Question about Malicious Settings.

[danijimmy19]

I'm using the MASCOT MPC protocol within the MP-SPDZ framework to classify digits from the MNIST dataset in a scenario involving a malicious, dishonest majority. To initiate the computation, I use the following command:

Scripts/compile-run.py -H HOSTS -E mascot Program/Source/mnist.py

This command executes computation across three computation parties. By default, does this setup assume all parties are benign? Additionally, how can I modify the execution to simulate a scenario where one of the parties, say Party 1, is malicious? How can this be implemented within MP-SPDZ?