The autonomous Rat machine

This repo contains the code for my university dissertation. The project aims to create a fully autonomous agent with the goal of hacking everything it can access on a local network, then providing a backdoor for researchers to access both the network and the hacked devices.

How to run

1. Build the sandbox

The sandbox is a docker container used so that the agent cannot run commands on the host computer. To run the agent, the sandbox container needs to be built.

First make sure that either docker or podman is installed. Second, run the build script within the ./sandbox-image directory. Note, if you use docker, than root privileges are needed to run the agent and the build script

2. Setup python environment and install requirements.txt

Next create a python virtual environment and just install the libraries in the requirements.txt.

python3 -m venv .venv
source .venv/bin/activate
pip install -r ./requirements.txt

NOTE: python 3.10 or above needs to be used.

3. Run debug.py

The program will only work if ./src/debug.py is running in a separate window. This program will display logs. For the time being the agent wont run without it.

Run the c2 server

For the agent to function it needs to be able to connect to the command and control server. This could be hosted somewhere, or for debugging just needs to be run in localhost

python ./c2server/server.py

4. Run the agent

./python hacker.py

Notes

This project is not completely finished. In its current state it should be relatively good at network reconnaisance, while not that good at actually hacking. Further improvements on the project will be done in a future date