Low interaction honeypot designed for Android Debug Bridge over TCP/IP
The Android Debug Bridge (ADB) is a protocol designed to keep track of both emulated and real phones/TVs/DVRs connected to a given host. It implements various commands designed to assist the developer (
adb push, and so on) in both debugging and pushing content to the device. This is usually done via an attached USB cable, with ample mechanisms of authentication and protection. Turns out though that by a simple adb command (
adb tcpip <port>) sent to an already established connection (through USB for example), you can force your device to expose its ADB services over port 5555, after which you can use a simple
adb connect <ip>:<port> to connect to your device via TCP. However, unlike the USB protocol, the TCP one does not have any kind of authentication and leaves the device prone to all kinds of attacks. Two of them are as follows:
adb shell <shell command> - allows a developer to run all kinds of commands on the connected device such as ls, wget and many others.
adb push <local file> <remote destination> - allows a developer to upload binaries from his own machine to the connected Android device.
Coupled together, these two API calls can allow complete control over the device (legitimate or not) as long as the port is exposed over the Internet.
The purpose of this project is to provide a low interaction honeypot designed to catch whatever malware is being pushed by attackers to unsuspecting victims which have port 5555 exposed.
Right now you can
adb push and
adb shell into it. All of the data is redirected to stdout and files will be saved to disk. CPU/memory usage should be fairly low, any anormalities should be reported so they can be investigated.
Responses to shell commands can easily be added by editing the
responses.py file, currently only the
adb shell ls will return a unique response. All other commands will respond with
command not found
What doesn't work?
More advanced commands (like native directory listing and having and interactive shell) won't work. The main reason is that I haven't found any kind of malware to take advantage of mechanisms like this. I've also had to reverse engineer the protocol flow by hand, so please also provide a .pcap when logging an issue so I can look into it (or VERY exact steps for reproduction). Any improvements will be more than welcome.
OK OK, how do I get it started?
Just start the script in python:
nohup python3 run.py &
Just like that, shouldn't have any more complex dependencies.
The config file
adbhoney.cfg must be in the same directory as run.py or at
Or give the docker container a try, easiest with docker-compose:
docker-compose up --build -d
or without docker compose
docker build -t adbhoney:latest .
docker run --name adbhoney --rm -p 5555:5555 -v $(pwd)/adbhoney.cfg:/etc/adbhoney.cfg adbhoney:latest
You will probably want to save uploads and logs to the host machine, so add these volumes to the run command above
-v $(pwd)/dl:/ADBHoney/dl -v $(pwd)/logs:/ADBHoney/logs