BASS (pronounced “bæs”) is a framework designed to automatically generate antivirus signatures from samples belonging to previously generated malware clusters. It is meant to reduce resource usage of ClamAV by producing more pattern-based signatures as opposed to hash-based signatures, and to alleviate the workload of analysts who write pattern-based signatures. The framework is easily scalable thanks to Docker.
Please note that this framework is still considered in the Alpha stage and as a result, it will have some rough edges. As this tool is open source and actively maintained by us, we gladly welcome any feedback from the community on improving the functionality of BASS.
You need Docker (installation instructions) and docker-compose (installation instructions) installed. Even if your distribution has packages for those, we recommend you to install them as described in the installation instructions to have the newest versions available. Parts of our software might not work with old versions of docker and docker-compose.
Further, the client to speak to the docker cluster needs the python requests
package installed. This can for example be done with pip install requests if
you use python's pip package manager.
To build the containers, you need to export some environment variables:
IDA_BINARY=... #Make this variable point to your IDA Pro installation binary
IDA_PASSWORD=... #Set this variable to your IDA Pro installation password
IDA_WEB_PASSWORD=... #Set this variable to your IDA Pro restriced web password
cp ${IDA_BINARY} ida7/ida.run
export IDA_PASSWORD
export IDA_WEB_PASSWORD
You need to set the variables whenever you open a new shell that you want to use to build or run BASS.
Normally it should be enough to run docker-compose build in the repository
root directory to build BASS' containers.
If you have a VirusTotal key, export it in the shell where you run the docker
(e.g., export VIRUSTOTAL_API_KEY=xxx in bash).
Run docker-compose up in the project's root directory to start BASS.
Then use the client in client/client.py to submit samples and get the resulting signature.
For example, run python ./client/client.py sample1 sample2 sample3
to generate a signature for the cluster consisting of binaries sample1, sample2
and sample3.
The job object has an exception and exception_trace property which contain information about a raised exception if the job finished with an error status.
Debug logs may be found in the docker volume mounted to /tmp/bass_logs. In
particular it might be helpful to track progress in the most recent log file
via tail -f $( ls /tmp/bass_logs/*.log | tail -n 1 ).
The client is contained in client/.
The folders bass/, bindiff/ and kamino/ contain the docker containers for the specific tools.
Python APIs for the REST interface of kamino and bindiff are in ./bass/python/cisco/bass/docker/.
The k-LCS algorithm is implemented as a C library (source in ./bass/python/src/_lcs.cpp) which is interfaced with ctypes.
If you are looking for a starting point to the signature generation process, have a look at ./bass/python/cisco/bass/core.py.