your local settings to ensure the shell behaves correctly with interactive applications such as
John Hammond and I presented
pwncat at GRIMMCon. Our presentation, which can be found on YouTube here. This video demonstrates an early version of the API and interface. Please refer to the documentation for up to date usage and API documentation!
pwncat documentation is being built out on Read the Docs. Head there for the latest usage and development documentation!
pwncat requires Python 3.9+.
pwncat now supports windows starting at
v0.4.0a1. The Windows platform utilizes a .Net-based C2 library which is loaded automatically. Windows targets should connect with either a
powershell.exe shell, and pwncat will take care of the rest.
The libraries implementing the C2 are implemented at pwncat-windows-c2. The DLLs for the C2 will be automatically downloaded from the targeted release for you. If you do not have internet connectivity on your target machine, you can tell pwncat to prestage the DLLs using the
--download-plugins argument. If you are running a release version of pwncat, you can also download a tarball of all built-in plugins from the releases page.
The plugins are stored by default in
~/.local/share/pwncat, however this is configurable with the
plugin_path configuration. If you download the packaged set of plugins from the releases page, you should extract it to the path pointed to by
Aside from the main C2 DLLs, other plugins may also be available. Currently, the only provided default plugins are the C2 and an implementation of BadPotato. pwncat can reflectively load .Net binaries to be used a plugins for the C2. For more information on Windows C2 plugins, please see the documentation.
Currently, there are two versions of pwncat available. The last stable version is
v0.3.1. There is a tagged commit for that version. It does not support multiple platforms or multi-session interaction. The documentation for that version is still available on Read the Docs as the
master branch is
v0.4.3. This version has overhauled a lot of the framework to support multiple platforms and multisession environments. Documentation for this version is available in the
latest version on Read the Docs.
v0.3.1 will not be updated further
Recently, the architecture of the pwncat framework was redesigned to encorporate a generic “module” structure. All functionality is now implemented as modules. This includes enumeration, persistence and privilege escalation. Interacting with modules is similar to most other post-exploitation platforms. You can utilize the familiar
info commands and enter module contexts with the
use command. Refer to the documentation for more information.
pwncat only depends on a working Python development environment. In order to install some of the packages required with
pip, you will likely need your distribution’s “Python Development” package. On Debian based systems, this is
python-dev. For Arch, the development files are shipped with the main Python repository. For Enterprise Linux, the package is named
pwncat is configured as a standard python package with
distutils. You can install
pwncat directly from GitHub with:
pip install git+https://github.com/calebstewart/pwncat.git
Or, you can install after cloning the repository with:
python setup.py install
It is recommended to install pwncat from a virtual environment.
python3 -m venv pwncat-env
python setup.py install
If you would like to develop custom privilege escalation or persistence modules, we recommend you use the
develop target vice the
install target for
setup.py. This allows changes to the local repository to immediately be observed with your installed package.
The setup script will install three binaries. They are all identical, but provide convenience aliases for pwncat. The three binaries are:
Connecting to a Victim
The command line parameters for pwncat attempt to be flexible and accept a variety of common connection syntax. Specifically, it will try to accept common netcat and ssh like syntax. The following are all valid:
# Connect to a bind shell
pwncat 10.10.10.10 4444
# Listen for reverse shell
pwncat -lp 4444
# Connect via ssh
pwncat ssh://user:[email protected]
pwncat [email protected]
pwncat user:[email protected]
pwncat -i id_rsa [email protected]
# SSH w/ non-standard port
pwncat -p 2222 [email protected]
pwncat [email protected]:2222
# Reconnect utilizing installed persistence
# If reconnection failes and no protocol is specified,
# SSH is used as a fallback.
pwncat reconnect://[email protected]
pwncat reconnect://[email protected]
pwncat [email protected]
By default, pwncat assumes the target platform is Linux. In order to connect to a Windows reverse or bind shell, you must pass the
pwncat -m windows 10.10.10.10 4444
pwncat -m windows -lp 4444
For more information on the syntax and argument handling, see the help information with
pwncat --help or visit the documentation.
The recommended installation method is a Python virtual environment. This provides the easiest day-to-day usage of
pwncat. However, there has been interest in using
pwncat from a docker image, so I have provided a Dockerfile which provides a working
pwncat installation. To build the image use:
docker build -t pwncat .
This will build the
pwncat docker image with the tag “pwncat”. The working directory within the container is
/work. The entrypoint for the container is the
pwncat binary. It can be used like so:
# Connect to a bind shell at 10.0.0.1:4444
docker run -v "/some/directory":/work -t pwncat 10.0.0.1 4444
In this example, only the files in
/some/directory are exposed to the container. Obviously, for upload/download, the container will only be able to see the files exposed through any mounted directories.
Features and Functionality
pwncat provides two main features. At it’s core, it’s goal is to automatically setup a remote PseudoTerminal (pty) which allows interaction with the remote host much like a full SSH session. When operating in a pty, you can use common features of your remote shell such as history, line editing, and graphical terminal applications.
The other half of
pwncat is a framework which utilizes your remote shell to perform automated enumeration, persistence and privilege escalation tasks. The local
pwncat prompt provides a number of useful features for standard penetration tests including:
- File upload and download
- Automated privilege escalation enumeration
- Automated privilege escalation execution
- Automated persistence installation/removal
- Automated tracking of modified/created files
pwncatalso offers the ability to revert these remote “tampers” automatically
The underlying framework for interacting with the remote host aims to abstract away the underlying shell and connection method as much as possible, allowing commands and plugins to interact seamlessly with the remote host.
You can learn more about interacting with
pwncat and about the underlying framework in the documentation. If you have an idea for a new privilege escalation method or persistence method, please take a look at the API documentation specifically. Pull requests are welcome!
pwncat would like to be come a red team swiss army knife. Hopefully soon, more features will be added.
- More privilege escalation methods (sudo -u#-1 CVE, LXD containers, etc.)
- Persistence methods (bind shell, cronjobs, SSH access, PAM abuse, etc.)
- Aggression methods (spam randomness to terminals, flush firewall, etc.)
- Meme methods (terminal-parrot, cowsay, wall, etc.)
- Network methods (port forward, internet access through host, etc.)
pwncat is trying to abstractly interact with any shell with minimal remote system dependencies, there are some edge cases we have found. Where we find them, we do everything we can to account for them and hide them from the user. However, some have slipped through the cracks and been observed in the wild. When this happens,
pwncat will do whatever it can to preserve your terminal, but you may be greeted with some peculiar output or command failures.
While BSD is a Unix-based kernel, in practice it’s userland tools are noticeably different from their Linux counterparts. Due to this, many of the automated features of
pwncat will not work or outright fail when running against a BSD based target. I have tried to catch all errors or edge cases, however there are likely some hiccups which haven’t been fully tested against BSD. In any case, the stabilized shell should function within a BSD environment, but I don’t provide any guarantees.
If I find some time later down the road, I may try to stabilize
pwncat on BSD, but for now my focus is on Linux-based distributions. If you’d like to contribute to making
pwncat behave better on BSD, you are more then welcome to reach out or just fork the repo. As always, pull requests are welcome!