 Agent Smith Army problem, How to keep Shadowrun street level |
  |
 Jan 12 2008, 04:41 PM
Post
#51
|
|
|
Cybernetic Blood Mage  Group: Members Posts: 3,472 Joined: 11-March 06 From: Northeastern Wyoming Member No.: 8,361  |
Well, I won't pretend to have put as much thought into the subject as FrankTrollman has, but why couldn't you bandage the problem as follows?
( 1 ) Give AR Modifiers out like candy. The reason that nearly everyone stays connected is that doing so makes their daily life easier. (Doesn't address the idea of corps keeping their networks offline, but I don't see corps forcing Runners to go on-site to do anything as a bad thing.) ( 2 ) Agents and IC use the resources of the node that they are currently present in, however there is something special about a Decker's personia that allows an Agent to break this rule. (Or simply get rid of the exception where an Agent can hack from a Decker's commlink.) ( 2A ) Anytime a Node's Rating is degraded it goes on alert. So multiple Agents are really only good for DoS Attacks, and DoS Attacks are a valid tactic in Matrix Warfare, at least until the corp in question is able to track you down. ( 2B ) In order to work together on anything other than a DoS Attack, you have to use Agents compiled from different source codes, copies simply won't work because they aren't different enough. ( 3 ) Hacking in 2070 simply can not be done without DNI of some kind, a Decker simply can not input commands fast enough to matter, so even while hacking via AR, BlackHammer can and will still kill you. |
 |
 
|
|
Jan 12 2008, 05:05 PM
Post
#52
|
|
|
Runner Group: Members Posts: 3,009 Joined: 25-September 06 From: Paris, France Member No.: 9,466 |
Most of these are in my system. ;)
Still, your solution leaves a few issues intact, such as hackastack (using a whole lot of commlinks and switching from one to the next to clear your condition monitor and for other purposes). And Frank would tell you that (2) doesn't prevent a drone from using the commlink so that the drone would have a persona. |
|
|
|
|
Jan 12 2008, 05:19 PM
Post
#53
|
|
|
Cybernetic Blood Mage Group: Members Posts: 3,472 Joined: 11-March 06 From: Northeastern Wyoming Member No.: 8,361 |
I wouldn't allow an Agent to use a commlink like that, so I personally don't see why a Pilot Program should be allowed too either.
As for hack-a-stack, a guess I never really understood the threat, if you try to log onto a Node with your stack of commlinks all at once it seems to me that you are likely to trigger an alert, and if you don't then every time you switch to another commlink you have to relog onto the Node you were in. Once you add the threat of BlackHammer even in AR Mode unless I've missed something I just don't see where hack-a-stack would be worth the risk, because with my bandage, you would be at risk of getting your brain fried through every commlink in your stack. |
|
|
|
|
Jan 12 2008, 05:57 PM
Post
#54
|
|
|
Runner Group: Members Posts: 3,009 Joined: 25-September 06 From: Paris, France Member No.: 9,466 |
Exactly. That's more or less the way it works in my system. :)
|
|
|
|
|
Jan 12 2008, 06:01 PM
Post
#55
|
|
|
Cybernetic Blood Mage Group: Members Posts: 3,472 Joined: 11-March 06 From: Northeastern Wyoming Member No.: 8,361 |
Ok, you win, I'll actually read your system and then come back to comment, although I'll give you fair warning, I place great trust in FrankTrollman's ability to weed out bad rules. :cyber:
|
|
|
|
|
Jan 12 2008, 07:44 PM
Post
#56
|
|
|
Cybernetic Blood Mage Group: Members Posts: 3,472 Joined: 11-March 06 From: Northeastern Wyoming Member No.: 8,361 |
*shakes head*
Umm, sorry Blade, but I disagree that my proposed off-the-cuff bandage is anything like your system, for starters I still can't wrap my mind around the idea that a corp node could gangrape an intruding Decker dead in just a few seconds anytime but doesn't. Plus I don't like the idea that the only thing really preventing Agent Smith Armies is that the knowledge to create an independantly acting Agent is being suppressed by the corps, if nothing else then I'd imagine that the AIs running around since Emergence would let the cat out of the bag rather throughly. |
|
|
|
|
Jan 12 2008, 10:24 PM
Post
#57
|
|||||
|
Runner Group: Members Posts: 3,009 Joined: 25-September 06 From: Paris, France Member No.: 9,466 |
Well, I've never said that your "bandage" was like my system, I just said that what you proposed was in my system, but my system isn't only made of these. ;)
I don't see what you're referring to. Can you explain, please?
Actually I haven't been through Emergence yet and it only seemed very likely to me that corps would do whatever they could to prevent Agent Smith. I couldn't find any technical reason that'd prevent independant agents if there could be non independant agents. At least, no technical reason that didn't lead to something I didn't like. So I decided that it was technically possible. But as allowing this would be far more damaging than rewarding to corps, it seemed likely to me that corps would do everything they can to prevent that, especially after the crash 2.0 virus. The best way to do so would simply be to benefit from the fact that they rewrote the Matrix from scratch to include security against these (2070 Captchas) in all possible layers. You're free to find what's special about a hacker's persona that an agent won't be able to emulate, but as it should allow for non-DNI use, I'm afraid it'll be a little bit complicated to find a reasonable technical reason. (That's what Frank was referring to with his drone using a commlink). Well, of course you can still skip the reason if it's an acceptable break from reality for you. It wasn't for me. |
||||
|
|
|
||||
|
Jan 12 2008, 11:33 PM
Post
#58
|
|
|
Runner Group: Members Posts: 3,382 Joined: 22-February 06 From: Shadowland Member No.: 8,297 |
Wow! An amazing amount of warm, yellow liquid on the floor ... <sniff> <sniff> ... Oh, I get it. There was a massive pissing match here I just missed.
;) |
|
|
|
|
Jan 13 2008, 12:22 AM
Post
#59
|
|||
|
Midnight Toker Group: Members Posts: 7,686 Joined: 4-July 04 From: Zombie Drop Bear Santa's Workshop Member No.: 6,456 |
Logging in all at once isn't such a problem as is logging in serially. Given a sufficient number of comlinks any character can continue cybercombat indefinitely. |
||
|
|
|
||
|
Jan 13 2008, 02:08 AM
Post
#60
|
|||||
|
Moving Target Group: Members Posts: 698 Joined: 26-October 06 From: Iowa, United States Member No.: 9,720 |
Ok, this Matrix causes large disputes... I've been been reading through and trying to come up with some theories, strictly from canon. Interpretations are different for everyone granted but here are some things that might not be coming together properly... Granted even the writers themselves didn't get everything together as we've all figured out.
First Item, looks like Agents must be on a node in order to work it, they can only exist in one node. Your persona can subscribe to multiple nodes, but agents are either with you, or in a node seperate of your persona. Taking this in, Agents reduce, which reduce response, which Pilot, which (No Rule, just theory) the agent would drop to 0, which should crash it "Your system does not meet the requirements to run this program.". So you could probably put crazy drain on a system, but at that point it would probably start dropping connections to improve performance "All server connections are busy, please try again later". Second Item, if the agents are with you, you are limited to (Response) number of agents, if the agents are independent (spread throughout the matrix), you are limited to your subscription list (System x 2). Agents look to always count towards your subscription list unless they are shut down completely. Though running the full System x 2, you would be issuing commands blindly since you couldn't be logged into a node yourself. There's my interpretations. Are they word for word in the book.. No, just my interpretation. But neither is how a ganger responds to getting shot at. This isn't a "Choose Your Own Adventure" book, its a guide. Whose houserule is best? Depends on whose house it is, and who's GMing (and what mood he's in). 8) |
||||
|
|
|
||||
|
Jan 13 2008, 02:25 AM
Post
#61
|
|
|
Moving Target Group: Members Posts: 698 Joined: 26-October 06 From: Iowa, United States Member No.: 9,720 |
Ugh, hackastack... More reading to try and theorize this...
To start with the hackastack is doing everything physically so the hackastack would be Wired not Hot-Simed otherwise its the independent agents as discussed above. (So I suggest a high AGI for commlink juggling) So options are 1. Lots of pairs of Image Linked Glasses using Simple action to draw your glasses (Quck Draw test?) and you still have to change your AR Gloves/Trode Net... Yeah this is ridiculousness. Smile it's Friday... 2. AR displays using two free actions, first to change your Image Link, then to change your AR Gloves/Trode net. (Change Linked Device Mode pg 135) Leaves you with a Simple Action to issue a command to the Agents on that commlink. As a group only. 3. Touch screen interface. Ugh how archaic, using your stylus or the built in keypad... So here he has a simple action to draw the commlink, and a simple action to issue the command. Assuming you're not going to penalize for trying to num pad or stylus enter commands. So 4 passes, each pass sending out whatever Agents are subscribed to that commlink (System x 2 per commlink). Hackastack very amusing concept for the juggling hacker. I suggest Adept thrower, so when your commlinks crash you can chuck them at people as well. "Thats right, brain hacking! The hard way!" (I'm actually a fan of brain hacking but the pun was needed.) [EDIT] I didn't want to add yet another reply in a row... So agents vs hacker skills. I think it is broken, and here's why. You need to get targeting, maneuver, clearsoft, electronic warfare.... But they all come with max Computer, Cybercombat, Data Search, and Hacking? Why aren't these autosofts as well? What really ticks me off... Sprites don't even come with all these knowledges and their supposed to be even smarter than agents. Autosofts should just be activesofts and none come included. |
|
|
|
|
Jan 13 2008, 02:48 AM
Post
#62
|
|||||
|
Cybernetic Blood Mage Group: Members Posts: 3,472 Joined: 11-March 06 From: Northeastern Wyoming Member No.: 8,361 |
Unless I'm completely misunderstanding what you are saying this would allow the corps to gangrape any intruding Decker dead in a few seconds the moment he is detected by hoards of IC. As for what is special about a metahuman persona that allows Agents to "ride" the signal, well I'm perfectly fine with having that special something being DNI, after all, I have no problem with requiring Deckers to use DNI for any meaningful hacking anyways and it also allows me to put the bite back into AR Hacking which in turn should help limit the usefulness of hack-a-stack. |
||||
|
|
|
||||
|
Jan 13 2008, 03:01 AM
Post
#63
|
|||||||
|
Midnight Toker Group: Members Posts: 7,686 Joined: 4-July 04 From: Zombie Drop Bear Santa's Workshop Member No.: 6,456 |
Actually, with a proper comlink chain it would be {(System)*2 + (n-1)[(system-1)*2]} agents with n being the number of comlinks in the chain, which is unbounded. A simpler hub and spoke setup would naturally max out at 4(system^2) but could be converted into an unbounded branching tree. Either way, the only limit to the number of subscribed devices is money, since you can always add new systems to the chain. |
||||||
|
|
|
||||||
|
Jan 13 2008, 04:03 AM
Post
#64
|
|||
|
Moving Target Group: Members Posts: 698 Joined: 26-October 06 From: Iowa, United States Member No.: 9,720 |
I'm not sure if I'm interpreting this correctly, but I think you're saying Agents can have subscription lists? Basically you send a command to your subscription list of commlinks, and each of those commlinks would have an agent with a subscription list and he would send the command to his subscription list. Since persona's maintain subscription lists, so there would have to be some mind there agent or person to send the command to their subscription list. My interpretation before was Agents could only work in one node or in your own persona, so I think that would rule out a subscription list of any type. Again my interpretation, taking this out and you would be quite right. |
||
|
|
|
||
|
Jan 13 2008, 10:34 AM
Post
#65
|
|||
|
Prime Runner Group: Banned Posts: 3,732 Joined: 1-September 05 From: Prague, Czech Republic Member No.: 7,665 |
Sure can. Every Commlink in the Hackastack has a unique persona and subscription list whether there is currently any person pressing buttons on it or not. It has a subscription list and loaded programs even while not having commands sent to it. If one of those programs happens to be an Agent, then your Hackastack just became Agent Smith. -Frank |
||
|
|
|
||
|
Jan 13 2008, 11:17 AM
Post
#66
|
|||
|
Runner Group: Members Posts: 3,009 Joined: 25-September 06 From: Paris, France Member No.: 9,466 |
Actually, I like to compare a hacker hacking a Matrix systems to a face infiltrating a corp building. If the intruder is detected, you won't send all your security guards to catch him, because you have to keep an eye on the rest of the building. Sure, you might send some extra guys to take care of him, but you don't want to send everyone, except in small places or places with few guards. Well, they can do it if they want but in that case, the best way to attack a system would be to start a diversion attack somewhere in the node while you hack. |
||
|
|
|
||
|
Jan 13 2008, 04:51 PM
Post
#67
|
|
|
Cybernetic Blood Mage Group: Members Posts: 3,472 Joined: 11-March 06 From: Northeastern Wyoming Member No.: 8,361 |
Two things come to mind, one, in the meat world your security can't just teleport over, splatter some poor sod and then teleport back to whatever they were doing a few seconds before, and two, if a corp building could have as many sec-guards as your system allows IC, then there is no reason not to send them out in hoards, in fact it's a far safer tactic as it would be virtually impossible for a Decker to sleaze past the hoards of IC set up to scan "their" portion of connection slots.
|
|
|
|
|
Jan 13 2008, 06:27 PM
Post
#68
|
|||||
|
Moving Target Group: Members Posts: 715 Joined: 4-September 05 From: Metaplane GEPLK136 (The one with the lizards. You remember the lizards, don't you?) Member No.: 7,684 |
10 agents could be distinguished from 10 hackers by a matrix perception test. This implies that there is a fundamental difference between the two. Based on the idea that they are in fact distinguishable, I'd propose the following rules:
The above rules also apply to IC based on the same code as well. This would seem to solve (or at least improve) the Hackastack (and by extension) Agent Smith problem. It also allows users to try and exploit multiple agents as well if they feel the drawbacks are worth it -- and for cheap corps to try the tactics as well against unsuspecting or unperceptive hackers. (Running four copies of your identical Blaster IC will keep out script kiddies, but not experienced hackers with these rules.) I like this better than either the "ruled out by fluff" or "player-GM MAD standoff" approach. AR Hackastacks will be very dangerous options -- a large number of poorly-rolling commlinks controlled by a hacker (not an agent) will make you a physical sitting duck, as IC will be able to track you very quickly unless you use a large number of actions to shut down your commlinks in a hurry. Sure, you *might* get past nasty IC with that many commlinks. If you've got wired reflexes, you *might* even do it before they shut down remote access to the node. You won't do it before you're seen though. Having a room full of commlinks with identical agents will not be very helpful. Against a single (perceptive) target, they will do little damage. Against several targets, they may do a good deal of damage (at the expense of being more easily tracked), but they will be shut down fast by perceptive attackers. A very rich (or very prolific coder) could purchase or write a large number of different agents, and use them with impunity. This is fine, actually -- because it is not unlimited (time and resource limits apply) and is no different than spending the equivalent resources to hire multiple hackers. A very skilled hacker could hack into multiple nodes (high rating nodes, or he risks his agent being degraded!) and avoid the track penalties by running his identical agent army on a number of different nodes. His army of agents will still be vulnerable, though. And, he's limited by the number of nodes he can hack and remain undetected in. This again seems fine. The system also gives a very slight penalty to running even a single agent. That seems fine to me as well. This creates a question for fluff -- why do Agents running on different hardware create identical personas, while a metahuman using the System-supplied persona running on different hardware creates different personas. I'm sure it could be explained away with some technobabble about quantum computing and interactions with the human brain. One thing it doesn't solve is a room full of anthropomorphic drones using AR commlinks (well, the track penalties would still apply.) Really, though, if you allow drone Pilot programs to run commlinks in this manner, you're probably already setting yourself up for a world of hurt. I'm sure we could think of enough exploits for Pilot that advanced to rename this game AI: The Awakening. |
||||
|
|
|
||||
|
Jan 13 2008, 07:30 PM
Post
#69
|
|
|
Great Dragon Group: Members Posts: 5,537 Joined: 27-August 06 From: Albuquerque NM Member No.: 9,234 |
Even if you do have to buy them, agents are not that expensive. And compared to a 225,000 :nuyen:/year 24/7/365 security hacker, they are darn cheap. And I think that 15 rating 6 agents with rating 6 SW patrolling are likely to be more effective that 1 hacker with rating 6 SW patrolling.
|
|
|
|
|
Jan 14 2008, 08:48 PM
Post
#70
|
|||||||
|
Shooting Target Group: Members Posts: 1,512 Joined: 26-February 02 Member No.: 392 |
Few things to point out to pay Devil's Advocate here: 1) The RAW states that you can issue the same command to multiple drones/agents at the same time. So you can have System*2*a bazillion agents working for you at once. Basically a rating 3 commlink could have 3 different hordes of agents doing 3 different tasks all at once. 2) Proponents of hackastack seem to take the RAW definition (from the sidebar on Matrix Terminology) of Node as any Drone, Commlink, Device or Network. The important one to note there is Network. I remember getting into this debate with Seven-7 and the interpretation I was told is that since a Node = Commlink = Network then you can have Multiple Commlinks = Network. Thus it only takes a Free Action to switch to your new Commlink with its fresh Condition Monitor. And since your Commlink = Network = Node then you are still in the same session. Now I don't allow the Agent Smith or the Hackastack in my games. Here's why and how: 1) First off Agent Smith is going to get you caught. If you have an arbitrarily large number of Agents all trying to brute force a node then the same laws of statistics that ensures you get in also ensures that you will get a critical glitch. Edge can't negate this when it is an agent acting on your behalf. This means that you will be tracked by security and in SR4 it is REALLY easy for corps to set up covert surveillance until the strike team gets there. 2) The way I interpret the Agent rules for agents working independently limits their effectiveness. Even if you hack into an arbitrarily large number of rating 6 nodes and plant 1 agent per node they still operate at Response = to system on which they are operating (target system). As a result most corporate nodes in my game are 3/3/6 (firewall is cheap). This still allows most worker bees to get their job done without slowing them down while also restricting what Agents can do. All of a sudden they can only use 2 programs at a time without going down to Response 2 and System 2. And they are only rolling a max of 6 dice. 3) All security systems are rigged so that Matrix Alert = Physical Alert = Magical Alert. Congratulations your hacker just got the rest of the team killed (if infiltrating the facility as you are in most runs). 4) Networked (ie actively subscribed) Commlinks =/ Network =/ Node. Sorry but you have to do it the long way. Hack in. Log out. Switch devices. Log in. Lather, rinse, and repeat. A Node = Network for situations where you are considering a Corporate Node. The Node is not just one device. It is numerous devices throughout the facility that are set up (through system architecture) to act as 1 Node. |
||||||
|
|
|
||||||
|
Jan 14 2008, 11:00 PM
Post
#71
|
|
|
Great Dragon Group: Members Posts: 5,537 Joined: 27-August 06 From: Albuquerque NM Member No.: 9,234 |
The counterpoint is the guy setting off the burglar alarm four times a night for a week. At what point do you stop having the huge reaction and start assuming that something is broken? And if you are spending all your effort reacting to X, how much can you direct against threat y?
And you can program this stuff a long time in advance and scrub the data trail. |
|
|
|
|
Jan 15 2008, 05:11 AM
Post
#72
|
|
|
Shooting Target Group: Members Posts: 1,512 Joined: 26-February 02 Member No.: 392 |
To counter your argument how hard is it really for SR4 security forces to be on alert? AR is pretty darn flexible. Maybe you are right and it comes down to corp security ALWAYS being in full combat armor and packing heavy weaponry.
|
|
|
|
|
Jan 15 2008, 02:34 PM
Post
#73
|
|||
|
Technomancer Group: Retired Admins Posts: 4,638 Joined: 2-October 02 From: Champaign, IL Member No.: 3,374 |
No, but it might come down to a small group of five to ten individuals "on-call" and prepared to engage with a target while the rest of the forces suit up. It's usually not numbers that ruin the team's day in my games but reinforcements..... |
||
|
|
|
||
|
Jan 15 2008, 03:59 PM
Post
#74
|
|||||||||||
|
Prime Runner Group: Banned Posts: 3,732 Joined: 1-September 05 From: Prague, Czech Republic Member No.: 7,665 |
1) This is an established methods of incentivising people to have a connection to the Matrix. But so long as you can segregate networks one from another you can put your important stuff off Matrix. And the corps can and will do this as well. One computer for incoming information and another for storing important data and running your vital systems. Your cyberarm will have all the files that run your cyberware and it will have the pay data in it, and you'll have a pan that constantly gives you positional information that is subject to hacking, but they won't directly conect at all. In this model you've upped everyone's dice pools across the board. And you've given people the ability to be hacked. But the hacker is just eliminating people's dicepool bonuses, because the actual good stuff is still unhackable as long as that is physically possible.
Sure. But I can just have more than one persona so long as it is functioning in AR and I have no neural connection to it. And as long as each of my personas can run agents, Agent Smith exists. I don't really know why anyone thinks that Agents counting as being run on the hardware they are attacking makes anything any better. The important part, the part where they hack themselves a connection in the first place, doesn't work that way and can't. If you have even one agent in a node then it can spend an action to give you access and boogy out. Or not, heck if piling in more Agents degrades the target host it's basically just an "attacker wins" scenario anyway because no matter how nova hot the target server is or how black the IC you can always drop it down to your level by adding more Agent Smiths.
Yeah. I'm not feeling you there. It just means that any hacker can automatically crash any node whenever they want and no IC can theoretically be tough enough to matter. All agents and IC on both sides will automatically be the same rating and the attacker will always be fielding more.
The costs involved in an Agent Smith assault are so small that making them twice as much is meaningless outside the realm of starting characters. It may make a difference to a particular campaign, but the abstract effect to the campaign world is essentially non-existant.
AR is without a direct neural interface. If you make computers mandatorily connected to human brains then you've eliminated Drop-Out. That is, since you can't have no connection to your brain, people will hack you and the hacker always has something to do. But that's my system. It's an extremely massive overhaul to the way absolutely everything works and it doesn't even play nicely with the other stuff you have here. If you go stick approach (where hackers can hack your stuff and you need firewalls to try to stop them rather than allowing people to be unconnected and then entice them into sticking their vulnerables into the nets where they can be crushed), then you don't need or want any of the other things you have written down here. It's not simple. It's not clean. It completely changes everything. But it does work. -Frank |
||||||||||
|
|
|
||||||||||
|
Jan 15 2008, 04:35 PM
Post
#75
|
|
|
Awakened Asset Group: Members Posts: 4,464 Joined: 9-April 05 From: AGS, North German League Member No.: 7,309 |
Network seggregation of critical systems can not be avoided. It is a viable tactic, and good for running as it requires the team to go on-site.
Not having important data on the matrix is also doable for private persons, always was. Anything with a Network-on/off switch will work fine. But things can be wireless without being connected to the matrix. A corp that pays for wire gains some better protection from outside intrusion, but can´t save on other security measures because the hacker might gain access to their wire. If wireless security is deemed sufficent, it will be used. Drones and personell need a wireless setup. If you encounter a purely wired system (no drones, no networked security), little will stop you from going onsite and installing wireless capability. The hacker will have to just tag along for a short while. If he can´t fight, combat will be a problem. He got 400 BP like everyone, his choice. Not doing combat is very valid IMO. Conclusion: Opt-out is no problem, so you don´t need a fix. |
|
|
|
|
|
Lo-Fi Version | Time is now: 31st July 2025 - 07:20 AM |
Topps, Inc has sole ownership of the names, logo, artwork, marks, photographs, sounds, audio, video and/or any proprietary material used in connection with the game Shadowrun. Topps, Inc has granted permission to the Dumpshock Forums to use such names, logos, artwork, marks and/or any proprietary materials for promotional and informational purposes on its website but does not endorse, and is not affiliated with the Dumpshock Forums in any official capacity whatsoever.