CI not catching failures reliably

[IAmMarcelJung]

As mentioned in #316, the current CI does not check for all failures, e.g. if a command returns instead of throwing an error so that the FABulous shell does not terminate.

Since we should not rely on the code throwing errors for the current test setup (since this can cause the checks to falsely pass) we should implement another way to check if the test run was actually successful. In the future we should also completely update our test setup, but for now maybe there is a simple soultion for this like checking a log or checking for the existence of file.

[IAmMarcelJung]

This could be a duplicate of #147?

[EverythingElseWasAlreadyTaken]

For a quick fix, I'll add a workflow, which also runs the simulation standalone, without the usage of the CLI.

In general, we should maybe divide down the CLI and the script interfaces, all functions called via scripts should throw errors, the CLI functions should rather be wrappers around them, that should be able to catch these exceptions.