Test and Evaluate Unintended Behaviour

[sandorstormen]

We should test whether different unintended behaviors is allowed in the MAL-suite of python tools. These include things like:

• Attack steps having defense steps as children
• Attack steps pointing to themselves
• Associations with the same object at both ends

We should then evaluate if this behaviour should be allowed. Maybe the mal-toolbox should disallow attack steps having defense steps as children, for example?

[mrkickling]

Also adding defenses that point to or-nodes since that seems to not have any real consequences (which I think is questionable but that's a different story).