I wonder if at some point someone will build a SciFi style countermeasure agent.

Essentially, when intrusion detection triggers, an offensive AI is unleashed against the attacker, trying in turn to hack it back.

It would be fun and interesting (in the ancient Chinese curse sense).