I’ve spent the last few days studying and understanding how to build my new open source project.
It’s time to code.
I’ve a few days before holidays, let’s make them profitable.
I want to have a basic version of #shellSentry ready before leaving
#buildinpublic
While I was studying how to make #ShellSentry I run in to a little limit I wasn’t aware.
I’ll need to check if a sequence of commands is usual for a user or not.
Ofc an isolation forest can’t do that, until now I can say if a command is or not usual for the user.
So a new rabbit hole opens: Markov models.
#buildinpublic