I remember a month or so back, someone proposed a modified Bust-a-move robot which, for 80k, was a world-class hacker, kicking it with r6 Pilot, Hacking Autosofts, Fuzzy Logic, and such. I liked the idea, though not at such an extreme, and designed a fuzzy plush spider drone for a gnome mage hunter I'm sending after my team. But it got me to thinking about how a Pilot program would hack. I know that an Agent, when it's doing its robohacks, will transfer itself over to the system it's accessing in order to look about inside. I really don't see a Pilot doing this and just abandoning the drone it's made for to go wander to another node.

So, how would a Pilot, with a Hacker skill-equivalent, operate? Would it be just like a metahuman hacker with an embedded com where it keeps it's own home node response and system, or like an agent where it must deploy in the subverted system and deploy its payload programs there?

Just give it a Replicate Autosoft.

technically a agent can do the same as a hacker, that is access multiple nodes at the same time.

yes, its a topic where SR4(A) contradicts unwired, but unless one follow unwired, agents and AI Alike (as the latter use the same rules) are very restricted).

So basically, if the subscription list of the drone allows it, it can send its icon into other nodes to hack, without the need to upload.

and for reference, this is based on unwired p110, "node movement and accounts".

or maybe its not so much a contradiction, as a confusing use of terminology. Basically, unwired talks about accessing and moving, the former being a login, much the same as a hacker, while the latter is the actual act of moving the agent data between nodes.

rechecking how SR4A is worded on the subject, they never actually talk about the need to upload a agent (tho they make a mention of it may requiring the agent to run on a independent node, but not specifically the node its accessing). So i guess it comes down to vague working in SR4A, with unwired being more specific with its use of access and move to separate the two activities.

your basic agent-in-a-box really is just stealth and then the stuff you need it to use. A confusing point however- can your agent use a fetch module? it could simplify things a lot if it could, but that seems like the kind of thing it may need to load command for (to control the add-on to the commlink) as its not a part of the agent.