OK - so this came up in last night's game session:

Our hacker "Apex" wants to hack into a stack of different things at the same time ... he's got a Response 6/System 6/Firewall 6/Signal 5 commlink, and has three agent programs at rating 6 also ... He has access to every hacking/common use program in the BBB at rating 6 also.

He loads up three of those agents loaded with exploit and stealth and tells them to begin hacking into three different targets (while he uses his own actions to do other things). How many programs is he running on his commlink (for purposes of response drain)?

• Three - just the three agents count as active programs
• Five - the three agents, plus exploit plus stealth, but exploit/stealth only count once
• Nine - the three agents, plus an instance of exploit and stealth for each agent, counted per-instance
• Something else - why?

The bonus round:
What if Apex now wanted to use Exploit/Stealth himself as part of what he was doing? Would he have to activate new instances of those programs, running more programs, or would the already-running instances of those programs be sufficient?

Last night I ruled with "five" but the rules in the BBB were a little vague, so wanted to check after the game ... It seemed a bit strong, being able to start up five rating 6 agents all at once, sending them in different directions a once to go hack the bad guys.

As I understand the clarification in the FAQ. The answer is 9.
Bonus round answer would be 11.

From FAQ: http://www.shadowrun4.com/resources/faq.shtml

"Do all the programs loaded into an agent affect the Response of your commlink if the commlink is running the agent?
Only if the agent is using those programs."

Now, this doesn't really answer your question about them all sharing a program, but since in your example, all agents (and possibly the hacker) all wanted to use the same program at the same time - I believe they would basically all be running a separate instance of that program, and the commlink's response would go down the pooper.

Thanks - good stuff ... I'll let Apex's player know that he won't be able to get away with that again ...

Get Unwired. It goes further in depth with Agent programs and whatnot. Like I bet he's not using an Unrestricted Agent, and I bet he doesn't have a short script written for it. Also, the Ergonomic program option GREATLY increases versatility of agents and commlinks.

Yup, the ergonomic option can up to double your program loudout if you don't mind spending a few nuyen.

That said, the real trick to using agents is to load them on the node you are hacking. That way your programs on running on your node, and your agents are running on your opponents node, and any slowdown that occurs from the agents happens to your enemies and not you.

The problem with that is you can only load one agent at a time into a node. As per Unwired, copied agents all share the same MSP that you can't spoof (unless you recode them), and a node automatically prevents copied agents from accessing the same node (to prevent "Agent Smith" scenarios).

EDIT: You can totally get like 50 agents and have them all do stuff. it'd be expensive though. Botnet attacks would be an easier way to handle that though.

Sure but remember, unless there is a really fragging good reason for the node to stay active, the moment the operating system detects a slowdown one of the first responces should be to reboot.

Actually the first response would probably run an Analyze and notify SysAd, not reboot. If it detected a virus, prolly notify SysAd immediately, then Shutdown until SysAd could run a Virus Protection.

Boldfacing Added

Care to explain where we actually disagree Neraph?

There are easy ways around that, though. An Agent Smith army is entirely possible, and practical. About the only difference is that it can't be done accidentally.

Care to enlighten us?

Basically, agents run on your persona don't use their access ID, they use yours. So, you just run them on your persona. You can cluster a bunch of commlinks together, or use Frank Trollman's HackaStack method, to easily run a hundred agents at once. In either case, you have to set out to do this, you can't get away with this by accident.