Okay, so in a campaign I'm running we have a character who, for all intents and purposes, is a combat twink. You know, human-looking orc with bone lacing, dermal sheathing, pain editors, blah blah blah. Dumb as a rock, too - Logic 1. Yet he has a rating 6 agent which, for all intents and purposes, is as good as most other hackers out there.

So how do you handle this type of issue, as a GM? Do you care? I'm thinking of having him use a Computer+Logic roll to interact with the agent (with a threshold based on the complexity of the command he's issuing), but that runs the risk of making his multi-thousand nuyen investment effectively worthless (and possibly upsetting the player who will feel gipped). So... what to do, DS?

Is this character ruining anyone else's fun? If it isn't no big deal. If it is ruining other people's fun, then why is it? In practice any character archetype or build can ruin other player's and GM's fun, it's not the Hacker in the Box by itself that can do this.

If you do want to rein things in a bit, number one would have been looking at allowing this to begin with. Sounds too late for that. Next is to realize there are limitations to Agents, make sure to know what they are and use them. e.g. No Edge, lack some skills to do certain tasks. It takes time and effort to manage and order the Agent effectively, while you are doing that you can't do other stuff as well.

I'd definitely enforce the need for the Agent to make a Rating + Response roll to understand any orders given (per pg. 111, Unwired) and give it a higher threshold because the person giving the orders is as dumb as rocks.

Have him write out a script, a sample is given on page 69 of Unwired. And remember: scripts are followed to the exact T.

EDIT: How'd he start with a r6 Agent? And I hope it's an Unrestricted Agent. In that case, it's availablility 19 (IIRC). If it's a restricted (read: normal) agent, he can't hack with it; it's hard-coded not to hack.

Agents, even if they have a Stealth program, are not exactly "subtle" in undertaking their tasks in the Matrix. Agents can be Spoofed to carry out different tasks than the other wanted (or outright stolen).

Make sure that you establish with this character how he is communicating with the Agent. If he is simply giving it commands and not receiving any real-time communication back from the Agent, then that's dangerous. It's equivalent of giving important tasks to an underling and then never calling them back to see how its going. The Agent can blunder into unexpected situations and won't know how to deal with them, they could be captured or killed and the owner wouldn't know.

If the character is receiving feedback and giving a constant stream of orders back, then he is Subscribed to the Agent. This has the added problem that an enemy Hacker could (quite easily) Trace the Agent back to your Logic-inept combat monster. This would put the character, and potentially the whole team, at risk. This would include all that precious cyberware that your combat monster relies on to do their job. If your character argues that he has secured his cyberware blah, blah, then point out to him that he has a Logic 1 and its probably a miracle that he got the stuff to turn on in the morning. This is unless the team has another tech support specialist that has configured his PAN for him.

Overall, remember, that Agents are not AIs or even Semi-Autonomous Knowbots. They can do specific tasks particularly well when ordered to do so. But they do not handle complex tasks or unexpected situations well at all. Don't give your player/character a whole bunch of "my Agent would know that" kind of slack. The Agent has "task" knowledge, not "concept" knowledge. If you tell the Agent to Hack a system and grab some files it will do it. Is the Agent able to figure out that the System was "way too easy" to hack and that its probably walking into a trap? Not in the slightest.

Well, you could always have the the character purchase the Adaptability Autosoft (rating 3) and Virtual Person for his 'Link and should allow this mental giant to use that uber computer in head for more than porn. With the Autosoft and the Virtual person you could have a "NPC" like Dixie Flatline from Neuromancer.

A dp of 12 doesn´t make a hacker. No edge, no hot-SIM bonus, no specialisation, no hacking implants, no 5 IPs. The role of "hacking" is not filled. If your group doesn´t have a hacker, be glad someone cares about the matrix.

The "dumb orc without skill" issue is on a different page - someone is apparently metagaming, and skill is only required to write agent scripts. Since agents are pretty dumb, too, you can easily have things happen that are not reflected in the agents orders. Multiple conditional orders = software test.

See, I knew I could come to you guys and walk away with some great suggestions! I think that, definitely, there should be some software tests involved with this program.

Perhaps I've been playing it too loose with what this agent will return - I suppose I was misinterpreting it as being a "smart" system whereas the majority of you seem to point out it's not. That could make things interesting, not for the things that the agent finds out, but for the things the agent DOESN'T find out.

Also, a good point about it being a registered vs. unregistered agent, and an easily traceable stumbling block. I suppose I should make the character write a script if he wants the agent to erase its tracks as it searches for data.

Good points, all of you!

Generally the problem with dumb user & smart program is that without having any knowledge the player commanding the agent shouldn't be able to use it anywhere near as well as a true hacker.

Say for example the orc wants a door opening and orders the agent to hack in and open the door. Fair enough the agent gets the door open, but because the orc in question doesn't ICly know much about how computer systems work he shouldn't be able to clear up the logs afterwards or take any of the precautions a smart hacker would to avoid being caught.

Personally if the character using the agent doesn't have /any/ skills to back it up I wouldn't even give them the option of making up scripts. Because they're not going to have a clue what a script is let alone how to set one up.

While that might sound a little cruel I'd personally just use it for a plot arc where the agents hacking bites the team in the ass. If only to prompt them to spend the odd bit of karma on basic matrix knowledge before they try play at being a hacker again.

Ah, I forgot about the Access Log, good point! Everything the Agent does will leave an entry in the Access Log of the system. This probably won't make a difference during the course of the run itself, but if your dumb Ork is using the Agent to Hack everything (via brute force) with the Agent, the system analyst examining the logs after the fact should easily be able to pick up on the Agent's Access ID. Then, next time the Ork starts to use the Agent, the corp Security Hacker can pick up on the trail and then you have the "Agent Traced back to the dumb Ork" situation again.

Once a Trace has completed on the Ork's Commlink, the corp will have his Access ID until the Ork decides to change it. Don't let meta-gaming knowledge creep in here! An Ork with a 1 Logic and (possibly) no computer skills wouldn't even think to change his Commlink's Access ID, if he even knows such a thing exists. Think of your techno-inept grandma and ask yourself how long it would take her to think of altering her computer's MAC Address, seriously. So now that said corporation has the Ork's Access ID, they can determine his physical location everywhere he goes. Time for a little GM retribution to teach the character a lesson about being sloppy.

Agents are useful tools, but they are not a replacement for a true Matrix character. Telling an Agent to go hack every obstacle encountered is the Matrix equivalent of sending a Drone with a rocket launcher to blow up every locked door that is encountered. It's going to work, but its going to leave an awful mess behind.

Um. Are we forgetting that a Logic 1 hacker, dumb as bricks, can hack the bejesus out of the matrix because the test is Hacking + Program, and Logic has nothing to do with it?

Or has Unwired fixed that?

Unwired has an optional rule (actually very heavily used from what I can see) where you role Logic + Hacking, with a hits cap equal to the program rating.

Plus the Logic 1 hacker would still have a whole lot of computer skill and there'fore know what he has to do to cover his tracks. Well, actually most likely he wouldn't know why he had to do them, just have "Do this after this, but before this" pounded into his head.

Ahh excellent. Can't wait to get my hands on Unwired. I'm tired of brick-headed script kiddies.

Another variant of that rule (which I am currently using) is to continue to roll Skill + Program but limit the hits to the character's Logic. It achieves the same result without altering the existing rules as much.

I rather the Logic+Skill (Program Cap). It is how the rule is written and it just sits better in my mind. But yea, ether one is much better then the BBB Raw.

Has the ork written a script to order the agent to protect itself? Has he also loaded it up with attack & armor programs? Because Agents aren't sentient. They don't think to protect themselves unless programmed to do so. And if the hacker has said, "hack this place" and the agent is attacked, what will it do? Likely keep hacking, unless the ork has put in some more complicated scripts... Of course, a Registered agent couldn't even protect itself in Matrix combat. So let's say dumb-as-rocks ork has asked the agent to do a data search. I'm a hacker. I see this lovely rating 6 agent running around Shadowsea (or whatever datahaven), seemingly on it's own... hm. Looks like it's my birthday! (as a side note, this is a great way to suck nuyen out of your team as they continue to replace the agent - if you're feeling evil)

Seriously. A rating 6 agent on it's own is fair game to any hacker out there. And if it ever sets off an alert while hacking (if it's unregistered) it'll be toast to the first security spider that comes along. If it does set off an alert (and even with a rating 6 Stealth program, hacking higher-rating nodes on the fly will set off alerts frequently), has he written a script so it knows what to do? Or will it even tell him there's an active alert out there? The team can get seriously messed up if they walk into a place where the agent has set off an alert, and they don't even know if...

What if it comes to something with a data bomb? Or something encrypted? How many programs has the ork bought for his agent? Has he told it to do a Matrix Perception Test before it hacks a node or loads a file? It doesn't have any sense of self-preservation. So even if it does see a data bomb on a file, it'll still follow his orders to get the file, triggering the bomb and... bye-bye agent. Has he told it what to do if it comes up against an encrypted node? Or does he say, "hack this node, get this file" and hope for the best? Because that agent is likely to just go to the node, spin its wheels until recalled by the ork, and good luck for a logic 1 character to figure out why the agent couldn't hack the node (remember, it isn't sentient. It is likely to simply say "unable to comply with request" over and over again, not say "the node was encrypted and I lacked the proper programs to proceed").

Agents are very useful tools. But agents in the hands of someone without any Matrix knowledge (and a logic of 1) are just dangerous. At best he'll just loose the agent. At worst... he'll lead corp security right back to him and his team mates.

Make your Ork take a Logic + Software test to write a decent script for that Agent that covers some of those bases like Tiger Eyes is talking about.

As a professional programmer I can say that writing a script for a program like that is very similar to coding software itself with all of the IF... THEN... ELSE conditions that must be thought of. So calling for a Software test is perfectly fair game.

While I would prefer an Attribute+Skill mechanic for the matrix on general principle, doing so while using (program rating) as a cap on hits is not as minor a change as it may seem. You now need two hacking tests to beat anything beyond threshold 6, and are in trouble in case of rating 3 admin and rating 4 security accounts, should the node run Analyse. (16 dice against threshold 6 (best stealth): 45% chance of success).

Nice reference. I always meant to simulate this back since SR1.

The problem with some of these suggestions is that 1) Unless he's taken Incompetence, he still has *some* skill at computers. Would you make a hermetic mage with Logic 6 make the same roll? If not, that's singling out the player. Also, for 2) By RAW, he's not the one who has to make the Comprehension test, it's the agent. You can say that he can't write a script without Longshotting it (Logic 1, -1 die for defaulting, = 0 DP), but that won't stop him from giving the thing orders, and using its brains to figure them out. As for attacking the agent, the problem here is that 3) An agent is capable of cracking its own copy protection, so you can flood the matrix with copies of it.

Agents have many limitations, but quite honestly, they tend to not lie with the user's Logic score. I do think that if this isn't causing a problem, you should let it go. If the agent is stealing spotlight time from a PC decker, then you have a problem. If you don't have a PC decker, then it's not an issue: just run it like any other matrix-based NPC.

Amen, brother.

Cracking copy protection is still a software test. What skills do agents have again? Can you default on Software tests (for writing a script, in this case)? German RAW suggests otherwise.

And on the comprehension test: Orders have to be straight-forward, or scripted. "Hack that node" would be interpreted as "use Exploit on that node to get (average of last 50 hacks access level)"

heh, read some strips of dilbert, especially those where dilbert has to explain something to the spiky-haired boss.

or for that matter the interaction between sales man stef and the geek squad of user friendly.

the agent may be smart, but its probably autistic style smart. that is, it will understand anything, as long as its explained in the ways of a computer

Dunno about the German RAW, but the English BBB says Pilot in agents stands in for most Computer skills.

That's kind of the problem; I've had this issue stealing time from the hacker in the past. I like the suggestions of making the agent "dumb" in the sense that, if it does break into a place, it won't know to erase it's trail unless specifically told to do so.

I'll handle the metagaming aspect of, "Well I told my agent to delete its tracks" later. I mean, honestly, if his character has a logic of 1 and no software or hacking skill, there should be no way he'd even know to say that in the first place.

Going on to list everything but Software skill?

Most, but not Software, which is what is needed to crack copy protection. BBB p. 214:




Exactly. A Logic 1 character with no skill in Computer, Data Search, or Hacking, is your grandma who's never touch a computer before. Ask them if they know how to edit a server's system logs or how they can alter a device's MAC Address.

Thankfully they spell that out since it seems kind of ridiculous to create a computer program capable of programming by itself. If that were the case, every agent is one mishap away from becoming the next Deus.



That's what I think I'll be going with on this one, along with the aforementioned "autistic" agent routine.

Good calls, ladies and gents.

Depends, is the Orc named Blaster and his comlink named Master? That earns a drink on the house!

Personally I'd probably ask what's up. It's really how the player is looking to handle the character. If it appears he's just min-maxing to try game the system that's going to show up elsewhere anyways, time to head that off directly...and some argument about rules or trying to patch over any potential "loophole" isn't going to address the real root issue longterm.

Otherwise *shrug* game on.

If the player doesn't roleplay a logic of 1 steeply reduce his Karma awards. I calls it selection of the fittest! [roleplayers, that is...]

Now, this may not be entirely germaine to the discussion at hand, but I kind of feel the need to point out that Logic 1 does not automatically mean "dumb as a brick." What it means is "incapable of memorization, quick learning, and technical knowledge." As long as that character has a high intuition, they're probably not dumb, per say - just ignorant. Look at it this way: what use does a street sam who came up as a ganger in the Barrens have for Logic? Almost none, honestly. Not only does logic not help them in their daily life - or not as much as say, being tough and thinking quickly - but they're not likely to have developed any skills that are even keyed to logic unless they're into cars or something.

I actually made an NPC fairly recently who was an ork with a logic of one, but an intuition of 5, because he was precisely that sort of street sam - a guy who plays from the gut, thinks on his feet, and lives by instinct in that sort of 'zen killer' sort of way, just to prove that it wouldn't be as detrimental as one might think to not have a lot of 'book schooling' on a hired gun. A character with low logic and high intition would be the sort of person who doesn't know a lot of technical facts, would have problems understanding complex theories, and who couldn't grasp WHY something works... but would also know how to think quickly on their feet, adapt rapidly to changing situations, and would be able to understand how to use something to accomplish a goal, even if they didn't understand how it works.

Now, specifically in this case? Yeah, I might have the player roll Logic+Computer to give the agent a particularly complex order or set of orders, especially if there were contingencies and whatnot, just because it might be difficult for them to know exactly how the agent would interpret the orders. Malachi and Oenone have pretty good input on the sorts of problems a low-Logic character would run into with computers: they wouldn't think of a lot of the more intricate hacker tricks, not because they're dumb, but because they just wouldn't know you could do them. Once they learned about the concept of "always spoof your datatrail" ICly, though, I wouldn't have a problem with them telling the agent to do so. But I sure as hell wouldn't give them any hints to do it beforehand.

I know people with a very high Logic who can't do any of those. In 2070, however, this sort of thing is common knowledge. After all, unless this ork is Incompetent, he has an actual skill rating. It might be zero, but skill zero is not the same thing as "completely oblivious". It means he's got enough computer skill to function in the Sixth World-- he can run basic searches, he can set his TiVO, and he can order an agent about.

The only question here is how well he can give orders to that agent. Since he's becoming a problem, having the agent take everything literally is probably the best. If he wants to give it complicated commands, he'll either need to be right next to it in the matrix, or scripting it. To be fair, you have to make the decker do this as well; but assuming the decker has a decent skill and attribute (e.g., not a script kiddie) he should be able to do fairly well. The scripting rules, plus frequent Comprehension tests, should keep the agent from becoming a problem.

Additionally, someone pointed out earlier that he couldn't get a Rating 6 agent at chargen, and definitely not an Unrestricted one, which he'd need in order to hack systems. You might need to do a retroactive audit on his sheet, and dial back on some illegal gear.

Actually Cain, you CAN get an Unrestricted Agent at Chargen by taking the Restricted Gear Quality. A Unrestricted Rating 6 Agent has an availability of 20... just barely within the Quality.

True, but given that this is a secondary trick for his character, I highly doubt he'd blow an Advantage on it, when there's lots of other things a combat character would be focused on. At any event, this still calls for a sheet audit; it's causing a problem, and it'd be easy to see if it was bought according to RAW.

I personally hate slapping the restricted/unrestricted agent thing down on a player, particularly one who didn't have Unwired when he created his character. It seems unfair that an agent who was bought to hack suddenly can't do it, just because a new book has been brought to the table. But in this case, it seems called-for.