Printable Version of Topic

Posted by: Bearclaw Jun 1 2011, 10:48 PM

OK, so my team, including myself, has all of our commlinks slaved to my external commlink running battle-tac.
The street sam has skinlink on and all cyberware set to DNI only. Meaning that his cyberware only accepts commands from his commlink, and his commlink only accepts commands from my commlink.
The only way he can be messed with wirelessly is to hack the master commlink, right?

As in, spoof doesn't work against his gear, because it's secured, and you can't get to his commlink without going through my commlink. Sounds too simple.

Posted by: Ghost_in_the_System Jun 1 2011, 10:54 PM

A only accepts commands from B, and B only accepts commands from C. If you spoof to look like C, then B accepts commands because it thinks the spoof is C. Also, chaining the slaving is pointless, because if a hacker gets control of C or spoofs C, it commands A just as easily as it commands B.

Posted by: Makki Jun 1 2011, 11:08 PM

slaved devices are not unhackable or unspoofable afaik. slaving just gives a negative dice modifier

Posted by: Ghost_in_the_System Jun 1 2011, 11:14 PM

They are unhackable actually. They have to be spoofed. Well, unless you can get a direct physical line to the device in question, and even then it throws on a serious penalty.

Posted by: Bearclaw Jun 1 2011, 11:15 PM

I was under the impression that you couldn't spoof commands to commlinks. It doesn't have a pilot rating. Is it considered a peripheral device?

Posted by: Ghost_in_the_System Jun 1 2011, 11:21 PM

Basically every device is a node, yes. A commlink is a node, a cyberarm is a node, a fridge is a node.

Posted by: Bearclaw Jun 1 2011, 11:22 PM

Sorry, I edited before you answered

Posted by: Fortinbras Jun 1 2011, 11:25 PM

One can Spoof commands to his commlink if they have your Access ID, something they can get with a Detect Wireless Signal check.
The downside here is that you can't spoof your Access ID because, if you do, the other commlinks won't be able to do anything. They do all their stuff through your commlink and, if they no longer recognize your commlink(because it's running a different Access ID) they aren't receiving or giving any commands to anything they recognize.
The other drawback is that your subscription limit is severely clogged up with all the team's commlinks.
It also means that if someone gets into your commlink, they have access to everybody on the team and don't have to hack one person at a time.

As with anything, when you increase security, you decrees access. Same goes for physical and Astral security.

Posted by: Makki Jun 1 2011, 11:25 PM

everything's a node. and Unwired p55 says the slave is a node and can be spoofed. and there isn't even the modifier I believed there was.
So if B is slaved to A, I can spoof A's ID. If B is a commlink I can tell it to shut down/restart/play AR-Porn/etc

Posted by: Laughing One Jun 1 2011, 11:25 PM

Unwired, pg55 under "Slaving".

And one more thing: DNI is a device-to-brain interface. The samurai dont need to connect his cyber to any commlink.

Posted by: Ghost_in_the_System Jun 1 2011, 11:28 PM

Posted by: Bearclaw Jun 1 2011, 11:30 PM

My impression is, to spoof a command, you have to be sending the command to something that accepts commands. The pilot in a drone, or the equivalent in a coffee maker or or taxi. If there isn't a pilot type program running, you can't tell it what to do, so you can't send it a spoofed command.

Example: You can't spoof a command to a gun. There's nothing there to accept the command. You can spoof a command to a smartlink, it's got just enough of a brain to accept commands like, full-auto or eject clip.

I don't believe a commlink has anything built in to accept and act on commands, but I'd be happy if some one showed me where I'm wrong.

Posted by: Ghost_in_the_System Jun 1 2011, 11:32 PM

Everything in a commlink is built to accept and act on commands. Shut down, load attack program, give admin privilage to access ID 4372469, change master to ID 4372469, etc.

Posted by: Bearclaw Jun 1 2011, 11:33 PM

From pg. 236 BBB

Posted by: Fortinbras Jun 1 2011, 11:39 PM

You are turning the commlink to a peripheral device by making it slaved.

Posted by: Ghost_in_the_System Jun 1 2011, 11:44 PM

Well, 'Peripheral device' isn't defined anywhere in the book, other than mentioning a couple of particular examples. I take 'peripheral device' to mean 'anything capable of accepting and acting on electronic commands', which would include commlinks along with everything on your list. I think the reason that it mentions pilot/agent in particular is that they aren't devices, they're computer constructs, and so wouldn't be affected by spoof if it only said devices.

However, regardless of trying to define peripheral device, as myself and three other people have quoted, the rules say very directly that spoof is an acceptable means of defeating slaving. This could mean that a commlink is a peripheral device (Which I believe to be the case) or it could mean that a slaved commlink is a particular exception of some kind.

FYI I define peripheral device as any device that is connected (or could be connected) to another device in some way.

Posted by: Bearclaw Jun 1 2011, 11:48 PM

So....having cyberware is stupid?
If any jackass with a rating 3 commlink and an off the rack agent can completely shut you down, what's the point?

Posted by: Fortinbras Jun 1 2011, 11:54 PM

You need to check out Unwired. It gives a great many detail on things like this.
Much of your cyberware can be made without wireless.
Spoofing also means that you are only giving one command per action, as opposed to hacking, in which case you have complete access to the device.
You also have to be within mutual signal range of the device you are spoofing. Most cyberware, when it is wireless, has a Signal range of 0.

That being said, having cyberware makes you vulnerable to hackers. In this way hackers are not toothless in combat.

Posted by: Ghost_in_the_System Jun 2 2011, 12:08 AM

Mutual signal range includes any and all devices that you can bounce the connection off of, it doesn't have to be a direct mutual signal range.

It isn't bad to have cyberware though, all you have to do is turn off the wireless when you don't actively need its wireless to be on (such as running a diagnoses every so often).

Also, while you can only spoof one command at a time, your first command can be 'accept me as the new master' and then you have full admin access. So slaving a device can be potentially dangerous.

Edit: Well, what I said about mutual signal range isn't quite right, but there are -very- few applications in which routing can't accomplish the same thing as mutual signal range, and spoofing access to someone's cyberware isn't one of those exceptions.

Posted by: James McMurray Jun 2 2011, 12:22 AM

It's actually defined pretty clearly. A commlink is most definitely not a peripheral device.



A commlink is a device intended for matrix interfacing and processing, ergo it ain't one.

Posted by: KarmaInferno Jun 2 2011, 12:41 AM

Basically the only really effective way to not be hacked is not have data connections to outside your PAN.



-k

Posted by: Fortinbras Jun 2 2011, 12:41 AM

Only if the Access ID you are spoofing has admin privileges. Even then "Accept me as the new master" is a bit of a misnomer.
The first thing you would need to do is delete the enslavement. Then you would need to hack in and create a legit account for yourself, probably an admin account. Once you have a legit admin account you can slave the device to yourself.
This cannot be done in one command.

You can't just give the command "Be enslaved to X Access ID" because X Access ID doesn't have an account.

Posted by: Fortinbras Jun 2 2011, 12:43 AM

Unwired more clearly defines what you can and cannot Spoof that overrides the BBB.

Posted by: James McMurray Jun 2 2011, 12:45 AM

Cool. I was responding specifically to the claim that peripheral devices aren't defined (hence my only quoting that portion).

Posted by: Fortinbras Jun 2 2011, 12:50 AM

Yeah, Peripheral Nodes are clearly defined on p. 48 of Unwired.

Posted by: suoq Jun 2 2011, 12:56 AM

I don't understand.
1) hack in directly requires a physical (wired) connection to the device. If this is the option available to by the "jackass" you're in enough trouble already.
2) hack the master node. If your master node can be hacked with a rating 3 commlink and an off the shelf agent, you need to quick spending money on your hairstylist and upgrade your gear.
3) spoof the access ID of the master node and then spoof commands to the slave. - Assuming your got the access ID, you're facing a Device Rating x 2 check at -6 dice. (because I cannot, for the life of me, think of a single reason why the admin bonus would not apply here). Again, the rating 3 commlink and off-the-rack are not going to be successful here)

I will agree that of the 3, #3 is by far the problem child. Spoofing was written for a power level in line with SRA4 and with every expansion the power creep has upset the balance of power even more.

Posted by: Fortinbras Jun 2 2011, 01:19 AM

If you choose not to have your broadcasting Access ID have Admin Access to aforementioned cyberware, then those Spoofing that Access ID can't delete your enslavement.
Not the best reason, but a reason.

Posted by: baronspam Jun 2 2011, 03:03 AM

Because of the vulnerability to spoofing I think the strategy outlined in the first post is a poor one. Don't slave everyone's comlink to the hacker. The hacker should have an access code in case any matrix shenanigans start to go down and he has to jump in there and trash someone, but I think the best overall strategy is for EVEYRONE to always run Firewall 6 on their comlinks, to have the very best response and system they can possibly afford, have the best Analyse program you can afford and run it 24/7 to check for attempts to break in, and add a Steath program for when you are on a run and need to go into hidden mode. With firewall 6 and a high end stealth program its going to be hard for the badguys to even find your comlinks in the first place if you are running in hidden. Getting into a well secured comlink is not as easy as it sounds, and once in you are in you are subject to connection termination or an old fashioned system restart. Plus, if they never tried to hack you how is your own hacker supposed to feel like a hero?

In general, don't connect cyberware to your pan. Once in a while you should to get firmware updates and such, but there really isn't a reason to have cyberware subscribed mid run. You have a DNI to all your own cyberware, turn the bloody wireles off when not in use. The main exception is cybereyes for smartgun systems, but even then just skinlink the system and skip the wireless, or go old school and run a cable through a harness under your armor and plug into your datajack.

If you really, really, really are paranoid about your comlink getting hacked, don't forget you can run your own IC on it as well.

Another trick, do all the above to your main comlink, keep it in hidden, and run a second comlink with moderate defences with a fake SIN and no useful data when you need to have a comlink non-hidden in high security zones or social situations that call for it. The hacker busts into the active comlink, finds a empty address book, a bank acout with 50 nuyen, and nothing subscribed. By this time the wireless on your main link is deactivated, and the group's hackers is joining the party to earn his keep by making the intruder's ears bleed.

Posted by: suoq Jun 2 2011, 04:04 AM

Are you saying that if the broadcasting access ID has some access OTHER than admin access to the cyberware it's somehow creates a more favorable situation for the target? Could you explain please?

Posted by: Fortinbras Jun 2 2011, 04:12 AM

Nevermind. Once you slave something, the master has admin and only admin access. Unwired p.55.

Posted by: suoq Jun 2 2011, 04:15 AM

To do this you either have to:
1) Not run a tacnet.
or
2) Run a decentralized tacnet instead of a Centralized Tacnet (PG 125 Unwired). To do this, you must
a. Convince everyone on the team to invest in a Tacsoft
b. Make sure everyone's commlinks are as solid as the hackers.

This is a non-trivial investment. If your team has the it's worth it, but teams with limited funds are going to be drawn to that centralized tacnet and therefore slaved to a single commlink and vulnerable to spoofing.


Edit:
3) Simply get everyone good commlinks. It occurs to me that if everyone's on a device rating 6 commlink, then the spoof check is a target of 12 at -6 dice even with the centralized tacnet.

Now someone please refresh my memory, is the Device rating of a commlink System or the average of System & Firewall or the Average of System/Firewall/Signal/Response. I keep finding conflicting suggestions when I try to look it up.

Posted by: Ghost_in_the_System Jun 2 2011, 04:24 AM

By definition a master has admin privileges, but does not necessarily need a particular account. So the first thing you would have to do would be to get the access ID that is the master (which is more like the 0th thing you have to do). You can then spoof the slave -any- command. That command is 'Access ID x is your new master', which then gives Access ID x admin privileges because it is the master. You don't need to hack an admin account first, you're basically just claiming to be the system administrator, then saying you have a new system administrator, so your switching who owns an account, not having to create a new one.

As for peripheral device, guess I missed it, was trying to do a quick find through my PDFs and wasn't coming up for some reason. But as the Unwired Spoof says it operates on electronic devices, which includes nearly everything you would want to use spoof on, and the other things are also listed.


Why would you take the -6 dice? Spoof doesn't have anything saying that the access privilege required to perform the command has any bearing on the check, does it? My memory on some of the hacking rules are fuzzy, so could be wrong about that.


I think he means slave device A to device B with access ID B, then change device B's access ID to C. Now device A only accepts commands from an access ID that isn't broadcast anywhere. Device B can change its access ID back if it needs to (Pulling the proper ID from the hard drive or being entered from metahuman memory). Thus Device A is virtually 100% secure, but also a pain to access and requires a time delay.

Posted by: LurkerOutThere Jun 2 2011, 04:33 AM

Actually to do all three.

You should slave your comlink to the hacker so he can monitor the info sec side of the house (provided you trust him)

You should run the best independent security you can in case something breaks the slave linking (spoofing, jamming, hacker getting taken out by panther, actual or cannon)

You should put no more of your ware on your PAN then you need to do to get the job done. Even without tacnet I really like having an audio/video feed to my teammates so i usually keep that up.

Generally speaking cyberware hacking is dumb but there are some easy and relatively cheap precautions you cna take against it. Generally speaking I as a GM honor any reasonable precautions my players want to put in ahead of time. I encourage people at convention games to write down what precautions their taking on their character sheet. If i say to them "Hmmm and it looks like your comms are hacked" and they say "Nope, i've done XYZ" and can show me prior documentation of same I congratulate them on their preparation. Unless said methods involve clustering at which point I start to work on a way to feed them to ghouls.

Posted by: Fortinbras Jun 2 2011, 04:46 AM

Firstly "Access ID X is your new master" isn't a command you can give. A slaved account can only have one master at a time as it will only accept commands from that node. So the first thing you would have to do is delete that enslavement, otherwise the node will still only accept commands from it's original master.
Secondly, the Command "slave to Access ID X" cannot work because Access ID X has no valid account. A device does not recognize commands from any Access ID that isn't in it's account database. That's what hacking is. You are fooling the device into thinking you have legit access.
To slave the device to your Access ID, you would need to create an account. To do this, you need to hack into the system and create one using a Software or Hacking + Edit test.

You can't create any kind of network without access to all the nodes within that network.

EDIT: You could always just Spoof the device to add an account, but that is more difficult.

Posted by: suoq Jun 2 2011, 04:51 AM

Unwired, Pg 99. I think it's somewhere else as well but searching the PDFs gets frustrating.

-----------------

While we're on the subject. Does a loaded agent program count as IC or not? Any guidelines? Do I need to buy an agent program that's ONLY good for IC for it to count as IC? I get confused with that (much like I get confused with device rating and commlinks). If you're running IC as a hacker, what do you have loaded and running on your commlink at any one time?

Posted by: Fortinbras Jun 2 2011, 05:05 AM

All IC are Agents. Not all Agents are IC. For the most part, treat them as one and the same. They are both SK personas and treated the same for rules purposes.

As for device ratings, the BBB says that if the device is playing an important role, then it's Matrix attributes should be filled out. Device ratings are really just for GMs.

Posted by: Ghost_in_the_System Jun 2 2011, 05:08 AM

Sure it is. First off nothing in the book says a slave can't have more than one Master (But that doesn't really mater, never wanted it to). Second, Access IDs don't have accounts, so you can't set up an account for Access ID X, and thus don't require an account for Access ID X. Third, you don't need to delete the old enslavement, because that enslavement is how you are accessing the slave in the first place, and so deleting it would only be useful if you then planned to do a normal hack, but that's fairly pointless since you already more or less have admin access.

Device A is slave to Device B, Device C spoofs as Device B, Device C changes the value of the Master's Access ID to that of Device C. Device A is now slave to Device C. There is no need to make a new account, because accounts are just username/password, accounts have nothing to do with access ID. It may be true that after making this switch you'll need to crack the password on the 'Master Account', but it is a bit fuzzy from the description means that it requires an admin account (in which case you're just taking over the one that the old master used) or that the master is automatically given full admin access without an account (It basically gives admin level access automatically to anything coming from the master). I don't see spoofing working if it is the former though, because you'd send a command as the master, but it would be blocked if the master didn't happen to be logged in at that moment.

It seems instead that the slave goes "Oh look, an incoming connection/command, if it isn't from my master, I'll ignore it and send my master an update. Oh, it is from my master, I'll go ahead and do whatever it says, because it is from my master. Oh, it says that my master is now Bob instead of Joe. That's weird, but it is my Master, so I'll do what it says. Oh look, a command from Bob, I better do what he says."

Posted by: suoq Jun 2 2011, 05:09 AM

Alas, tests like Spoofing go vs Device ratings so I need to convert Matrix Attributes back to Device Ratings. Can't recall if there's a consistent way I should be doing it.

Posted by: Ghost_in_the_System Jun 2 2011, 05:16 AM

Add the stats together and divide by 4? Might create some oddities with high signal stuff being oddly hard to spoof, but whatever. Might also consider doing a direct substitution of firewall for the device rating.

Posted by: CanRay Jun 2 2011, 05:24 AM

If you slave a CommLink, you better watch out for the Abraham Lincoln AI... He's a nasty bugger!

Posted by: Fortinbras Jun 2 2011, 05:36 AM

Unwired does. "the slaved node does not accept any Matrix connections from any other node but the master." If it could have more than one master, it could accept more than one connection.


Access ID Account. Unwired p. 52.
In this scenario you aren't connected to any other node with access to the device, nor are you inputting a password, so you've got to have an Access ID account.


Yes, but you have to beat that device in a Spoof check for every Command you give. Hardly full access.
In addition, deleting the enslavement doesn't delete the Admin access of the master's ID. It simply frees up the device to be slaved to another. You can still Spoof the original Access ID.


You can't say accounts have nothing to do with Access IDs when Spoofing a device requires an Access ID.

I'm not saying it's not something you can do, it's just not something you can do in a single action. The device will indeed do whatever Joe says, but if the command from Joe is "Only listen to Bob" you are giving it conflicting info. One of it's basic structures is "Only listen to people on this list." That is integral to all Matrix interaction. Then someone on that list gave the command "Only listen to Joe."
You first have to put Bob on that list. Then say "Don't only listen to me(Joe)." Then say "Only listen to Bob."
Devices need to have things spelled out and can't have conflicting information.

Posted by: KarmaInferno Jun 2 2011, 06:17 AM

Have the TacNet on a separate network than your other gear. Just sensors, Smartlink, stuff that NEEDS to communicate with your teammates or outside devices.

Second personal network, hardwired, with everything else you carry that needs to talk to the PAN but not outside the PAN.



-k

Posted by: Eimi Jun 2 2011, 08:45 AM

Also, for people mentioning a smartlink being the one weak spot in an otherwise non-wireless activated cyberware character, remember the cyber safety from Augmentation. If you have one of those implanted, as well as an implanted smartlink (in cybereyes or on its own) and a gun with the appropriate safety chip, any wireless smartlink input is overriden, making your smartlink immune to hacking.

Posted by: Yerameyahu Jun 2 2011, 01:09 PM

Nah, the smartlink should just be skinlinked anyway, like *all* your gear.

Posted by: ravensoracle Jun 2 2011, 01:33 PM

Why not just add a script to your slaved device. If the Master tells you to switch Access Id's, ignore the command and start an Alert?

Posted by: squee_nabob Jun 2 2011, 01:49 PM

I don't consider spoofing to be that much of a problem. I've only had NPCs do it to me once, and it's more likely that I am spoofing something. Thus it is actually in the players best interest for it to be possible to spoof anything.

You can make unhackable stuff, but I am not convinced you can make something unspoofable via slaving. Slaving something gives it a Master, and it accepts all orders from the master. If you pretend to be the master, and then give it an order it obeys. All of the scripting to prevent that is what the opposed rolls represent.

Posted by: suoq Jun 2 2011, 02:05 PM

I'll admit. I'm not a fan of my gear being skinlinked.

Personally, I like the thought that when I have to leave my gun at the door, I've left a sensor at the door. It's my understanding that the camera on the smartgun (SR4A 322) is a guncam. (Please correct me if I'm wrong, but it certainly appears to make sense.). Since the guncam is trideo, that means I have a decent microphone reporting to my PAN as well as a camera that will let me know if someone is moving my gun. With a skinlink, once your gun is out of your sight, it's vulnerable.

Posted by: Ghost_in_the_System Jun 2 2011, 03:03 PM

MMmm, except that if it has more than one master, it still isn't accepting a connection that isn't from a master, but it does say the master, not a master. It isn't exactly spelled out, but I suppose only one master makes sense. Although that just supports my below argument all the more.


Okay, my bad, so that means I was indeed right about how the slave responds to a master.

True on the first part, and for the second part, my argument below.



You seem to be amazingly underestimating what a spoof command can do. You don't think that it can do three painfully simple things in one command? When commanding a drone you don't have to send thousands of commands of "move your left leg, now your right leg, now your left leg, now lift your gun, aim it at the target, pull the trigger, pull the trigger again, etc", you just send it the command of "Go there and engage enemies" Similarly you don't have to send the commands "Create account for ID X, Demaster from device B, master device X" You just say "Change master to X"

Also, I'm fairly sure that the account that device B would be using is simply the "Master account" ie. whoever is listed as the master always uses that account. So it isn't that Device B has an account and happens to be the master, it is that Device B is the master, and thus is using the account that is automatically used for the master.


Because if you really try, it's fairly easy to make something entirely unhackable, or a facility impossible to break into, but that really isn't in the spirit of being a game. Also, it isn't hard to imagine that the spoof program is able to bypass/defeat such a simple script.

Posted by: Yerameyahu Jun 2 2011, 03:49 PM

suoq, skinlink doesn't remove the wireless option. It's just a better option when you do use it.

Ghost, I'm not sure. Drones have a Pilot. Commlinks don't. It's bad enough Unwired added the 'spoof commlink' loophole (apparently 'devices' and 'slaved nodes' are now valid targets for Spoof—they never were before).

Posted by: Ghost_in_the_System Jun 2 2011, 04:34 PM

No, but it does have an operating system, and in SR those seem to be designed to be as user friendly as possible, so that you have to do as little input as possible to get results. Like I'm sure you just hit 'call bob' and it automatically opens the phone app (or whatever), retrieves the number of bob, and dials it'. Similarly, I think if you hit 'Slave this device to X' it would simply do all the setup that might be required to do so, it wouldn't have you create accounts (Though I still contest that you don't need an account to do this, and simply uses a 'master' account for whoever happens to be the master), or add IDs to the accepted list (Again, not something I think really happens, because it would make all systems virtually unhackable on the fly), or do anything else manually. It'd simply do what is required to make X the master. Kind of like when you install a program, you don't have to unpack individual components and manually install them, you just hit install and the computer does the rest for you (with a couple questions that you can ignore like where to put the program).

Posted by: Yerameyahu Jun 2 2011, 06:09 PM

Yes, but the fact that Pilots exist and *themselves* have trouble interpreting commands ('go kill that there') overwhelmingly implies that something lacking a Pilot shouldn't be able to do what a Pilot does.

It certainly depends on the command in question, but I'm saying it's better to err on the side of 'less spoofable' for commlinks. It just seems more fair. After all, Issue Command is a 'tweet-sized list' at max (often, more than you need), and Spoof Command is "Each Spoof Command action applies only to a single command; multiple commands require multiple Spoof Command attempts."

Posted by: Fortinbras Jun 2 2011, 06:09 PM

You seem to be under the impression that there is a "master account." I'm not sure that is a thing. It's not something I can find in either the BBB or Unwired. There are only 4 types of accounts. Public, user, security and admin. If the device is slaved, then it's master has admin access and it won't receive wireless commands from other users. The slaving of it takes up a subscription slot, but that is by no means a "master account."
Therefore giving the command "Change Master Account" can't be done because there is no such thing.

The three things you are asking it to do aren't "painfully simple." To suggest that a device would perform multiple actions after one command is like suggesting that performing the action "Download all files with X." That is also something you can't do. You first have to find the files with Browse. Then you have to Decrypt them, then you have to download them.
Similarly you can't give the command "Kick off all users that aren't me." or anything like that because that is not a Matrix action. The actions you can take are represented by the Matrix rules.

Also, a System can't perform actions, it can only do things it is told to do. A Pilot program can be told "Go shoot that guy." and can interpret that info, usually after a Pilot x 2 + Adaptablility + Fuzzy Logic roll. A System has no such roll. It doesn't have a dog brain like a Pilot program. Moreover, it has no IPs, so it's can't take the multiple actions you are suggesting it do.
Even if it could interpret your Command, it would first have too delete it's former enslavement on it's first action, then add your Access ID to it's data banks of acceptable users on it's next action, then slave itself to your commlink on it's third action. Thus you still couldn't do this in one action as you suggested.
But, as Systems don't have actions, this is something the Spoofer has to do. A Pilot program might be able to do this if it had the necessary programs like Edit, but not a commlink.

As I said, this is something that can be done, but it can't be done in one command or one action. Things are only truly un-hackable if you start creating your own Matrix actions and ignore the rules system set in place to interpret such complex systems as the 2070 Matrix.

Posted by: Ghost_in_the_System Jun 2 2011, 06:50 PM

I think you're misunderstanding what I mean by master account. I mean an Access ID admin account which is automatically set to whoever the master happens to be.

And to suggest that a device can't perform multiple actions is ludicrous. As I explained, you only have to hit one button to install a program, despite the fact that the installation process requires (among other things): to create multiple folders, to extract files, to install files in multiple locations, to add itself to the program registry, to make a desktop shortcut, to make a start menu shortcut, possibly add itself to the 'run on startup' list. This is exactly comparable to whatever steps might be required to slave a device. Both require a single command. "Install program X" or "Make this device a slave to device X". If you want to claim that changing the master of a slaved device requires more than a single command, then you need to realize that everything you do in the matrix would require dozens or hundreds of commands.

Also, I'm fairly sure there is a 'disconnect all users' action specifically spelled out in the matrix rules, though I could be wrong. Yep, there it is, p223 BBB. But that is entirely beside the point. You also don't need to decrypt a file before you download it, you just need to decrypt it before you can do much of anything with it.

From the description, I figure this is more or less exactly how slaving works, so all you need to do is change what the master access ID is. Since you're the one complaining about actions not in the book, I'd like you to find where it says "A master requires an admin account" as opposed to "A master has admin access".

Also keep in mind that when I say "send the command 'change master to X'" I don't mean that is literally what you send. I mean you send whatever the code equivalent of getting that done would be, there is no need for the commlink to interpret anything.

Posted by: Fortinbras Jun 2 2011, 07:43 PM

Again, this isn't something I see anywhere. I don't think this is a thing in Shadowrun.
It would be counter intuitive to even have such a thing. You need Admin Access to slave a node, therefore you just give it a subscription slot. Why would you make a script that would give admin access and a master subscription to another user? If such a line of code existed, and I can't find anywhere that it does, why wouldn't you delete it?


Except you aren't talking about obscure Matrix concepts, you are talking about specific Matrix actions that must run in real time with the rest of combat. Load Program can be done in one Matrix action and all the addition computer stuff is presumed to occur within that action. Changing and deleting subscriptions are very specific Matrix actions, all of which take a Complex action. Specifically they are Edit actions. One cannot perform more than one Complex action per Initiative pass. These require multiple Commands.


If you are referring to Terminate Connection, this can only be done against one person at a time and only if they don't have a legit account. That is it's limitation and why "dissconnect all users that aren't me" isn't a Command you can give. It is too broad and doesn't take into account all the things happening in the Matrix that the rules are there to interpret.
If you are referring to System Reset, then that takes more than 1 IP and does far more than dissconnect all users.
In the Decrypting scenario, one is presuming the files are Encrypted. You still need to find them first, which takes multiple Matrix actions.

Unwired p. 53
"Admin privileges empower individuals to...
assign access levels to accounts, and assign privileges to account levels."

p. 98
"To add, alter, or delete an account, you would almost certainly
need to spoof a command from someone with admin privileges."

I think adding a subscription falls under the heading of altering an account and creating a master/slave connection certainly falls under the heading of assigning access levels.



I understand what you are wanting to do, but you can't do it in one command. You are giving the system multiple conflicting orders with that order. Much like the command "Download all files about X" or "Give all icons that aren't me User Access only" or any other million things you can imagine.
It's just like combat. You can say "I toss the guy shooting at me into the other guy shooting at me" and that is something you can do, but first you have to roll to grab the guy, then roll to throw him then roll to hit the second guy.
These are doable things, but you have to follow the rules in order.

The FAQs try to throw out the baby with the bath water and say you can't Spoof nodes, but not only does this contradict Unwired, it's more limiting than I think most people want to make it. In order not to throw off game balance, you have to use the Matrix rules as they stand.

Posted by: Bearclaw Jun 2 2011, 07:50 PM

In your code example, you make it clear that the system will only accept input from the master.

If I understand your point, you are saying that you can send a single command that will edit the "slave.ini" to replace the previous access ID with your access ID. Which of the commands listed in the Shadowrun rules will do that?
It seems to me you still would have to get yourself in somehow before you could edit that file. As in, you would have to spoff a command to un-slave. Then, hack the system to give yourself an admin account, then slave it to yourself.

Your idea only works if that file were a text based .ini that you could use your one command to copy on to the system (which I also don't think would work, but let's not worry about that right now). As apposed to an encrypted hex file that you would need to steal, decrypt, and pick through line by line til you found the line, in hex, that contains the Access ID. Edit that line. Re-encrypt using the same software and key, then copy the file back to the target, and telling it to over-write the origonal. Right?

Now, even if the file is an unencrypted ASCII .ini file, how do you, with one spoofed command, edit it?

Posted by: Fortinbras Jun 2 2011, 07:53 PM

Coding in Shadowrun isn't text based. It's too complex. It's metaphor based.

Posted by: Ghost_in_the_System Jun 2 2011, 08:15 PM

You're extrapolating a ton of stuff. Read through the book, there is no action at all, of any kind, that causes a device to become the slave to another device. There are no listed requirements for a device to be a slave to another device. Any requirement or action you come up with is just as much 'not a thing in shadowrun' as my idea that you can spoof 'change master to X'. I say again that if you'd like to show me where it says the requirements of making a device a slave, or the actions required to do so, I'd love to see it. Because until then, by your own arguments, making a slave is in fact impossible because it isn't listed as an action. Sure, it says that you can make a device a slave, but it never says how, and since it doesn't say how, it isn't an action in shadowrun, just like changing the master isn't an action.

No, it is pulling the variable of the master_access_ID, which is something that you would generally be able to easily change with a single command. There is no need to edit any kind of .ini file any more than there is a need to edit a .ini file to do anything at all on a computer. As I've said a dozen times before, you just have to send the command that changes the master's access ID to whatever you want it to be.

Let me say it this way:
Somehow it is possible to make a device a slave to another device. We don't know the exact processes involved in this, just like we don't know the exact processes involved in running a program or installing a program. We do however know that it is possible. And just like installing a program, it is likely done (since everything is geared towards user friendliness) in a fairly quick format that requires as little input from the user as possible, just like installing a program is done by double clicking the .exe file and hitting yes a few times. Thus there is no reason that you shouldn't be able to send a single spoofed command to run that same setup process to change the master to something else, just like you should be able to send a single spoofed command to install a program.

Posted by: suoq Jun 2 2011, 08:57 PM

PG 59 Unwired. That's the action, slaving the node.

Don't ask me where Bearclaw (or anyone else) is getting his idea that the slave action can't be spoofed. I don't get it either. As far as I'm concerned the easiest thing to do if that happens to one of your devices is physically turn the device off.

From a programmer standpoint, I need to be able to move slave nodes to other nodes and it makes life a LOT easier if the master node can do that. Now if someone wants to write on their character sheet that the only way to deslave a node is (for example) to reboot it, sure. However I've found people tend to regret excessive security measures. Make a device secure enough and eventually it will lock you out. That's why most security isn't secure. No one puts up with being locked out of their car or home and discovering a locksmith can't unlock it in under a minute. They want security that a locksmith can bypass but criminals can't. For fun, sell people those lug nuts that can't be removed except by the unique tool that (I guarantee you) they'll be unable to find when they have a flat tire. Oh man, those people get mad...

Posted by: Ghost_in_the_System Jun 2 2011, 09:10 PM

Right, but my point is that it is something you can do, not an action. It isn't a simple action, isn't a free action, isn't complex or extended. It doesn't list what if any kind of roll is required to do so, or what kind of access is required to make it happen on a device. All it says is that slaving a device is possible, which by all rights means that changing the master of a slaved device should also be possible, but similarly lists no exact actions, conditions, or requirement for doing so.

So, Fortinbras' constant argument that changing the master isn't an action spelled out in the book applies equally to making anything a slave in the first place. It also means that his insistence that being a master requires an access id admin account equally untrue, as that is also not listed as a requirement. It also means that any steps he lists as being required to change the master are entirely of his own fabrication.

Posted by: Yerameyahu Jun 2 2011, 09:17 PM

I didn't read everything yet, but *any* change on a device should be assumed to be at minimum a Change Linked Device Mode (Free action), or something like it. Obviously.

It's pretty reasonable to assume that slaving requires an admin account, though. The alternative is just incredibly bad computer security, and would instantly be changed if it were the case.

Posted by: suoq Jun 2 2011, 09:19 PM

Take a deep breath.

It's Shadowrun.

If you do find an exact action, condition, or requirement in the rules, treasure it, especially if you look and you can't find a conflicting action, condition, or requirement somewhere else in the rules.

Eyeball it. Wing it. Roll with it. It's not neurosurgery. It's a role playing game in a world that fractally doesn't make sense*.



*anything that doesn't make sense is supported by other things that also turn out not to make sense which are supported by other things that....

Posted by: Fortinbras Jun 2 2011, 09:25 PM

Slaving requires a Subscription.(Unwired p.55)
A slave may only be subscribed to one master(same)
Subscribing to a node is a Complex action(SR4a p.224 & 231)
Unsubscribing to a node is a Simple Action action(SR4a p. 229)
When one subscribes to a node, one does so at a particual access level, from public to admin(SR4a p. 225)
To slave a node requires Admin Access(Unwired p. 53 & 98)

One must terminate the ongoing subscription by spoofing that simple action. Then one must create or hack a subscription at Admin access, at least one complex action. Then one must make that subscription slaved. This is most likely a Spoof Command Complex action if one is still spoofing or a Change Linked Device Free Action if you are hacked in.
Even if you sent this as a script for the System to execute, it couldn't be done in one turn.

Just because you can't find the rule, doesn't mean it doesn't exist.

Posted by: KCKitsune Jun 2 2011, 09:40 PM

Why? Clustering allows you to run a whole bunch of IC. My combat medic mage has 13 independent pieces of 'ware. If they were all alphaware, then I could run 26 rating 4 programs on the resulting Node. That would be the following:

1. Agent (Rating 4) with Adaptability (Rating 3), Cascading (Rating 3), & Expert Defense (Rating 3)
2. Attack (Rating 4)
3. Armor (Rating 4)
4. Blackout (Rating 4) or Medic (Rating 4) for the Medic Agent

This would be multiplied 5 times (4 Attackers and 1 Medic), so when your NPC hacker comes in he gets a VERY rude surprise. This also doesn't take into account what IC is running on my 'Link...

Posted by: Yerameyahu Jun 2 2011, 09:48 PM

That all seems pretty clear, Fortinbras.

Because doing it with *cyberware* is pure evil, KCKitsune. Hehe. Clustering *real* nodes in order to run more IC is fine (although you could also just run them on those nodes and link in, which is the usual method with spiders).

Posted by: LurkerOutThere Jun 2 2011, 10:05 PM

Yes, but you'd scream bloody murder if the corps did that by comparison. Every single player I have had use clustering on me at a con that i've gotten time to reverse the roles has flipped their lid.

Clustering was obviously written by someone with only a passing understanding of computers or even electronics in general. This notion that there is all this extra CPU cycles on your ware or your toaster to make it comparable to a dedidicated computer device is patently rediculous. It's like someonje read about seti at home and suddenly thought it was a viable project in microcosm. It's not clever it's just bad rules.

Posted by: James McMurray Jun 2 2011, 10:17 PM

So wait... You're saying players don't want to be turned to paste by their own ideas? Then why do they keep coming up with such nasty stuff?

Posted by: LurkerOutThere Jun 2 2011, 10:23 PM

I might even give a little bit of credit if it was something that was obscure or difficult, but nope, it's "slap a bunch of ware together, get infinite response". Idiocy. Ok i'm better now.

Posted by: Fortinbras Jun 2 2011, 10:28 PM

Thank you.

I ruled that if you are clustering cyberware, then that cyberwaer now "plays an important part in the adventure" not "a passing role" and is no longer subject to being a simple device rating. I just use the sample peripheral devices from Unwired as a base to give it a full compliment of Matrix stats. Same System, lowered Response, like a credstick.
This way no one is running around with single cybereyes in their pockets with Responses of 6.
Does that seem fair or Draconic?

Posted by: Ghost_in_the_System Jun 2 2011, 10:50 PM

I understand that, I wasn't directing those comments at you, I was directing them at Fortinbras as he was the one saying that if something wasn't an explicit action it couldn't be done.


Read again. It says that the slave is linked to the master, it does not say that a subscription, active or otherwise, is required.
Also not stated, it says that it doesn't accept any connections that aren't from the master. Presumably if there was more than one master, it would accept connections from more than one master.
Irrelevant, as a subscription is not required.

p53 makes no mention of slaving, and page 98 indicates that you need security access, not admin access, to slave a node to something, but does not say that the ID that you are slaving it to requires any sort of account at all. It says quite plainly that you can use a riggers account (Presumably security since the section is talking about what different levels of accounts can do, and the rigger account portion is between the basic and admin descriptions) to slave that drone to your commlink. There is no mention of 'slave it to your commlink after booting off the rigger and hacking yourself an admin account'. No, it says that by spoofing a security access ID you can slave the device to your commlink, period. No other requirements.

So, thank you for providing that page reference that I'd missed to prove my point.

Posted by: Ghost_in_the_System Jun 2 2011, 10:53 PM

Seems fair to me

Although system can't be higher than response as I recall, so that would make a high system/low response device fairly pointless. Unless maybe it is optimized... can you optimize system?

Posted by: Bearclaw Jun 2 2011, 10:57 PM

From pg. 55 Unwired

Posted by: Yerameyahu Jun 2 2011, 10:58 PM

It can, for peripheral devices. So it'll be okay right up until you cluster.

Slaving should obviously require a subscription. If it's true that the book says it doesn't (or, perhaps, doesn't say it *does*), then it's just one more erroneous Broken Rule to ignore.

The fact that it's *the* master seems pretty clear; one master.

Posted by: Bearclaw Jun 2 2011, 10:59 PM

Don't you need a subscription to execute ANY commands on a node?

Posted by: Fortinbras Jun 2 2011, 11:04 PM

You read again. It's the big box in the middle of the page under "Actions Needing Subscription."

You said it yourself "the master" not "a master." That means one and only one master.

p. 53 makes note of what admin access can do and what security access can't do.
p. 93 says to in order to add, alter or delete accounts(altering being the key one as you must alter it to give it a subscription to a separate node) you need admin access. Nothing is mentioned about security access.

I'll ignore that the page is referencing drones and not commlinks, and simply say that the text states it can be done. I've never claimed it can't be done. It simply can't be done with a single action.

As much fun as it would be to simply say "I do this." you run into the old Cowboys & Indians problems of having the other guy say "you missed me."
That is why the rules are there. To regulate actions accordingly. They have to be followed, else Matrix actions become a series of escalating arguments between people who know computers really well while the rest of us who don't simply throw that aspect out of the game.
Then Hacking becomes Decking. And Rule #1 is the decker always dies.

Posted by: Ghost_in_the_System Jun 3 2011, 12:48 AM

Why? Device A only accepts connections and commands from Device B, why does either device need an active subscription for this to happen? That's like saying a password is only needed if someone else is already logged onto an account.


And I'll say for the millionth time that nowhere does it say that slaving requires an account of any kind. I'll also repeat myself that it says very clearly that a rigger's grade account access via spoofing (security by virtues already explained) can slave a device. You continue to fail utterly to provide a single line from the entire book that says that slaving a device takes more than one action, or requires any kind of account.

I've provided evidence directly from the book that says: Security access is good enough to slave a device (drone is the specific example, but a drone is simply a type of device as far as the matrix is concerned). Since security level access can't add/remove/adjust accounts (as you've gone through great pains to point out), this directly indicates that an account is not required for master/slave relationship. You've failed to show where a subscription is required for the master/slave relationship. I'll concede that a subscription is required to do much of anything besides send single commands, but it isn't necessarily required for the relationship to exist.

And as much fun as it is to simply say "This is totally required even though there is no evidence for it book, and evidence to the contrary" you run into the old Lack of Evidence problem. That is why the rules are there. To regulate actions accordingly. They have to be followed, else Matrix actions become a series of escalating arguments between people who know computers really well while the rest of you who don't simply throw that aspect out of the game.
Then Hacking becomes Decking. And Rule #1 is the decker always dies.

Posted by: Yerameyahu Jun 3 2011, 01:52 AM

Um, seriously. It's there in black and white on p55: slaving requires a subscription. As we all know it should, anyway. It literally could not be clearer.

Posted by: Ghost_in_the_System Jun 3 2011, 02:07 AM

Geez x.x kept missing that box. Still don't know why A couldn't be a slave to B even though they aren't connected, as it seems somewhat silly that as soon as B turns off or loses wireless, A suddenly decides to accept connections from anyone. So really there is a 4th method to getting into a slaved device, and that is to disrupt the connection between master and slave, via a jammer perhaps.

Posted by: Yerameyahu Jun 3 2011, 02:29 AM

AFAIK, you can probably 'program' arbitrary conditions into a node (like an Access ID whitelist). Half of slaving is forwarding connections; between that and having admin control from the master, that explains the subscription.

In ideal-ville, the rules are supposed to be abstract, simplified, and balanced, compared to reality. If you want to make changes (reality-based or otherwise), do go for it.

Posted by: deek Jun 3 2011, 02:13 PM

Ghost, wouldn't you at least need to find the Access ID of the comm you want to slave? I follow you on saying you don't need an admin account to slave, that security will do just fine, but if there are say, 3 Access IDs that are slaved to one master commlink, I'd think you would either need to perform three actions to slave each of them to a different Access ID or take two actions to do one (that being an analyze to get the one Access ID and a second to reslave it).

Granted, I think by spoofing, you could simply get in the master commlink as a user account and then just start spoofing commands to the slaves, instead of worrying about getting some sort of real control over the hacker's comm.

Posted by: Ghost_in_the_System Jun 3 2011, 02:32 PM

Oh yes, if you want to change over multiple slaves, that is going to take multiple actions. And yes, you would need to find the slave you want to change ownership of. I've been operating under the assumption that you know the slave already, and are looking at how to get into it, as opposed to looking at the master and wanting to take its (unknown) slaves.

Posted by: deek Jun 3 2011, 02:40 PM

I think the only way you'd know the slave already is if you had its Access ID before it was actually slaved. Since any traffic directed at the slave goes straight to the master, I don't see how you could even do a Matrix Perception on a slave as its effectively invisible. Even if you knew its physical location, you couldn't get an Access ID wirelessly, right?

I'm just thinking that going into all of this, like in a combat or whatever, you won't know the Access ID of the slave, so you do have to get in the master first. That is kind of the point of slaving comms to the hacker for security. The slave turns invisible and the master has to be hacked first, which should be harder to do...

Posted by: sabs Jun 3 2011, 02:43 PM

You could get the accessID with a EW+Sniff test to pick up the packets, and then a decrypt to decrypt the packets.

Posted by: Ghost_in_the_System Jun 3 2011, 02:46 PM

Sure you could, that is basically the purpose of the scan program. How else would you ever connect to anything?

Posted by: deek Jun 3 2011, 02:55 PM

Ghost, the way I am reading it is all traffic (which should include scanning) would be forwarded to the master to handle. So your scan on a slaved commlink would go to the master. Again, this is only for slaved commlinks, not anything in general. Slaving a commlink provides an extra level of protection.

sabs, that makes sense. You could pick up the packets. And if you did so before engaging in combat or other interaction, you'd have less actions to worry about. If not, well, you've got some extra steps to perform to get the Access ID, which may actually be quicker to get into the master comm and not putz around with the slaved commlink at all externally.

Posted by: Ghost_in_the_System Jun 3 2011, 02:58 PM

You could still get the access ID from scanning, just like you can still get it from scanning a hidden node.

Posted by: deek Jun 3 2011, 03:10 PM

Wouldn't the slave have to respond to the scan? That seems like a contradiction to forwarding all traffic and control to the master. I don't think a slaved node is the same thing as a hidden node. The way I read it, a slaved node, for all intents as purposes, turns into the master node, so that's the only node you are able to interact with wirelessly. Again, I thought security was the whole purpose of slaving, making you have to go through the master node (which is assumed to be better secured than any of the slaves) for everything.

Posted by: Bigity Jun 3 2011, 03:20 PM

I agree, I get the entire 'chokepoint' vibe from slaving devices. You have to go through the chokepoint to reach anything else. You can spoof commands to slaves, but only from the master node.

Of course, I have one read-through of the matrix stuff in SR4 under my belt, so I am still pretty much helplessly confused.

Posted by: Ghost_in_the_System Jun 3 2011, 03:21 PM

Yes, and you do have to go through the master node. I don't know how scanning works, but my guess is that it takes the wireless signal that a device puts out (even the slaved device puts out signals to the master to maintain the required subscription) and pulls the access ID from that. It's kinda like a specialized form of intercept traffic. At least that's how I think of it. However it works though, scanning makes no mention of the device being searched for needing to accept connections or anything of the like, and slaving makes no mention that it prevents the device being scanned.

Posted by: Tymeaus Jalynsfein Jun 3 2011, 03:23 PM

You cannot spoof a Slave unless you have access to the Master Node. You cannot find an Access ID for a Slaved Node, because you do not detect the slaved Node, you detect the master. Etc...

Here is the quote for Slaving and accessing Slaved Nodes...

You only have three choices for deeling with a Slaved Node.

1. Hack it Physically through a WIRED CONECTION (Or through its own hardware)...
2. Hack the Master Node to gain control of it, and thus any Slaved Nodes.
3. Spoof the AID of the Master Node, and THEN Spoof commands to the Slave.

That is it... Nothing else will work.

Posted by: suoq Jun 3 2011, 03:43 PM

I don't see how spoofing the access ID of the master node = "have access to the Master Node".
A Trace User test will provide the access ID. A Capture Wireless Signal Test will allow you to do a Trace User.

Posted by: Ghost_in_the_System Jun 3 2011, 03:47 PM

No one was disputing any of that, I don't see why you quoted it. Though actually, there is a 4th method. You can break the connection between master and slave (Via jamming for instance), which would cause the slave to no long be slaved, as being slaved requires an active subscription, and thus make it hackable via normal means at least until a connection is reestablished (after which it isn't stated if the slave/master relationship starts up again automatically or needs to be manually reinstated). Even so, I don't know why you quoted ways to break into a slave when we were talking about obtaining an Access ID which doesn't require a connection to the node in question, or a connection of any kind.

Posted by: Tymeaus Jalynsfein Jun 3 2011, 03:52 PM

If you can spoof access to the Master node, then you can also spoof access to the Slaved Nodes, but it is a two step process. Don't try to relate it to anything in real life, that way lies madness...

No it wont, it will supply the AID of the MASTER NODE. Unless, of course, you are using the Slaved Node to communicate Wirelessly. Why exactly would you do that?

Posted by: Ghost_in_the_System Jun 3 2011, 03:56 PM

You would do that because it is required in order to have a subscription with the master node which is required for the master/slave relationship.

And no, you spoof the access ID of the master node so that you can give commands to the slave. You don't 'spoof access to the master node'. You may be mixing up what spoof is/does. If you spoof the access ID of the master node, you are basically pretending to be the master node. You don't have to get any actual access to the master node, or indeed interact with the master node any more than is required to obtain its access ID.

Posted by: Fortinbras Jun 3 2011, 03:58 PM

But if both nodes are communicating wirelessly, aren't both their AIDs being sent through the air?
And if they are, can't you capture the wireless signal of both?
(Not being Socratic. Honestly asking.)

Posted by: Tymeaus Jalynsfein Jun 3 2011, 04:01 PM

I am not mixing up my rules here. Read the Rules I quoted... and I quote, again...



Pretty evident there, don't you think?
Once you have SPOOFED the AID of the Master Node, You still have to SPOOF the commands to the Slave, uisng that Spoofed Master Node AID.

Why is that so difficult to understand. You are NOT the Master Node in that Case. You are pretending to be the Master Node, and then Pretending to Give appropriate commands to the Slaved Node.

Posted by: deek Jun 3 2011, 04:02 PM

Yeah, for spoofing, you just need to know the Access ID of the Master and the Node, then you can "pretend" and send all the commands you want. But, you do need to get both of those Access IDs to make it work, so its multiple steps again.

Posted by: Tymeaus Jalynsfein Jun 3 2011, 04:04 PM

It is not so much that you need 2 AID's. It is that you need to convince the Slaved Node that you are indeed who you say you are, which is a Spoof Command. You only need the AID of the Master node.

Posted by: Ghost_in_the_System Jun 3 2011, 04:07 PM

Hmm, that's weird, I assumed that spoofing a command included making it look like it was coming from the correct access ID. Otherwise why wouldn't you spoof your access ID to look like the master and then just use normal commands instead of spoofs? I mean if you use spoof to look like the master, why do you also need to use spoof to make commands look like they came from the master? Hang on, let me read up on what exactly spoof does to get a clearer picture.

Posted by: Ghost_in_the_System Jun 3 2011, 04:11 PM

Yeah,

I think the problem here is that spoof is a term, as well as a program. You spoof the master as part of the action of sending a spoofed command. It only requires one action to do so (not counting actions required to obtain the access ID of the master in the first place).

Posted by: Tymeaus Jalynsfein Jun 3 2011, 04:11 PM

The Problem is that you need to Spoof the Master Node for the Access to be able to send additional commands down the pipe. And, because you are NOT the Master Node, You need to Spoof any commands (using the Master' Node's AID) to the Slaved Node. It is often just easier to Hack the Master node and then you have access without needing to Spoof. But Spoofing does not usually carry the risk of actually hacking the Master Node (You are pretending to be a legitimate User from the "Outside" because you have a Valid AID).

Posted by: suoq Jun 3 2011, 04:39 PM

No. You don't. The rules you're quoting say nothing about spoofing the master made, just spoofing the access ID of the master node

Here's the classic example from http://forums.dumpshock.com/index.php?s=&showtopic=26473&view=findpost&p=804857 trimmed down to just Slamm-O!'s actions.

Posted by: Bearclaw Jun 3 2011, 04:51 PM

I think you are mostly right.
I believe you must Sniff the Master Node's signal for the Master Node's Access ID. Then, you use the Spoof program to send a command to the Slave system. The Spoof program convinces the slave that the command came from The Master. But, each time you want to send a command, you must fool the system again. Because you haven't hacked it. You are just using The Master's AID.

Like the scene in Terminator 2. The T1000 (the hacker) used John's (the slave) Foster Mother's (the master) voice to convince John to come home (spoofing a command). But, the T800 (John's Firewall) realizes it's a fake and John ignores the command.

Posted by: Tymeaus Jalynsfein Jun 3 2011, 04:59 PM

The problem is that he cannot Directly Spoof the Drone if it is SLAVED to the Master Node, which in the scenario above is not stated. For him to spoof orders to the Drone, in general, he must have the AID of the Connected Subscription from which it is receiving commands. In the case of a Master-Slave Scenario, the only way in is through the Master Node, whcih is not the case in the example that you used. The Drone is not SLAVED to the Rigger, the Rigger just has a Subsrciption to it.

I would have the Drone Slaved, personally, but it is NOT required.

You MUST know about the Slaves available before you can command them, since you cannot tell if the node you are in is Master or not without actually observing its connections. You are not notified when you are forwarded from an attempted Connection with a Slave, because all you see IS THE MASTER. You need to analyze the Node to discover any Slaves present. This is why you need to Spoof access to the Master.

Posted by: deek Jun 3 2011, 05:02 PM

How would you send a spoofed command to a Slaved Node that you don't have an AID for? Without two AIDs, you can't communicate.

Posted by: Tymeaus Jalynsfein Jun 3 2011, 05:05 PM

You do not need the AID of a Slaved Node, you need its Master's AID. You still have to Spoof the Slaved Node (Using the AID of the Master; but as long as it accepts the Master's AID, it will do what it is told to do) of course. The Slaved node is already accepting commands from the Master, and you do need to know of the Slaved Nodes Presence (Using Analyze, of course), but once you have found the Slave, all you need to do is convince it that you are the Master Node.

Posted by: Ghost_in_the_System Jun 3 2011, 05:13 PM

Slave or not in no way matters in the example provided (assuming that the rigger is the master, which she would have to be to operate the drone). Since you are trying to break into the slave, knowledge of the slave seems to be an already established fact, so there is no need to spoof access to the master (which isn't a thing).

The spoof program is basically identical to the command program in all respects except one, which is that instead of giving your own Access ID, you can give any Access ID you feel like. So, if something will happen when a command is sent from access ID X, then it will happen if a spoofed command is sent to look like it is from access ID X. Access to the Master is in no way required for this. Now, finding what slaves a device might have, that is an entirely different matter.


Well, technically an Access ID isn't required for communication. I can go "Okay, I want to connect to that node right there." without actually figuring out what its access ID is. Kind of like how you can connect to dumpshock without knowing its IP address.

Posted by: suoq Jun 3 2011, 05:18 PM

Of course he can.

Do we agree that he can directly spoof the drone if it's NOT slaved. If so, can you show me where a slave connection makes it unable to be directly spoofed because the rules you're quoting only require the access ID of the master node, something Slamm-O! already has from his Trace User.

I can't find that rule anywhere. Actually, I can't even find "Connected Subscription" in the rules. Do subscriptions even have Access IDs?

Not unless it's a wired connection.

Posted by: Ghost_in_the_System Jun 3 2011, 05:20 PM

I think he meant the access ID of the device it is connected to through the subscription, which the example provided has.

Posted by: Bearclaw Jun 3 2011, 05:30 PM

The rules say you can. They say exactly and without quesiton that you can, because it's one of THREE, not FOUR, ways you can control a device that is slaved to a commlink that isn't yours.

pg. 55 Unwired

Posted by: Ghost_in_the_System Jun 3 2011, 05:51 PM

Hey, when faced with a slaved node, you also have the option of ignoring it. You also have the option of convincing the owner of the master node to do stuff for you. You have plenty of options that aren't listed, or are you saying that those things magically aren't options? That if you find a slaved node you must hack into it because ignoring it isn't an option?

Posted by: Fortinbras Jun 3 2011, 05:58 PM

My team's technomacer with her Black Hat Paragon has to.
One of these days she'll get those 3 hits. Well, one of these days she'll try to.

Posted by: sabs Jun 3 2011, 06:00 PM

remember, the slave node is still part of the adhoc network matrix system.

It receives your 'packet' checks to see if your AccessID matches the one of it's master.
Yes: Do what you tell it to
No: Forward request to Master

It is possible btw, to edit the AccessID of your commlink, via a hardware modification, or a computer+edit roll. If you made your accessID the same, then you could hack the node. Especially if you jammed the other commlink's signal.

Think of AccessID as a MAC address.

Posted by: Ghost_in_the_System Jun 3 2011, 06:03 PM

Isn't there something preventing two devices with the same access ID from being around at the same time? I mean, I'm sure you could override it, but it would cause all kinds of problems, wouldn't it? I mean, you'd get everything that was intended for the other ID: Phone calls, matrix connections, even data requests by the original device. The other person would also quickly notice anything you're up to, because anything sent to your device would also be sent to theirs.

Posted by: Bearclaw Jun 3 2011, 06:12 PM

Isn't that exactly what Spoofing is?

Posted by: suoq Jun 3 2011, 06:13 PM

No.

The protocol says DON'T DO IT so programmers don't do it and hackers do. There's no law of physics or game rules that prevents it as far as I'm aware of. Access IDs aren't particularly secure.
Define "problems"....
That may meet your definition of problem. Not mine.

Why would I have anything sent to that access ID I don't want the other person to see?

Posted by: deek Jun 3 2011, 06:27 PM

TJ, yeah, I wasn't thinking of broadcasting. You are right, as long as I have the master's AID, any slaved nodes that can get my spoofed command will try and perform it. Now, there is risk of being found out quicker if I don't specify only one slave and the other slaves can't perform the action. But yeah, if I broadcast with the right AID, then I don't need to know any of the slaved nodes's AIDs.

Ghost, I'm not sure that I can just pick a slaved node, though. That's just software abstraction that allows you to pick. Just like dumpshock, I may not know the IP, but the IP has to be connected to dumpshock. My impression is that a slaved node does not appear anywhere, as all traffic gets forwarded to the master. Now, that may be a way to figure out a node is slaved, as I could be looking at a comm 3 feet away and try to connect to it and find my traffic is travelling to a comm that is 30 feet away. Again, the way I am reading slaves is that once the node is slaved, it disappears (wirelessly) from any view and only the master is "visible". Granted, the traffic is still being forwarded so that is accessible to scrape AIDs from the intercepted traffic, but you are not interacting with that slave unless you physically connect, go through the master or spoof the master.

Posted by: Fortinbras Jun 3 2011, 06:50 PM

If you try to show up in a node which already has that Access ID in it, the node would refuse to allow it because a persona with that access ID already exists.
It's the same thing that prevents Agents from making infinite copies of themselves on the same node.

Posted by: Tymeaus Jalynsfein Jun 3 2011, 07:08 PM

It makes all the difference if a Node is standalone or slaved... What books have you been reading?

Unfortunately, an AID IS required for communications. You cannot communicate without transmitting an AID. It is just not something that you generally worry about, because your own AID is used when connections are made. Which is why you can be traced. And your dumpshock relation is inaccurate. You leave a log of what actions you take on Dumpshock, and specific information about you is recorded in the log. Among that data is likely the "Access ID" that is provided to Dumpshock that has the details of your account. If you are anonymous, then they likely just get your IP address, or other such information.

Posted by: sabs Jun 3 2011, 07:12 PM

Also remember that spoofing is 1 way. You can send Spoofed commands to a node, but you cannot spoof receive information from that node. You can tell the drone to shoot target X, but you can't know if it's going to do it or not, until it does.

Posted by: Tymeaus Jalynsfein Jun 3 2011, 07:12 PM

Indeed, but not in all situations.



Of course he can Spoof an Un-Slaved node. That is why Slamm-O! went to the trouble of obtaining an AID that was not his. However, when he tries to spoof a Slaved Node, he is directly routed back to the Master Node, because the SLAVE DOES NOT ACCEPT ANY OTHER COMMUNICATIONS. They MUST come from the MASTER NODE. Because the Hacker has been routed back, the Slave does nothing. You must connect from the Master to the Slave. To do so requires a 2-Step Process, as I indicated above.

I have provided said rules twice now. They are in posts above.

Posted by: Tymeaus Jalynsfein Jun 3 2011, 07:13 PM

Ironically, these are the same rules that I have already posted twice now. There is NO 4th Option listed.

And no they do not. You cannot DIRECTLY SPOOF a Slaved Node. UNLESS You have direct, physical access. In the examples we have been bandying about, there is NO direct physical access.

Posted by: Tymeaus Jalynsfein Jun 3 2011, 07:15 PM

Very, Very True.

Posted by: sabs Jun 3 2011, 07:17 PM

Option 3:
You can Spoof the AccessID of the MASTER NODE, and then spoof commands to the slave.

That's pretty clear cut
If you have the Access ID of the Master Node, you can totally spoof the slave.

Non Slaves:
You send a command to a device or agent, pretending it is from an authorized source. You must have an access ID from which the target accepts commands (usually by making a Matrix Perception test on the authorized source or by tracing its icon). To spoof a target, choose a command (per the Issue Command action, p. 229) and make an Opposed Test between your Hacking + Spoof and the target’s Pilot + Firewall (System + Firewall for peripheral devices); if you succeed, the target accepts the command as legitimate. Each Spoof Command action applies only to a single command; multiple commands require multiple Spoof Command attempts.

Posted by: Tymeaus Jalynsfein Jun 3 2011, 07:23 PM

But only FROM the Master Node, not from an independant node, because the Slave does not accept any connections except those from the Master Node. You cannot forge a connection except from the master node. Command Connections (The only thing a drone will listen to) need a Subscription. Unfortunately, you cannot forge a Subscription to a Slaved Drone with a Spoof Command. So therefore, you need to use the connection that already exists. The Master Node's connection.

Posted by: Ghost_in_the_System Jun 3 2011, 07:24 PM

Do you have even the vaguest idea of what spoof does? When you try spoof a slaved node using the master's access ID, the slave will accept them because the instructions look like they are coming from the master. That is THE ENTIRE POINT of spoofing. A device can't tell the difference (without winning the opposed check) from a spoofed command saying it is coming from the master, and an actual command coming from the master. All a command is is a packet of information sent out through the wireless matrix with an access ID attached to it which gives authentication. All a spoofed command is is a packet of information sent out through the wireless matrix with a fake access ID attached to it which gives authentication.

The rules you're posting just say that you have to use the master's access ID as opposed to any other access ID. So, if you have the master's access ID, spoofing a slaved device works exactly like spoofing any other device.


Read again, you cannot hack the slaved node without direct physical access. Spoofing is not hacking.

Posted by: Bigity Jun 3 2011, 07:27 PM

I'm not seeing the point of slaving then, if you need exactly the same information to spoof a node either way. I guess it would only make it unhackable, and no other benefits.

Posted by: sabs Jun 3 2011, 07:30 PM

No you're completely wrong on this one. Because, if you can only spoof FROM the Master Node, then option 2 doesn't make ANYSENSE WHAT SO EVER.

Lets look at each option:

1) Hack Directly to the slave with a +2 TN modifier, requiring a hardwire line.
2) Hack the Master Node
3) Spoof the Access ID of the Master Node and then spoof commands to the slave.

If you've hacked the Master node, you don't need to spoof anything. You're already there.
When you use Spoof with the ACCESSID you're actually pretending you HAVE the subscription. That's how spoof works.

Remember, the subscriptions are 'virtual' they're in a table somewhere, linking access ID with subscription information. When you spoof the correct AccessID it thinks it's comming from the Master Node.
That's what having the AccessID allows you to do.

Posted by: Ghost_in_the_System Jun 3 2011, 07:31 PM

The benefit is that there is only a single access ID that works, as opposed to a large number. The other benefit is of course the unhackable thing.

Slaving is not "I win the matrix" it is a tactic to delay intrusion. Also, for a large part, you're going to want to hack into something, not simply spoof it, so you'll need to go through all these steps to be able to spoof the slave, then you'll need to use spoofing to break the connection or switch you to being the master. Then you'll need to actually do the stuff (possibly with hacking if you didn't make yourself the master). So yeah, slaving buys time against hackers (Which is often something very important) but it doesn't grant immunity to them.

Posted by: Bigity Jun 3 2011, 07:32 PM

But how is that different from spoofing any old node? Basically that way slaving has no benefit at all against spoofing, which seems odd.

Which really means nothing to a hacker, because they will have to sniff out an access ID either way. I guess slaving in this case would make it harder to get an access ID out of someone with legwork?

Posted by: Bigity Jun 3 2011, 07:34 PM

double post

Posted by: Tymeaus Jalynsfein Jun 3 2011, 07:35 PM

And you continue to go in circles.

Yes, I know exactly what Spoof Does. However, Spoof begins to break down when Slaving is used. Why? Because you cannot forge a link with your damn spoof command against a Slaved Node. You must use the Master Node (and the Master Node's AID) to do so. Yes, You can Hack the Node, and subsequently have complete access to everything; OR you can Spoof the Master Node as a Legitimate User (Acting as them, remotely), and then Sppoof Commands form there uisng the Master's AID. That is exactly what the Slaving rules tell you. Since you are not hacking the Master Node, using the AID for the Spoof Action does not boot you, because you are not putting another persona/agent/program onto the Comlink/Nexus, you are using it as a bouncing point. Once you have spoofed your access, then you spoof the slaved node (Uisng the same AID). Now, the slaved node will take the command and run it.

When a Node is NOT slaved, then any communications with the correct AID will work, because it is accepting such communications (as the Slamm-O! example provided above indicates). Any external communications with the correct AID will cause the Target to react. Unfortunately, if it is slaved, you cannot just connect, becuase your initial try will just shunt you back to the Master Node. You are NOT connecting to the Drone (in this case) node, you are connecting to the Riggers Node (Comlink or Nexus), and a Comlink/Nexus has no Pilot Program to execute the command (in the Drone Example).

Spoofing is useful for some things, and Hacking is generally more useful for all things. Spoofing IS NOT HACKING... In fact, You would likely not even be able to analyze the Master node that you have spoofed to determine any slaved subscriptions, because you cannot run any such programs on the master node with a Spoofed ID, because you are NOT ACTUALLY IN THE NODE. You have to hack it for something like that.

Anyways...

Posted by: suoq Jun 3 2011, 07:36 PM

Note that in unwired, a slave node is considered as running under admin. In unwired attempts at spoofing that require admin get -6 dice. So, in the above example, if the drone was slaved, Slamm-O! gets -6 dice to his spoof attempt under Unwired's rules.

Posted by: sabs Jun 3 2011, 07:37 PM

Because for example any old node might be setup to accept any number of access Id's.

Lets say you have a Node that controls a mechanical arm on your warehouse management system.
It accepts commands from: the maintenance guys access ID's, the access id's of all the security riggers, the command and control node, the safety manager.

If you slave it to the Command and Control node, then the ONLY way to control it is via the Command and Control node. The Security Riggers would need to connect via the command and control node, and sends their orders from there. The maintenance guys can't do any maintenance on site, etc...

Posted by: Tymeaus Jalynsfein Jun 3 2011, 07:37 PM

Never said it was... Slaving is a tactic to make the Target more difficult to hack. That is really all it is used for. Slaving has a nice side benefit of also making it a bit more difficult to Spoof.

Posted by: Bigity Jun 3 2011, 07:38 PM

Ah right. Thanks for pointing that out

Posted by: sabs Jun 3 2011, 07:40 PM

You do not need a subscription to the node, for spoofing to work. That's actually the whole point. But spoofing only lets you send 1 way commands. It's of limited use against anything but an agent or a drone.

You do not need to spoof the AID of a user, and send that to the master node. What you do is spoof the AID of the master node, and send that to the slave. It's for once actually fairly clear english.

Posted by: suoq Jun 3 2011, 07:40 PM

Apparently anyone at your table must use the Master Node, but (trust me on this one), we don't. Clearly we're not getting through to you either through sabs clear line of reasoning "If you've hacked the Master node, you don't need to spoof anything." or Tiger Eye's example of spoofing.

At this point it's becoming a pointless shouting match, all the way down to the "damn"s.

Posted by: Tymeaus Jalynsfein Jun 3 2011, 07:43 PM

It makes perfect Sense. SPOOF is a poor Man's Hack, and that is ALL it is.
Yes, If you are Hacking the Master Node, you don't need to spoof. Again, Spoofing is not hacking.
For a Regular Node, you are correct. You pretend the orders ocme from the controller. Unfortunately, against a Slaved Node, You are not pretending that you have a subscription, because you do not have a Subscription. You must use one that is already in place.

AS for your ideas on the Subscription List, It is only Virtual, ON THE LINK IT HAS BEEN CREATED ON, not floating somewhere in the Matrix. It is specific to each Master Device. If you do not have access to that List, well, then you are SOL. You either need to hack the Node to acquire the List (which changes constantly, I presume, as things are added and deleted from the list), or you spoof the Target with the AID of the USER, or for a Slaved Node, You spoof the Master, then Spoof the Slave (with the Correct AID).

Posted by: Tymeaus Jalynsfein Jun 3 2011, 07:44 PM

This is correct... In the Frst Paragraph, the node is not slaved at all. In the 2nd, it is. What is the Problem here?

Posted by: sabs Jun 3 2011, 07:47 PM

The subscription list is on the SLAVE.
it says: AccessID 8763 is my master, I must listen to it.
I spoof directly to it: I am AID 8763, execute shutdown command.

If my spoof roll is good enough, the slave sends back to accessID 8763 okay (but I don't see that) and begins shutdown procedure.

The slave has a list of all nodes that have a subscription to it.
The Master has a list of all nodes IT has a subscription to.

To fool the slave, you only need an AID on the access list. You don't give a damn about the Master.

Posted by: Tymeaus Jalynsfein Jun 3 2011, 07:47 PM

Which is what I said. They are two different Scenarios.

One is for a Normal (Non-Slaved) Node. Spoofing works great against Non-Slaved Nodes.
Two is for a Slaved Node. You can only spoof the Slaved node from the Master Node, however, as it is the only one with a valid connection, all others will be refused.

Posted by: sabs Jun 3 2011, 07:48 PM

That there are some job functions where that's not useful. Where that level of security is TOO MUCH for the job needing to be done. You don't want everyone accessing the Control Arm to have Admin access to the damn thing. You want the maintenance guys to only have access to the maintenance functions, you want the safety shift supervisor only to have access to the emergency stop.

Posted by: sabs Jun 3 2011, 07:50 PM

Except you're wrong! On the Internet!
OMG

I'm sorry but Scenario Two is not RAW. That maybe how you play on your table, but it's completely not RAW. And it requires a fundemental misunderstanding of the Slave/Master and Spoof rules to come to the conclusion you have.

Posted by: Tymeaus Jalynsfein Jun 3 2011, 07:52 PM

I think you confuse Subscriptions with Slaving. All your arguments work for standard Subscriptions that are non-slaved.

There is NO ACCESS LIST for the SLAVE. There is Only 1 Connection that it recognizes.

For standard Nodes, You would be correct. Any member on the Access List could send commands to a non-slaved connection. For a Slaved Connection, YOU ONLY HAVE A SINGLE CONNECTION, and commands are ONLY accepted through that connection. It is a Direct Link. A Slaved node will accept no other links whatsoever, because any attempt at such reroutes them directly back to teh master Node.

Spoofing a Slaved Node is difficult. Spoofing a Non-Slaved Node is cake.

Posted by: Tymeaus Jalynsfein Jun 3 2011, 07:54 PM

But that is a Security Choice. Many things will likely not be slaved, as it is intensive and requires more effort, even for those intended to use the system. Just becasue it is less useful does not mean that it is of no use. Some things will be Slaved (I ALWAYS Slave my drones, because it is so much more difficult to get into).

I understand your concerns, but that is not how the rules have been written.

Posted by: Ghost_in_the_System Jun 3 2011, 07:55 PM

What is a subscription? A subscription is Device A having somewhere on it something saying "I have a subscription to device B" and device B having somewhere on it something saying "I have a subscription to device A" and them sending data back and forth to confirm this.

What is an Access ID? An Access ID is something saying that Device A is Device A, and not Device C.

What is a command? A command is a packet of data sent out into the matrix that contains 1) instructions to do something and 2) an Access ID to act as verification that the instruction should be followed.

What is a spoofed command? A spoofed command is a command which has 2) from above artificially altered.

What happens when a device receives a command? It checks the Access ID provided by 2) against the list of Access IDs that it should accept commands from. It then looks at the command and checks that the Access ID provided by 2) has the proper level of authority to perform the command. If this checks out, the device performs the command.

What happens when a device receives a spoofed command? Exactly the same thing as a non-spoofed command, except that the checking of the Access ID provokes an opposed check to see if the device believes the forged Access ID.

What happens when a slaved device receives a command? Exactly the same thing as any other device, with the exception that the list of accepted Access IDs only contains a single possibility, and the additional instructions that if the Access ID doesn't match, it send a warning of some kind to the Access ID that is listed.

What happens when a slaved device receives a spoofed command? Exactly the same thing as a non-spoofed command, except that the checking of the Access ID provokes an opposed check to see if the device believes the forged Access ID.

In no way is it required that the master be involved in spoofing except that its Access ID needs to be found, and indeed the idea of sending a spoofed command from the master is ridiculous, because if you can send a spoofed command from the master, you could send a normal command from the master, which would be accepted. It is also ridiculous because to do that, you would first have to hack into and gain total control of the master, at which point you have total control of all slaves regardless.

Also, to whoever said slaving doesn't provide much advantage against spoofing, you're largely right. But nothing claims that slaving is an improvement against spoofing. Slaving is put forth as a security against hacking, which it is, and it does help somewhat against spoofing due to the decreased accepted Access ID list, and the fact that failure is more directly acted upon.

Posted by: sabs Jun 3 2011, 07:56 PM

the slave HAS to have an access list. Otherwise it doesn't know who to send all it's requests to. Or How to send them.

That Single Connection isn't really a single connection. (unless you use a wire, and turn off wireless)
The Slave gets a signal on the wireless, and it says, "are you AID 7843" If the answer is yes, it accepts the input, if it's no it says, "please contact AID 7843, rerouting connection".

So, if I happen to be able to spoof AID 7843, then, when I send my command 'as AID 7843' the slaved node says, "oh.. okay." But when it goes to send a response, that gets sent to AID 7843 as registered on the ad hoc network, which I can't do, unless I have 'permanspoofed' my AID on my commlink. (which causes some issues)

EDIT:
Ninjad by Ghost in a more coherent fashion.

And my point was that, you can't slave everything a s asecurity concern. Somethings do get slaved, but not everything under the sun.

Posted by: Tymeaus Jalynsfein Jun 3 2011, 08:01 PM

Sorry, I am enjoying the Discussion, and I have been using caps, because it faster than other formatting... Sorry for the "Shouting."
Not everyone likes penetrating a Slaved System. I get that. If you have HACKED something, then the Spoof Command is totally useless (Unless youwere detected, at which point you must spoof everything). I agree with that. But there are others who only want to just Spoof a command or two, and Hacking is not needed for that. Unfortunately, a Slaved node presents certain obstacles that MUST be bypassed. There are only 3 ways to get around them, after all, and they are very explicit. Unfortunately, you need to also understand the nature of subscritptions and slaving in concert, if you want to Spoof through the Master Node. Not saying that you don't, but apparently, you do not like how they interract, and that is quite okay.

Again, Tiger Eye's Example shows that you can indeed spoof a non-slaved node with ease. I never argued that you could not.

Posted by: Yerameyahu Jun 3 2011, 08:01 PM

I don't see why it's so hard. Slaving requires admin access, so spoofing a slave is at -6; not easy, but not impossible. You can attempt it as long as you're within signal range. There's no 'spoofing *through* the master', only spoofing *as* the master. A subscription is not a prerequisite.

An encrypted link does require a subscription, and it seems like you should easily be able to set up an encrypted slaved link and be done with the whole question.

Posted by: Tymeaus Jalynsfein Jun 3 2011, 08:06 PM

The slave does not need a list, all it needs is the Master Node's AID to reference. (I guess a list of 1 is still a list, though).
And since the Connection IS a Subscription directly with the Master node, it is as good as hardwired, in any event (Any communications must travel along that specific port, or it will be rerouted). Now, a non-slaved node can accept any other subscription request it likes or receives. Unfortunately, a Slaved node cannot. It will re-direct any such request back to the Master node, which may or may not accept such a request, depending upon its configuration. You cannot force a subscription request onto a Slaved Node.

I agree that not everything is going to be Slaved on a network. That creates an unwieldy system in a lot of respects.

Posted by: Tymeaus Jalynsfein Jun 3 2011, 08:08 PM

A subscription is a requirement, because a Spoof Command issues a Command, and only subscriptions can support such things according to Unwired.

Subscriptions and Data Requests are two completely different animals.

Posted by: Yerameyahu Jun 3 2011, 08:10 PM

I dunno. If a subscription is a requirement, spoofing anything is impossible. I read that bit about commanding drones (not devices) as referring to a persistent "sensible" connection that you *maintain* with your drones. It's not the same as being able to send a single command to a given device.

Posted by: sabs Jun 3 2011, 08:13 PM

Except that subscriptions are only for two way links. They are a Fast, two way, maintained connection. Completely unnecessary for Spoofing.

Posted by: Yerameyahu Jun 3 2011, 08:14 PM

Right.

Posted by: sabs Jun 3 2011, 08:16 PM

TJ: Normally I think you have solid grasp of the SR rules. But I think in this one, you have a fundamental disconnect.

Posted by: deek Jun 3 2011, 08:18 PM

I see where TJ is going on this. Its kind of like a VPN connection. When the master and slave negotiate the connection, they are authenticating each other and making a secure tunnel between each other and only traffic using that secure tunnel, is valid.

This assumption, I think, is more between the lines then RAW. We can certainly argue both points. TJ is supported by the book saying a slaved node can only accept commands from the master node and the text saying you need access to the master node to send commands to the slave. That "feels" like there is a secure connection between the two that cannot be spoofed.

Ghost has a case, because nowhere in the book does it say spoof can't spoof, well, anything.

I think it comes down to how powerful you want to let the spoof command be and how tough you want to make slave nodes from being hacked...

Posted by: Ghost_in_the_System Jun 3 2011, 08:20 PM

I think it comes from the fact that it says you need to spoof the master then spoof the slave. It is using two different definitions for the word spoof. "Spoof the master" means 'spoof' as in 'pretend to be' while "Spoof the slave" means 'spoof' as in 'send a fake command'. When put together you 'pretend to be the master then send a fake command' which is covered in a single spoof action.

Posted by: sabs Jun 3 2011, 08:20 PM

except that VPN is not a slave connection.

VPN is an encrypted connection, which is something completely different.
The way to cockblock spoof, is of course to encrypt your slave connection.

Then you need to:
Sniff the traffic
decrypt the traffic
spoof the AID
encrypt properly
spoof your command using your spoofed aid and the new encryption.

Posted by: sabs Jun 3 2011, 08:22 PM

the exact wording is:
You can spoof the access ID of the master node, and then spoof commands to the slave.

It does not say:
You can spoof the access id fo someone with access to the master node, and spoof commands to the master that will be relayed to the slave.

Posted by: deek Jun 3 2011, 08:22 PM

You would still need to use spoof if you didn't have a security or admin access on the master node. It kind of goes back to what I was saying. You could hack into the master node as a user and then spoof commands to the slaves. Or, you could create a legitimate account in the master node and just send commands without fear of setting an alarm. Hacking the master node for a security or admin access to begin with is making it more difficult to get in without setting off an alarm.

Posted by: sabs Jun 3 2011, 08:24 PM

READ page 55 of the book. It doesn't say any of those things. It says, hard hack the slave, hack the master, or spoof the slave.

You could also user hack the Master, and then use spoof to send commands to the slave.

Posted by: Ghost_in_the_System Jun 3 2011, 08:25 PM

How is this tunnel created? What prevents another device from sending something that looks like it came from the master? What method of identification above and beyond Access ID that cannot be duplicated is being used to confirm that something actually came from the master and not another source?

Now, I did just realize that for added security, you could encrypt the connection between master and slave, which would then require the encryption to be broken so that the hacker could properly encrypt the spoofed command so that it gets recognized properly by the slave.

Posted by: Ghost_in_the_System Jun 3 2011, 08:30 PM

Almost correct, but you don't need to spoof your AID as a separate action, that is part of what spoofing a command does.


That doesn't change anything. Just replace 'master' with 'access ID of the master node' and you still get 'pretend to be the access ID of the master and then send a fake command' which is still a single spoof action.

I never said anything remotely approaching that.

Posted by: Bigity Jun 3 2011, 08:31 PM

Sometimes, just sometimes, I miss the days when everything in the matrix was very standard, and you had rectangle data stores and hexagon CPUs and little triangle access points

Posted by: Bigity Jun 3 2011, 08:31 PM

double post again, freaking internet here is bonkers today.


On a side note, is there a way to delete a post?

Posted by: sabs Jun 3 2011, 08:33 PM

Ghost, we're on the same side on this

I was talking to Deek and TJ who both DID say that.

Posted by: Ghost_in_the_System Jun 3 2011, 08:33 PM

I do admit, having a more abstracted system like that has its advantages, because you don't have people comparing it to real life. It also means somewhat less freedom though, so there is that.

Posted by: suoq Jun 3 2011, 08:41 PM

I just want to make sure we're down to the following impassible barrier.

Side 1: "Spoof the Access ID of the Master Node" = Spoof the Master Node
Side 2: "Spoof the Access ID of the Master Node" = Get the Access ID of the Master Node

Any other points of contention?

Posted by: sabs Jun 3 2011, 08:42 PM

No you got it in 1

Ghost and I think that #2 is the right interpretation, TJ and Deek think #1 is.

Posted by: deek Jun 3 2011, 08:48 PM

I was just adding some information to support TJ and make a case that it made sense and could be interpreted this way. I would say that the tunnel creation is a part of the slaving a node process. That's all part of the setup and the effects of slaving a node. Obviously, its not spelled out in detail, but some of the wording used in RAW supports that.

I could also see you bypassing the whole thing and taking sabs (I think) point and taking a -6 to your spoof if the target is a slaved node. Then you could send whatever command you want using spoof. Just at a -6 die pool. And, you'd also have to know the master nodes AID, which is easy enough to get. A -6 to spoof is likely not a trivial matter for most.

Posted by: deek Jun 3 2011, 08:52 PM

I'm not saying either is right actually. But I do agree those are the 2 points of intention.

I can see where TJ is coming from and can read into some of the text to support that. I can also see where Ghost and sabs is coming from.

Personally, I'd prefer my players to feel safer when they slave their comm to the master hacker and not have to worry about being spoofed. That puts the spotlight on the hacker if someone is screwing around with hacking comms. And, I also feel that was the intention of the rules introducing slaving as an option. Not to make anything unhackable, but make it a single point of entry on a assumedly harder target.

PS: I do like the tag team debate going on here. It is rather enjoyable.

Posted by: Ghost_in_the_System Jun 3 2011, 09:08 PM

You need to define spoof in side one, because spoof has two major meanings 1) pretend to be 2) send fake command with fake credentials to. So, to further clerify:

Side 1: "Spoof the Access ID of the Master Node" = Send fake commands to the Master Node
Side 2: "Spoof the Access ID of the Master Node" = Get the Access ID of the Master Node (so that it can be used as part of the "Spoof a Command" action)
Side 3: "Spoof the Access ID of the Master Node" = Use Spoof to make yourself temporarily have the access ID of the Master node (And yet still have to spoof commands instead of sending regular commands)

I'm not entirely sure if TJ is on side 1 or 3 or both. I know that I am firmly in 2 because that is how the rules for spoofing a command work, and nothing in slaving says that spoofing works any differently from normal.

Note: I believe you would suffer the -6 penalty for the spoofing because you are using an admin account.

Posted by: Bearclaw Jun 3 2011, 09:14 PM

I think you are misunderstanding what spoofing is.
In Shadowrun, spoofing and hacking are two different things. In real life, of course, spoofing is one of the main tools of a hacker.

When you spoof, you are attempting to convince the recieving node that the sending node is a different node, such as the system that it's slaved to. It already has a subscription.
If you were to have a subscription you would have had to hack it, and there would be no point in spoofing.

Posted by: Tymeaus Jalynsfein Jun 3 2011, 09:49 PM

Again, I am not confused here. I do know that the two tasks are completely different in Shadowrun. You cannot treat both Normal Nodes and Slaved Nodes the same for purposes of Spoof. And the Rules support that.

And Ghost_in_the_System, Essentially Option 3 above (which is exactly what Unwired says). You are creating a temporary access remotely (using the Master Node AID) that will use the Master Node to communicate with the Slaved Node. The reason that this is necessary is because the Slaved node will not accept any other connections except that of the Master Node. You then use this temporary poiont of access (the Spoofed Master Node) to send that Spoofed COmmand to the Slaved Node. It is NOT a Hack, because you are not actually implanting anything on the Master Node at all. You are using its own protocols and subscriptions to gain your own ends.

The reason the Slaved node will not accept commands outside of the Subscription port that is subscribed, is because the Slave will forward ANY other connection attempts right back to the Master node. This is a result of how Subscriptions and Slaving interact. You cannot generate a Command connection without a successful Subscription. Slaved Nodes ONLY accept a such commands through the Subscription that the Master Node has established for itself (as it only needs the one).

Now, functionally, there are ways around that. The 3 discussed routes are the methods established on Page 55 in Unwired. Another possible method, which has been briefly talked about, is to completely disrupt communications in such a way that the subscription is disconnected. Then you can attempt to re-establish the connection using your own hardware and the Spoofed AID of the original Master Node. Thsi will, in theory, work, if you can pull it off. It is harder than you think it is to do so, however.

The only Surefire ways to do so are the 3 ways described in Unwired.

In any case, you should still suffer the -6 penalty to the Spoof attempt upon a Slaved Node becasue you have to do so with Administrative Access.

I continue this debate because, otherwise, Spoofing a Slaved node is no harder than spoofing a non-slaved node, and this is obviously NOT the intent of the rules as presented in Unwired. The Story in the SR4A Book, with Slamm-O! and Netcat, only uses Basic Rules, which does not include Slaving. If they were the same difficulty, then why have slaving at all, as there would be no functional difference?

Anyways...

Posted by: Ghost_in_the_System Jun 3 2011, 09:52 PM

Sure you can. Tell me where it says you can't. Spoofing is not a connection, and does not require a connection, so don't try pulling out the connection thing again (Unless you can also tell me where is says that spoofing is or requires a connection)

Posted by: Tymeaus Jalynsfein Jun 3 2011, 10:08 PM

Yes it does, Spoofing a Slaved Node requires a connection to the Slaved Node, otherwise you cannot spoof it. That is the benefit of Slaving.

Posted by: Tymeaus Jalynsfein Jun 3 2011, 10:10 PM

Whoops... Double Post...

Slaving is a big benefit against Spoofing, and a Poor one agains actual Hacking. Why? Because if you hack the system, you do not have to hack ANY of the Slaved Nodes. You just command them. Unlike a normal system, where you may have to hack each node in sequence (some are set up that way, and some would not be). Slaving removes completely the need to hack into the slaved nodes.

Posted by: suoq Jun 3 2011, 10:22 PM

Just to be clear, I'm in the #2 camp. But I follow the "Thank you for Smoking" rules of debate.

I can't convince TJ and as far as I can tell, everyone else supports or at least understands the #2 camp so there's no "them" to convince.

Posted by: Tymeaus Jalynsfein Jun 3 2011, 10:28 PM

And that is okay...
And for the record, I never said I did not understand Position 2, just that I disagreed with it, and why.

Posted by: Ghost_in_the_System Jun 3 2011, 10:54 PM

Where does it say this?

Posted by: Tymeaus Jalynsfein Jun 4 2011, 12:10 AM

Slaved Nodes require a Subscription... It is on Page 55 of Unwired, Middle of the Page, in the Box that describes exactly what needs a Subscription. A Subscription is a Direct Link Connection.

Posted by: Ghost_in_the_System Jun 4 2011, 12:13 AM

It requires a subscription to be slaved to something, it doesn't require a subscription to be spoofed.

Posted by: Tymeaus Jalynsfein Jun 4 2011, 12:18 AM

You cannot Spoof a Slaved Node directly, Because it will not accept your communications attempt. The ONLY Communications it accepts are along the Subscription. Which is a Direct Link.

Spoof, by the basic book needs nothing else, because the bassic book knows nothing about slaving. In Unwired, there are only 3 ways to Spoof a Slaved Node. We have gone over those. Is this more difficult than a normal Node being Spoofed? Somehwat, though not heavily so. Is it often better to just Hack that Master Node in the first place? Most of the Time, Yes.

As I indicated before. Slaving gives more security against actual Spoofing, but removes a bit of the Security for Actual Hacking. There are tradeoffs.

Posted by: Ghost_in_the_System Jun 4 2011, 12:32 AM

I'm trying to think how I can possibly break this down any more simply.

A connection is established between A and B, if you spoof a command pretending to be B, it looks exactly like anything else that comes from B, and so is impossible to distinguish from 'communications along the subscription'. Also, a direct link is not required for a subscription, nor does having a direct link matter at all in this instance, because a direct link is no more secure against spoofing than a routed link.

We have not gone over 3 ways to spoof a slaved node, we've gone over three ways to gain access to a slaved node (You are once again mixing up terms). And there are no additional requirements set out in Unwired about requiring anything like a subscription or connection (which existed in BBB, and so should have been mentioned there regardless), despite an entire section being about spoofing.

And no, slaving is actually designed to make hacking more difficult, not less. And slaving makes spoofing... mildly more difficult, but not excessively so by virtue of a limited Access ID list and required use of an admin account.

You also have to realize that your interpretations are... asinine at best. In order to spoof a slaved device you first have to hack the master? Which means you already have complete control over the slave, so there is no reason to do so. Or something about... I don't know, bouncing the spoof off the master for which there are no rules for? I don't quite remember as I didn't quite understand at the time.

You also claim that there is some magical force that distinguishes a spoofed command from a legitimate command, despite that being exactly the entire purpose of spoofing a command.

Posted by: Tymeaus Jalynsfein Jun 4 2011, 12:37 AM

Your right... 3 Ways to Access a Slaved Node, Not Spoof a Slaved Node...
As for Slaving and for Hacking Difficulty, all it does is use a more robust system instead of the slaved system. Unfortunately, once you have cracked the Master Node, all slaved nodes are available with no further hacking required whatsoever. It makes Spoofing More Difficult, in the end, and Hacking Less Difficult in the End...

And I never said you had to hack the Master Node to get to the Slaved Node. Way to twist the words there. I said it is likely more efficient to hack the Node than it is to Spoof it. Completely different thing.

Quite twisting what I said... Especially since you know that it it NOT what I said.

And you and I are never going to agree on this one. So, No worries...

Posted by: Ghost_in_the_System Jun 4 2011, 12:42 AM

I'm not exactly sure how having to hack a more secure system to get at a less secure system makes hacking easier, unless you happen to care about the other slaved node, which you may or may not, and which may or may not even exist. It is harder to hack into a given node, but easier to hack the system as a whole.

And maybe I've gotten what you've said mixed up. Would you mind explaining to me what exactly you do have to do to spoof a slaved node then? Every single action, target, and requirement.

Posted by: Tymeaus Jalynsfein Jun 4 2011, 12:52 AM

The Slaved nodes are less secure in the long term, because once you have hacked the Master Node, they are wide open. If they were not slaved, you would potentially have to hack each and every node in succession to access them. When they are slaved, this is not required. SO you are trading a Heavily Defended Node for the Initial Hack, hoping that it will stop the hacker, but once he is in, any slaved nodes are easily accessible, since they are not required to be hacked.

Accessing any Node that is slaved has only 3 options. The 3rd is likely the easiest. To Spoof that slaved node, you must first gain access to the communications port that the Slaved node is using to communicate with the Master node (which is on the master node). You sppof access to that master node, and then you spoof your commands to the Slaved Node through its Subscription. THIS IS NOT A HACK. It is simpler than a Hack at its core, but takes more steps.

YOU Cannot access a Slaved Node through Wireless protocols (or even wired ones) Unless you are the Master Node, because it will only accept communications From that Master Node, On the Master Node's Subscription. Not sure why that is so hard to understand.

Spoofing A Slaved Node is not the same as Spoofing a Non-Slaved Node. You are trying to make them the same.

Posted by: Ghost_in_the_System Jun 4 2011, 01:13 AM

You keep bringing up 'spoof access to' but that isn't an action of any kind. There are no rules for 'spoofing access to' a thing, so I don't know what kind of roll is required. Is it opposed, is it a threshold, is it extended? It is difficult to perform an action which does not exist.


Yes, it will only accept communications that appear to come from a master node, but there is no way to tell the difference (besides the opposed device check of course) between a spoofed communication set to look like it is coming from the master node, and an actual communication from the master node. A subscription does not provide some magical unspoofable communication between two devices. Really, all a subscription does is set up that two things need to be ready to devote large amounts of bandwidth to one another. Other than that, there is nothing special about a subscription.


Because it is the same thing, you are trying to make them different when there is not a single word in the entire book that says spoofing them is any different. The only thing even remotely close is the single line on p55 which I have explained above does not constitute multiple actions, or indeed anything outside the normal parameters of a spoof.

Posted by: deek Jun 4 2011, 02:21 AM

Again, I think this comes down to how powerful you want the Spoof program to be at your table. If you want it to control a slaved node by simply having he master node's AID and sending a spoofed command for every action, then you can do that. I think it makes slaving nodes a lot less secure than Unwired intended them to be, but that may fit your style better.

Now this whole argument only holds up if there is no way to tell the difference between communications coming from the master node and spoofed communication set to look like it is coming from the master node. And this is in the realm of feasibility for the spoof program.

I just think it weakens the whole purpose for slaving a node. I think the spirit of the rules are that there is a difference between communications coming from the master node and communications not coming from the master node, whether they are spoofed or not. I think the whole purpose of slaving nodes is to make it harder to hack them (which, you could argue, spoofing at -6 meets that purpose).

Basically, you are saying the only difference between spoofing a non-slaved node and a slaved node is the -6 penalty. That may be enough for most. But I like the feel of making the initial hack harder to get into the master node and once there, able to have access to all slaves by default. I think the flavor of that, as well as the practical gameplay of it make good sense. The team's hacker then gets to focus on defending just his own commlink against hackers instead of worrying about all his teammates having crappy defenses. And he knows that if he doesn't defend his node, then the whole team has the potential to get hosed (matrix-ly speaking).

One more thing, it seems that ghost and sabs assume that hacking the master node is always done at the admin level. I still disagree with that, as it is quite viable to hack that master node at a user level and then use your spoof program to send commands to the slaves (as you'd have to use spoof because user level access doesn't allow control of slaves). I think that is the easiest way in.

Obviously, if you do hack the node at admin level, then you don't need to spoof any commands, nor risk setting any alerts, and can just start sending regular commands to any slave you desire. But I doubt that many hackers will want to spend the extra rolls with a +6 for admin and risk setting an alert. You set off an alert in some hacker's master node, and you can pretty much be guaranteed he's got countermeasures against you as well as cut links to his slaves to minimize the effects of the breach.

Posted by: Ghost_in_the_System Jun 4 2011, 02:45 AM

Well, actually the -6 is just from the fact that you're spoofing to an admin account, which could occur on any device, not just slaved devices, it is just that with slaved devices, it is the only option. For instance, if you wanted to spoof a command to create a new account for you to log into a non-slaved device, that would require admin access, which would face the -6 penalty.

Also, being able to spoof the device doesn't necessarily give control of the slaved device, it simply allows you to perform one command at a time. While that command might be able to bring over control, it doesn't necessarily. Also keep in mind that by encrypting the connection, you could easily force a hacker to spend several turns just decrypting the encryption, then applying the encryption to her spoofed commands. Slaving a commlink does make it more resilient to spoofing, just not as resilient as it does to hacking.

As for admin access on the master node, well, depends. Admin access may not be required to give commands to slaves, or it may, depends on the setup. And really, what you're doing is your going back to SR3 rules with your 'break in so that you can get access to the secondary devices'. There aren't gates like in the wireless system. Those can work on hardwired systems, because you can detect if a command is actually coming through a particular line or not. With wireless though, it is all information being broadcast out into the ether, and being plucked out of it. There is no true way to verify where it came from beyond it telling you as much, and faking that is exactly what spoof is all about.

Personally I think that spoof being able to bypass the master after a fashion makes the hacker's job more interesting. She now has to consider more than simply getting the biggest firewall with an IC that runs analyze constantly. She now has to encrypt connections and do her best to keep tabs on if people are trying to sneak past her. And if they do, well, now she suddenly has something to do, go find the theif, and hack into their system to retrieve the now lost commlink.

Hmm, just thought of an analogy. Pretend hackers are cattle rustlers. Normally cattle just stand around and are fairly vulnerable to being rustled by someone throwing a lasso over them and being dragged off (Hacked). With slaving though, you have them all on ropes being held by a ranch hand. The rustler now has the option of going right up to the cattle and attempting to remove the rope from its neck(hacking via a physical link), beating up the ranch hand and taking all the cattle's ropes from him (Hacking the master), or sneaking behind the ranch hand, cutting the rope, and then slipping away with the cattle (Spoofing the slave). Basically the rope is the master/slave connection. It doesn't matter much to the cattle who is actually on the other end of the rope, so long as it seems like it is the same person, which it must be because there is still that same rope. A cow can't distinguish one human from another (Okay, maybe it can, but for the sake of the example...).

It may not be a 100% accurate example, but I think it fits pretty well (And who doesn't love a strapping cowboy? )

Edit: In the above example, the cowboy would have his back turned on the cattle, with the ropes and possibly mooing being an indicator that something is wrong, just like a master device doesn't do much with the slave device unless something is actually wrong.

Posted by: Yerameyahu Jun 4 2011, 03:18 AM

This is pretty silly. There's no reason the list would include 'spoof the slave' if you couldn't do it (that is, if it required you have access to the master). As I said way back, if you want a dedicated 'locked' master-slave connection, encrypt it. Otherwise, the whole point of spoofing is precisely the reason that spoofing works: it pretends to be from the master. It doesn't have to actually be from the master, because it's spoofing. That's the whole point.

Posted by: deek Jun 4 2011, 11:35 AM

Which again, is an interpretation and setting a certain level of power to the Spoof program.

You could also take the phrase, "only accepts communcations from the master node" and interpret that as TJ has already done.

Please understand, TJ has said several times he understand what the argument is (as do I), but we don't agree with your interpretation. Which, really doesn't affect anyone at all unless you are playing at someone's tables that run a different rule for spoofing slaved nodes.

Ghost, I do like your analogy. And to add the "other" interpretation to your analogy, the cattle are covered with a giant tarp (meaning you can occasionally here a moo, but you really can't see any cattle, but you kind of know they are there) and you can't access any of the ranch-hands ropes until you are on the ranch (i.e. in the master node).

I'm with TJ on this. I understand what you are saying and I think its a valid interpretation. I just think that there is more than one valid interpretation here, based on the wording in the books.

Posted by: Yerameyahu Jun 4 2011, 01:05 PM

No, you really can't 'interpret' it that way. The way it knows who the communication is from is the Access ID, which spoofing spoofs. That's why spoofing exists, to appear to *be* a communication from the master.

It's not 'setting' the power of anything. The book has 3 options, and one of them is nonsensical if you change the rules in the way you're describing. Again, to require master node access is literally option #2 of those three option *already*. There is no reason why the book would offer the spoofing option at all if it was identical (or identical-but-inferior) to the 'access the master' option.

I appreciate your attempts to be evenhanded, I really do, but not everything is an opinion.

Posted by: Tymeaus Jalynsfein Jun 4 2011, 01:14 PM

Since I have had a reasonable amount of rest last night, Let me try this Tack...

1. You have a Slaved Node
2. You want to spoof Commnads to that Slaved Node.
3. You do not, yet, have the Access ID of the Master node.
4. You Trace a User on that Node to get Their AID to access the Master Node.
5. Now, you use the Users AID to Spoof a Command to the Master Node.

Now, why do you do that? Notice that the rule does not say that you Trace the Master Node to get its AID, because that is already assummed. Let me put it back up here again. It says...

So, why are you Spoofing the AID of the Master Node? I interpret this as a direct, temporary connection through the Master node so that you may communicate with the Slaved Node. Why? Because the Slaved Node will accept No New Connections from any nodes. Why? Because it already has a Direct Linked Connection, Via the Subscription that is already in place (See Subscription Rules, and implement Slaving Rules). It does not need to accept another connection, and in fact, will not accept any pther connection whatsoever.

6. Ultimately, you want that command to go to the Slave, but the Master Node will not Run that Command on its own, so...
7. Once the Master Node has been Spoofed, you grab the Master's AID and you now Spoof the Slaved Node, using the Master's AID.
8. The Slaved Node does what it has been commanded to do.

It is a completely different interpretation than you use, Ghost_in_the_System, because it has to be. Otherwise your Slaved Nodes and Your Standard Nodes are treated identically for purposes of Spoof, and that is not the intent, as outlined in the rules above.

Do both systems work? Of Course they do. But by using the exact same system for both types of Nodes, you have just invalidated the Slaving Rules. Why? Because you already Must have an Appropriate AID for the Node to be Spoofed. The reason a Slaved node is HARDER too Spoof than the Standard Node, is because the Slaved Node will not accept any other outside communications whatsoever, which is not true for the Standard Node.

Hopefully this is a bit more coherent than my sleep addled ramblings from last night...

Anyways. It really is academic, as Deek stated, because each table is going to use the rules as they see fit, no matter any other's interpretations.

Posted by: Yerameyahu Jun 4 2011, 01:29 PM

You're spoofing the master's ID:

That's simply not how subscriptions work in the wireless world. The Access ID is exactly (and solely) how it knows that incoming communications are from the master. Spoofing is not a 'new connection'.
This doesn't make sense, not as far as "spoof the access ID of the master node" is concerned. It doesn't mean (and doesn't say) 'spoof the master' nor 'spoof your way *into* the master'.

It is not the case the slaved nodes are supposed to be especially resistant to spoofing. It is not 'against the intent' for them not to be. Slaving has the purpose of not letting people *log in*. Your 'interpretation' is fundamentally denying the central function of spoofing and its reason for existing: to pretend communications are from the correct node or 'connection'.

Posted by: Tymeaus Jalynsfein Jun 4 2011, 01:38 PM

Actually it is not, but that is okay, you and I have different opinions. What else is new, right?

The Slaved node knows it is the Master Node for 2 rasons. It has the Master's AID, and it comes along an already established Subscription, through which it communicates with the Master Node. There are 2 parts there.

The problem with your argument is that you will not ever connect to that Slaved node wirelessly, (or wired without Physical access) unless you do so From (or through, if you will) the master Node, because the Master node already has a connection and a new one cannot be created while the Node is Slaved, Unless the Master Node Actually Creates It. You are completely ignoring that the Slaved Node will not allow further connections. At All. And since the Master Node already HAS a connection, there is no need to create a new one. So, unless you are using the Master Node's connection (that already exists), you are trying to create a new connection, which will be refused. You cannot argue that point, because that is what a Slave is supposed to do. Forward ALL Connection Attempts back to the Master. If you are not the Master, with an already Subscribed Link, you cannot form another one.

As I said. You have to look at the Subscriptiuon rules, as well as the Slaving rules. Anything else makes no sense.

Posted by: suoq Jun 4 2011, 03:50 PM

This is where you lose me every time. I don't see how spoofing an AID = spoofing a node. The problem is the rules say you have to "Spoof an AID" and there's no rules for spoofing an AID. You've decided it means "spoof a node" and we've decided it means "Get the AID (because we need it for the spoof action)".

Posted by: suoq Jun 4 2011, 03:53 PM

No. We're tricking the slave node into thinking our spoof is FROM that connection. It's a wireless signal. We're convincing it that our transmission is part of that connection. By spoof we mean we're trying to get it to authenticate as if it came from the right place, even though it isn't. We're using the right frequency, the right protocol, the right numbers, the secret handshake, etc. etc to make our wireless transmission appear to be an authentic wireless transmission from the master.

Posted by: KarmaInferno Jun 4 2011, 04:07 PM

There is no link between a slave node and master node that automatically tells the slave node that a communication is or is not from it's master node, besides Access ID.

It's not like some magic tunnel is established between the two nodes, that nobody else can breach. It's wireless. To send a command, a master node broadcasts a message packet into the ether stamped with it's Access ID. To receive a command, a slave node listens to the ether for message packets bearing the right Access ID.

A slave node receives a command. The slave node authenticates it, and if it's got the right ID it executes the command. As far as the slave node is concerned it DID receive a command from it's master node. That is the whole point of spoofing, to pretend to BE the master node.

Now, the master node may be able to tell the slave node is being spoofed, because the slave node is likely to send out a broadcast saying, "Command received". At which point the master node wonders what is going on because it did not in fact send a command.

Really, this should be easy to prevent, because it'd be trivial to establish a non-standard authentication scheme that does not follow any established procedures. Like altering your drones to only execute a command if it receives certain encrypted keywords along with the command, keywords that get switched with every communication. Or have the slave node only accept commands at specific changing intervals of time. But there's no rules for that kinda thing.


-k

Posted by: Yerameyahu Jun 4 2011, 04:21 PM

It's *not* a connection, as I told you. If you want a VPN (again, as I told you), you need an encrypted subscription. Otherwise, the *entire* purpose of spoofing is to pretend to be that original trusted source. The sole method slaving uses to accept/reject incoming communication is Access ID. There is no 'is this the magic connection?' test, because that's what Access ID is for. What makes no sense is for the book to give 3 ways to beat slaving, except the third one is the same as the second one (except worse).

suoq, in fact there are rules for spoofing an Access ID. I quoted them just now. It's not a required step for all spoofing, but it does save you time for multiple commands (and "commands" plural is specifically what the slaving-beating list mentions).

Those kinds of measures are indeed conceivable, though the FAQ deals with one such example. The basic idea is that you could spoof any relevant authentication, so it's just an escalating game of really annoying cat and mouse.

Posted by: Aerospider Jun 4 2011, 04:23 PM

Despite an innate dislike of them I believe I have an apt analogy to add. Jeff is a bouncer who's very handy on the door but has no common sense. The manager duly instructs Jeff to take orders from him and only him. Anyone else making enquiries must be referred to the manager's office, even if it only concerns Jeff. One day a magician uses his Mask spell to trick Jeff into thinking he is the manager. Does Jeff obey the spellcaster? Of course, because he only has to think it's his boss, it doesn't have to actually be his boss. How could Jeff ignore the order without resisting the spell?

In conclusion of this fairly arbitrary comparison, subscriptions/connections are not ethereal cables stretched out between nodes to keep communications exclusive. The only thing stopping a slave's node from being commanded is that it has orders not to if the command is not 'signed off' by the master. A spoofed command sent directly to the slave is simply a normal command with a forged signature. If you happen to have learned the required signature in advance you need not worry about the master at all.

Posted by: Ghost_in_the_System Jun 4 2011, 04:28 PM

But what is an already established subscription? I can tell you what it isn't. It isn't a physical link between two nodes. And if it isn't a physical link between two nodes, it can be faked through the wireless, because without a physical link, there is no way to actually physically determine the origin of one packet of data as opposed to another packet of data.

If a device has a subscription to Device B, and you send something that looks like it came from Device B, then it looks like it came along an already established subscription.

You still continue to fail to point out what kind of an action 'spoof access to' is, though you bring it up as a requirement



It's great that you interpret it that way, but there are absolutely no rules for doing something like that. None. Are you suggesting that half a sentence was intended to create a new action that has no rules for it what-so-ever?

The whole problem here is that you aren't allowing for the fact that the word spoof has multiple definitions. In the case of 'spoof an access ID' it can only have the meaning 'pretend to be' logically applied, because you don't spoof as in 'send a fake command to' an access ID, you do it to a node.


@deek Actually, my example already placed the rustler on the ranch. The ranch isn't the master node, it is just part of the matrix. The ranch hand is the master node. Interestingly TJ's idea of how things work falls apart completely with this analogy (because then it becomes something like 'first you have to put on the ranch hand's skin' or maybe 'You have to bounce a thrown knife off the ranch hand to cut the rope'). This doesn't make idea wrong, but it is something interesting to point out for an analogy that otherwise works fairly well.

Posted by: Yerameyahu Jun 4 2011, 04:34 PM

Spoof the Access ID of the master means one of exactly two things:

1. Actually alter your node's Access ID to match (per the rules I quoted above). This arguably allows you to just send commands, which is a quicker action than the Spoof Command action. (Note that, under normal hacking circumstances, logging in with a duplicate Access ID causes a security response; spoofing, however, is not logging in. Your persona is not present, and you don't have a subscription.)

2. The sentence actually means 'spoof the slave *as* the master', which is the simpler, normal method. While the wording is awkward, it fully makes sense to specify here that the only way to spoof the slave is as the *master*. Indeed, we should expect this.

In no case could this ever mean some unprecedented thing like 'temporary connection to the master and send commands through it'. That's called 'hacking the master', and it's *already* #2 in the list of 3 ways to beat slaving.

Posted by: Ghost_in_the_System Jun 4 2011, 04:35 PM

That's a very good analogy. When trying to determine who someone is, we can generally only tell by how they look and maybe how they sound or act, and the digital equivalent is an Access ID. Just like you wouldn't necessarily be able to tell the difference between Bob and someone pretending to be Bob with a Mask spell, a node can't necessarily tell the difference between something from Access ID B, and something from something spoofing access ID B.

The slave/master relationship could be viewed as the equivalent of someone pretending to be a friend of yours, instead of someone you don't really know particularly well.

Posted by: suoq Jun 4 2011, 04:41 PM

The RL term for that is "security through obscurity". It doesn't work as well as people want it to and it's expensive to maintain. It's not worth adding to the game because (let's be honest here) watching hackers hack in real life is even less exciting than watching chess players play chess. Players who use these sorts of things to protect themselves deserve to face these sorts of things as part of the challenge and they'll quickly get annoyed at having to do the legwork now required to hack the system. "The hacker will spend the first two days probing. Someone pick up some pizza while we roll dice."

Posted by: Bigity Jun 4 2011, 04:43 PM

That depends on what the manager tells Jeff to do, because Jeff is aware of magic and what it can do (even if artificially inflated by trids).

Nodes might not be able to use that kind of deduction though.

Posted by: Ghost_in_the_System Jun 4 2011, 04:51 PM

Well, I think the assumption is that Jeff isn't exactly the sharpest knife in the drawer, and is in fact much closer to a mallet. Still, you're right, still likely wouldn't obey the order to kill himself or something like that, but nodes don't have that kind of self-preservation logic or the ability to think 'hmm, does this really sound like something the master would have me do?' and question an odd order.

Posted by: Tymeaus Jalynsfein Jun 4 2011, 04:53 PM

Sure they do. IF they have been set to query each order with further authentication, then it fails, because the Spoofer does not receive the Authentication Query. It is even simple and easy to do.

Posted by: Ghost_in_the_System Jun 4 2011, 04:59 PM

Which is already part of the spoof program. That's why there is an opposed test in the first place. The original device requires some form of authentication (As stated in the books) already, which is part of what spoof has to fake. If a device didn't require authentication then it wouldn't get an opposed check against the spoof command.

Posted by: Tymeaus Jalynsfein Jun 4 2011, 05:07 PM

You miss my point... You spoof the COmmand (with appropriate AID and everything)... Node Receives Command... Node Re-Authenticates Command by requesting Re-Authentication, Spoofer does not receive Command, because he is not the Master Node. Master Node Gets Authentication Request... Master Node invalidates Request... Spoofed Command does not take place.

This tends to occur with Drones more than anything else, because the Pilot may have a Scrpt mandating this procedure, but a Normal Node may also have this script in place.

Yes, it is very ineffecient, becasue this takes multiple IP's to complete any requested action. But it can be done.

It really comes down to how difficult the GM wants to make the System/Node/Drone/Whatever to hack...

Posted by: suoq Jun 4 2011, 05:16 PM

You lost me there. The Slave sends out a wireless signal that the spoofer node antenna doesn't get but the master node antenna gets?

Posted by: Tymeaus Jalynsfein Jun 4 2011, 05:25 PM

That is indeed correct, because you only Spoofed the AID, you are NOT the AID...

EDIT: Here is a better comparison... You can Spoof Admin Level Commands; the fact that you Spoofed an Admin Level Command does not give you actual Admin Access Rights. If you had Actual Admin Level Access Rights, you would not have to spoof them.

Posted by: Bigity Jun 4 2011, 05:36 PM

You could change your commlink's access ID, and receive the authentication query that way I suppose, which just adds to the fact that slaved nodes are more of a pain to subvert?

Posted by: Tymeaus Jalynsfein Jun 4 2011, 06:19 PM

Usually... Once you physically change your own AID, you need to resubscribe all your subscriptions. As well, you cannot change your AID in the middle of a Matrix Run. You have to disconnect and then reconnect, which is generally a pain. That is why Spoof is so useful. Hopefully, you have no Agents out either, because once you change your AID, they lose track of you as well.

Anyways...

Posted by: Yerameyahu Jun 4 2011, 06:20 PM

The re-authentication Tymeaus is describing is specifically addressed in the FAQ, actually. That's what I was referring to.

Posted by: Tymeaus Jalynsfein Jun 4 2011, 06:23 PM

Ahhh... A FAQ reference... How is it worded?

Edit: Looked it up, so never mind...

Posted by: Tymeaus Jalynsfein Jun 4 2011, 06:51 PM

Okay Everyone...

I said I understood your points of view and I do...
Having not reviewed the FAQ in a LONG time (Mostly because of its inaccuracies), I have just done so. Thanks Yerameyahu...

I will have to agree with Yerameyahu (and by extension Ghost_in_the_System and Suoq.)

Leaving aside Options 1 and 2 for Accessing a Slaved Device, Option 3 is pretty Cut and Dried, and I will admit, my interpretation is (was) a little odd. And is obviously at odds with a lot of People. But there you go. However, after a review of the FAQ, I think I have determined where my interpretation went off course. So, let me clarify my current position.

Leaving aside that you can only Spoof Agents, Devices with a Pilot Rating (Drones usually), and Peripheral Nodes; You Must do two things to Spoof one of the Above Slaved Devices.

Well, 3 Things actually.

1st. You must obtain the AID of the Master Node. (Goes without saying, but did anyways to clarify my line of thinking)
2nd. You must set your hardware to emulate said AID. (This is tedious, and often not worth doing, as a Hack would take less time).
3rd. Then you may Spoof commands to the Slaved Node.

Note: You CANNOT SPOOF a Slaved Node that is not any of the above (because it is, for all intents and purposes a part of the Master Node, as I have indicated earlier in the series of posts), except in very specific situations (For Example: When an Alert is active, and all access for the AID you are using is termminated, you may then spoof commands at an appropriate access level, with the appropriate penalties. Assuming, of course, that you have access to the appropriate AID's to aid you in your spoof attempts).

I agree that my terminology was uselessly complex, and ultimately wrong on a few levels. But the core concept remains the same. It is verifiably easy to Spoof a Non-Slaved Node. It is often Uselessly complex to Spoof a Slaved Node (which is why you should use them for certain things).

Note also that you cannot use Spoof to change a Nodes Settings (Access Rights, Subscription Lists, etc.), as some have suggested.

Regardless... I have thoroughly enjoyed the discussion, even when it got a little heated. Thanks for the Discourse...

Posted by: Ghost_in_the_System Jun 4 2011, 07:45 PM

No, faking authentication is part of what the spoof program does, unless you're saying that it waits several seconds before it sends something to the master to confirm the choice (which is different from authentication). Otherwise that is all part of how a node operates, and defeating that authentication is all part of the spoof program.


2 is not correct. You do not need to mess with your hardware, and if you did change your hardware, you would not need to use spoof, you could just send commands directly.

Edit: If 2 was needed, it would say that you need to spoof your own access ID (in this case spoof meaning fake or change), Unwired p52.

Posted by: Yerameyahu Jun 4 2011, 07:48 PM

You *can* set up additional, 'handshaking' authentication. It's just bad. Again, it's in the FAQ.

Posted by: Tymeaus Jalynsfein Jun 4 2011, 07:49 PM

See, I disagree here. But that is okay, no need to ignite another argument.
That is how I am going to proceed from now on.

Posted by: Ghost_in_the_System Jun 4 2011, 08:05 PM

Oh, just found another way to beat slaving: The puppeteer trojan. Infect the master with it and then you can use it to send commands through the master. I suppose this might just be seen as a different version of option 3 though.

You do at least acknowledge that if you weren't dealing with a slave, you wouldn't have to perform step 2, correct?

Posted by: Ghost_in_the_System Jun 4 2011, 08:17 PM

I think the biggest problem with option three is that it uses unique terminology that is not repeated elsewhere in any of the books. There is no other reference to spoofing an access ID (except your own)

There is reference to spoofing a command, which is the most commonly thought of function of spoof.
There is reference to spoofing a node, which is simply another form of the above.
There is reference to spoofing an agent/pilot/sprite, which is simply another form of the above.
There is reference to spoofing your own access ID which means the long involved change like TJ is talking about.
There is reference to spoofing your own access ID to look like someone else's, which is a specific form of the above.
There is reference to spoofing your persona, which is used for tricking limitations.

But no where else is there anything where it uses spoofing someone else's access ID. It could be taken to mean that you spoof the node that access ID belongs to, it could mean that you spoof your own access ID to look like that access ID. It could mean that you are spoofing (pretending to be) that access ID as part of the regular usage of spoofing and is simply clarifying that you have to use the master's access ID as opposed to that of something else.

I suppose the easiest thing to do would be to spoof the master itself with a command to send a (totally legitimate) command to the slave. Some small chance of being noticed, but only if the master is watching the slaves like a hawk, in which case you're likely to get noticed quickly anyway.

Posted by: Tymeaus Jalynsfein Jun 4 2011, 09:08 PM

Indeed...

And yes, if you are not dealing with a Slaved Device, Step 2 is completely unneceessary. In fact, I never had an issue with non-slaved Spoofing at all.

Posted by: Ghost_in_the_System Jun 4 2011, 10:20 PM

Okay, so it is just that one undefined line that really can be interpreted in several ways. I think it is excessive to have to change your Access ID, and could argue against it by pointing out how it says it only accepts connections from the master as opposed to a master or the masters, but it isn't a big deal. If I'm ever at your table I'll just use spoof against the master to send a legitimate command to the slave and be done with it (Or maybe stay away from the matrix entirely )

Posted by: Tymeaus Jalynsfein Jun 5 2011, 02:17 AM

Heh... No worries...

Posted by: Bearclaw Jun 6 2011, 04:30 PM

Good analogy. I did the same thing about 4 pages back, but it got ignored, and I used T2, because Arnold is always awesome.