I'm designing an A.I. that Emerged from a tank's pilot program. Her personality is very easily summed up as "MOAR DAKKA!" So while the basic rigging stuff isn't too hard to work with (Piloting Origin, lots of nuyen blown on drones, guns, and softs), Matrix rules can sometimes confound me. It seems like the overwhelming majority believe that cybercombat just isn't worth it. First, because if your caught then you've failed, time to pop smoke and give it another go. Second, because, for A.I.s and technomancers, if you die in virtual reality, you die in reality, too. But I had trouble convincing myself that she would be gung-ho about combat encounters in the real world and a stealthy ninja in the virtual world. So, my questions are, can an A.I. reasonably gain mastery over cybercombat in such a way that she would not need Stealth (at least to get in - maybe to get out after leaving a smoking hole in the firewall)? Is it worth the necessary investment? Can she take advantage of Agent autosofts such as Cascading and Expert Offense to gain an upper hand against spiders and IC? Or should I leave the hacking bit to agents and the building-toppling to her? Thanks in advance.