How do sprites/agents get remote jobs done? Options

[Valatar]

If you have a spirit, you can tell it, 'Go to Mrs. Hendrickson's house down the street and set her cat on fire. Tibbles has used my tulip patch as a litter box for the last time!' And off it will float to go commit feline arson. No problemo.

But if you have an agent, and tell it, 'Okay Lassie, go get that girl's phone number out of her commlink!', it may bark and run off, but we have a problem. Whereas the spirit can just float right through Mrs. Hendrickson's house's walls to perform its dark deed, Lassie has to perform several specialized functions to hack the girl's commlink, search for the data, copy it, and deliver it back to its deranged stalker owner. This seems to put it outside the reach of smaller agents, and way outside the reach of sprites, in that a Data sprite can get the data but not break in, and a Crack sprite can break in but not get the data. Tell a Machine sprite to try to take over an opponent's drone and you'll just get a stupid look in reply; it has no CFs with which to do the job!

This is exacerbated by the fact that a system is a pile of nodes, each with security to deal with, like a hallway full of locked doors, and the tools for unlocking the doors are mutually exclusive from the tools to do anything when you've gotten past the door. You can't just send an agent to search for data for you once you're past a system's firewall; it'll walk into a data node, the local IC will say, 'Who're you?', the agent will say, 'I'd like that data please.', and the IC will reply, 'Frak you.' and melt the agent's face.

So I can only conclude that I'm missing something important here. The whole point of agents and sprites seems to be having a trustworthy little AI to handle side jobs while the hacker's dealing with the big stuff, but none of them actually seem equipped to handle even side jobs. If I have to hack a node, deal with the IC, decrypt a file, and defuse the data bomb before my agent can get the file, at that point I may as well have just picked the stupid file up myself and saved some time.

[Jaid]

agents can use any programs you make available to them. i'm not sure why you see this as being horribly problematic. if you're just referring to the fact that low rating agents aren't going to be able to hack into everything as a trivial task, well... what's your point? you bought a cheap piece of crap, did you expect it was going to turn into some uberhacker? you want to break into regular security with ease, you need better than average agents to do it. (but really, even a rating 3 agent can probably hack the average girl's commlink without too much difficulty... just give it stealth and exploit, and make edit and browse available to it).

with sprites, you may need to use a combination, or a fairly high rating sprite. but they aren't really designed to replace the technomancer anyways. they're supposed to supplement the technomancer.

the solution is to (for example) use a crack sprite to get in, create an account for a data sprite, and have that sprite find the information you need.

for machine sprites, their job isn't to take over machines... it's to control machines you have already taken over.

[kzt]

But if you have an agent, and tell it, 'Okay Lassie, go get that girl's phone number out of her commlink!', it may bark and run off, but we have a problem. Whereas the spirit can just float right through Mrs. Hendrickson's house's walls to perform its dark deed, Lassie has to perform several specialized functions to hack the girl's commlink, search for the data, copy it, and deliver it back to its deranged stalker owner.

That's not a deranged stalker. A deranged stalker would want her address, her alarm codes, the maglock codes to her place and her schedule. That doesn't seem to be any harder to get then her phone number.....

[DireRadiant]

Your sprite/agent/AI failings are that you view them in a sprite/agent/AI centric way. Take the poor node's point of view sometimes. It's sitting there running umpteen bazillion virtual cyberz rooms and there this teensy blip in of an authorized connection, that happens to be your Agent. Why should it care? The Agent is authorized, just like the umpteen bazillion processes it's got going. Those really important cyberz sessions of hot sim the node owner has going.

Your little spoof or exploit is not the only thing happening in the world.

[Falconer]

I think you underestimate the difficulties in sending out spirits as well.

The spirit comes from another metaplane. While highly intelligent, they're also reasonably ignorant.

They don't know who Mrs Hendrickson is, or necessarily what her cat looks like... you send the spirit to the location to torch a small animal and it may very well torch a rat doing her a favor. Generally you need to be able to provide a solid astral description of the target so that it knows the aura involved.


As far as an agent, once you've used your own software to identify the commlink via a sniffer program. You'd just set it off w/ instructions to break in and do whatever. The agent itself does not interact w/ the real world so has no idea which commlink it is unless you give it an address. Doing that requires you to use your own ewar/sniffer to give it targetting information. Not that hard IMO.

Generally when dealing with agents, just write up a simple 'script' type instructions.

Another case of above, would be ordering the agent to use the sniffer identify nodes in the immediate area, then tell you their locations. When it got to the girl it would have given you information you wanted, w/o even breaking in. (at least the commlinks ID... if not the login identity).

[Tymeaus Jalynsfe...]

I think you underestimate the difficulties in sending out spirits as well.

The spirit comes from another metaplane. While highly intelligent, they're also reasonably ignorant.

They don't know who Mrs Hendrickson is, or necessarily what her cat looks like... you send the spirit to the location to torch a small animal and it may very well torch a rat doing her a favor. Generally you need to be able to provide a solid astral description of the target so that it knows the aura involved.


As far as an agent, once you've used your own software to identify the commlink via a sniffer program. You'd just set it off w/ instructions to break in and do whatever. The agent itself does not interact w/ the real world so has no idea which commlink it is unless you give it an address. Doing that requires you to use your own ewar/sniffer to give it targetting information. Not that hard IMO.

Generally when dealing with agents, just write up a simple 'script' type instructions.

Another case of above, would be ordering the agent to use the sniffer identify nodes in the immediate area, then tell you their locations. When it got to the girl it would have given you information you wanted, w/o even breaking in. (at least the commlinks ID... if not the login identity).

This ^^^^^^^

Agents are very useful in this regard... for simple jobs, even a rating 2-3 agent will work just fine... the higher end agents can be almost as competent as a true hacker, thoguh a little bit limited in its decision making processes, as they have no true intuition to draw upon...

Keep the Faith

[Jaid]

I think you underestimate the difficulties in sending out spirits as well.

The spirit comes from another metaplane. While highly intelligent, they're also reasonably ignorant.

They don't know who Mrs Hendrickson is, or necessarily what her cat looks like... you send the spirit to the location to torch a small animal and it may very well torch a rat doing her a favor. Generally you need to be able to provide a solid astral description of the target so that it knows the aura involved.

if you've seen the cat, should be pretty easy. you tell the spirit to find and destroy the cat. the spirit uses the search power (which shouldn't be too hard, the cat is probably not all that far away), locates the cat, and then nukes the crud out of it. *bamf*, service completed.

but otherwise, i agree.

[Draco18s]

Your sprite/agent/AI failings are that you view them in a sprite/agent/AI centric way. Take the poor node's point of view sometimes. It's sitting there running umpteen bazillion virtual cyberz rooms and there this teensy blip in of an authorized connection, that happens to be your Agent. Why should it care? The Agent is authorized, just like the umpteen bazillion processes it's got going. Those really important cyberz sessions of hot sim the node owner has going.

Your little spoof or exploit is not the only thing happening in the world.

You're not viewing this from the sprite's perspective.

Hacker Sprite: "Copy? What's Copy?"
Data Sprite: "Spoof? Exploit? I can't reach the target."
Machine Sprite: "++?????++ Out of Cheese Error. Redo From Start." *

*Bonus points for catching the reference.

[zeborazor]

Not to derail this or anything but can someone explain sprites? Can a technomancer summon like a rating 10 sprite? Whats stopping them?

[Valatar]

Not to derail this or anything but can someone explain sprites? Can a technomancer summon like a rating 10 sprite? Whats stopping them?

The likelihood of failure and severe wounds. If a technomancer tries to summon above his resonance score, he takes physical damage for fading, just like a mage will take physical damage drain if they try to summon a spirit above their magic score.

[zeborazor]

And if I crash a node with a sprite in it what happens? Is it decompiled? Can a TM call it back to him? does he have to re-compile it?

[Valatar]

And if I crash a node with a sprite in it what happens? Is it decompiled? Can a TM call it back to him? does he have to re-compile it?

If the sprite's with the technomancer at the time, I think it's considered to be riding shotgun on his node and thus wouldn't be inconvenienced if the node the TM's standing in goes down. If it's off on its own, it would actually be loaded into the node it was visiting, just like a roaming agent, so dropping the node would kill the sprite. If it was a registered sprite, the technomancer could go on a quest to the resonance to retrieve it. If it wasn't registered, it'd just be gone.

That, at least, is my understanding of the rules.

[Da9iel]

Of course if it wasn't registered and it's not riding shotgun with the technomancer, it's on remote service and will be gone when it's done anyway.

[Jaid]

You're not viewing this from the sprite's perspective.

Hacker Sprite: "Copy? What's Copy?"
Data Sprite: "Spoof? Exploit? I can't reach the target."
Machine Sprite: "++?????++ Out of Cheese Error. Redo From Start." *

*Bonus points for catching the reference.

well, yeah, but funny story... a lot of people originally complained that sprites basically make the technomancer obsolete in terms of doing anything for themselves.

so your sprites aren't the greatest at doing things alone. put them in teams. nobody complains that a secure facility requires more than one shadowrunner to break into, and that's way more resources than sending 2 sprites.

[Saint Sithney]

the higher end agents can be almost as competent as a true hacker, though a little bit limited in its decision making processes, as they have no true intuition to draw upon...

Can you install Fuzzy Logic circuits in a comlink? (IMG:style_emoticons/default/grinbig.gif)

[Valatar]

I was surprised to find out that the humble Courier can actually handle entering a node to find data on its own, if it's at least a 6 rating to allow for adding Browse and Exploit. All this time I'd skimmed over Courier without paying it much heed, thinking, "Why would I want to compile a sprite to ferry some data? They have this great thing called e-mail already." Well, turns out that the delivery boy sprite has the holy trinity of Stealth, Exploit, and Browse. Sleuth sprites do as well, but aren't accessible by the bulk of technos. So while I still wouldn't trust a sprite to hack an entire system for me, it seems that it is at least possible to send one to fetch lightly-defended stuff on its own.

[Tymeaus Jalynsfe...]

Can you install Fuzzy Logic circuits in a comlink? (IMG:style_emoticons/default/grinbig.gif)

Not sure... probably not though... it would be interesting if you could though...

Keep the Faith

[Jaid]

I was surprised to find out that the humble Courier can actually handle entering a node to find data on its own, if it's at least a 6 rating to allow for adding Browse and Exploit. All this time I'd skimmed over Courier without paying it much heed, thinking, "Why would I want to compile a sprite to ferry some data? They have this great thing called e-mail already." Well, turns out that the delivery boy sprite has the holy trinity of Stealth, Exploit, and Browse. Sleuth sprites do as well, but aren't accessible by the bulk of technos. So while I still wouldn't trust a sprite to hack an entire system for me, it seems that it is at least possible to send one to fetch lightly-defended stuff on its own.

well, you also need edit (IMG:style_emoticons/default/wink.gif)