Ok, there's two tricks I want to try out in upcoming games and I though I'd run them by the Dumpshockers first. One is basically an AR "Invisibility" hack, the other is a defensive attack to screw with technos and hotsiming hackers.

The first trick is pretty simple. Take security guard Bob. Bob is standing at a critical access point that I want to get by. Instead of having the mage cast invisibility, I want to hack into his commlink (shouldn't be too tough) and upload a modified AR software called Reality Filter (AFB). Reality Filter allows you to tune out unpleasant sounds, visuals, and smells. I'm a pretty unpleasant fellow and so I've modified this AR software to consider me and my chummers "unpleasant". Now presuming this guy is using AR (which seems pretty safe) my chummers and I are basically invisible to this slob, right?

If I can get this to work, I'd really like to stick a trojan onto it, then hack into their system and upload it to a nexus as a "standard AR software update". Right setup and you're not only invisible to every slob hooked into their system, you've got a Trojan (like Rat or Vector) to wreck their system whenever you need. Sure, it might be a tough hack, but I can't imagine I'd ever need to hack their system again. Worst case scenario, I Vector in a bunch of Replicating Worms and crash/kill everyone hooked in.

The second one is a take-off on an old trick I saw in a 2nd ed book, I think it was called an Escher Loop. The idea was you access the wrong node/open the wrong file and pretty soon all you see is an Escher painting like the one with the stairs. You're trapped and can't even log off, you just hope one of your chummers comes by and disconnects you before you die of thirst.

My take on this is to use a ring of slaved nodes. It works off this premise. Say I slave node A to node B which is slaved to node C. Anyone trying to access node A gets booted to node C. So in this case I have node A (my hacking commlink) and nodes B-E (metalinks). Node A (my hacking commlink) is slave to node B. Node B, however, is slaved to node C, which is slaved to node D, which is slaved to node E, which is slaved to node B. You've got a complete loop with the actually valuable node hidden outside. Map it out and it looks like a lollipop.

Now someone tries to hack my Commlink (Node A). Because Node A is slaved to node B, they get kicked to node B. However, because node B is slaved to node C, they get kicked to node C, then kicked to node D, then kicked to node E, then kicked back to node B and the whole things starts again.

Presuming that slaving works this way (I'm like 95% confident), the intruder never actually lands in a node, they're constantly being forwarded to a new node. Here's where I'm less sure about what happens. Quite simply, I have no idea if someone can take Matrix actions outside of a node or while they're logging into a new node. If they can, it's just annoying, they logoff and log back on, then DDOS the loop to break it or something. If you cannot, however, then it gets fun. Someone hacking by AR is probably going to be frustrated for a bit, then manually reboot their commlink and try again. Hotsiming hackers are stuck though, they can't take non-Matrix actions (ok, technically they can at like -*8*) and the loop prevents them from taking Matrix actions, meaning they're trapped until either their chummers unhook them or they manage by sheer willpower (at -*8*) to rip their rig off. Technos are even worse off, as far as I know one of their chummers has to find him, figure out what's going on, then knock him out to disconnect him, then suffer dumpshock.

I'd appreciate any advice on these, especially if I messed something up and how I can fix it.

Slaved devices forward connection attempts to their master device. A full circle is simply a bunch of devices not connected to the matrix.

The master device could be 15 nodes deep, and surrounded by Glacier nodes though... But I am pretty sure the Escher Loop would not work because you MUST have a master device to which you are Slaved; As for getting out, AR users just stop the hack, and VR users just unplug (it is a -6 to non-matrix actions) either through a physical removal, or through a DNI Killswitch. Sure, they take dumpshock, but it is a small price to pay for getting out of an infinity loop, (or a Glacier Hole) which, as I indicated above, I do not think will work (Thought the Glacier Hole would).

Your first idea is, essentially, hacking someone's eyes. Cannon is that doing so in real time can't be done because the sheer amount of Editing power necessary to take an image, sound and smell, delete it and overwrite it with the background visual with enough detail that AR can be mistaken for real life can't be done in one second, or a single Intiative pass; the amount of time it takes to observe something.
It might be possible were one using multiple Agents and a UV node.
It is possible to overwrite that image with a non image, deleting it entirely or replacing it with an icon like The Laughing Man in SAC, but then your target knows something is up, even if he doesn't know what it is.
Your second idea is the first thing that comes into everyone's head when they read the Matirx rules and, were it possible, no one could ever hack anything ever. It's the Shadowrun version of wishing for a million wishes.
Do a search and you'll come up with a million reasons why not, but it's covered more extensively in Unwired.

The rule of thumb in Shadowrun Matrix rules is that if it isn't around, it isn't allowed. Unless you have the program or Matrix action that allows you to do the thing you want to do, you can't do it. Much the same reason you can't design a spell that kills everyone you hate with zero drain, even though that's a neat idea.
When you allow for things like daisy chains and one off pads and heavy Encryption of every node, then the Matrix becomes unplayable. That's why they got rid of deckers.

BUT... There is an ARE that does just that... It overwrites what the user does not want to experience. The ARE Program is in Unwired, Page 108. It is called Negator, and costs 100



So....
Of course, it requires the Target to be using AR, but that is almost ubiquitous these days...

The first idea is quite feasible. It depends on the guards having AR of course, but those odds are good (especially if they have smartlinked weapons!)

The second idea seems less feasible. It's essentially a continual connection forwarding loop that just keeps going in circles. But the trap is very simple in nature, and I'd say most OS would have some loop detection built in to prevent this. Some limits to how much memory and CPU resources the connection-forwarding system is allowed to consume, and some cycle detection - stuff that's quite easy IRL already.

Thanks, that's exactly what I was thinking of.

One of the other things is that in Street Magic, when it discusses the Matrix and Mages, it mentions that AR objects can block LOS for spells, which could be huge if you're subtle enough.

I'm gonna disagree with a lot of people here about how feasible this is. People seem to be saying, "Yeah, sure, that works." But there are a couple of reasons I think this won't work:

First, the guy is using AR, not VR, and AR is only additive. It can't remove sensation that's there without a Simsense Module, which by definition requires VR. Additionally, Cold Sim is definitionally not real enough to be mistaken for "real senses" - only UV hosts can truly be mistaken for the real world, and you need hot sim to access a UV host. So you can only add, not actually *replace*, unless the guard has a Simsense module and is actively using it. Possible, but not always gonna be the case.
Second, because AR is additive and not subtractive in terms of sensation, you can probably "paint over" his vision, but you can't make noise over whatever he hears without filling his ears with static or using active interference waves - which may be possible, but I would rule it's a penalty to auditory perception checks rather than outright "can't make any tests" territory. And as for feeling the breeze of you passing, the vibrations in the floor, and smells? Not really any way to disguise that. At least, I've never heard of any guard with an AR nose-hookup to spray perfume up his schnozz. But this is also the case with invisbility - you can't be seen, but you can be *sensed*.

Basically, I'd apply it as a penalty based on how complete his AR rig is, and how powerful your software is. It's not as good as invisibility, because seeing someone fucking with the light-up LEDs on your AR sunglasses is MUCH easier than if you're actively hacking their cybereyes. IF they're using AR through cyberware, this becomes a lot more reliable, IMO, because you can directly edit their sensory feed.

EDIT: Mechanically, I'd work it like levels of "camouflage", hindering the spotter, with the value varying based on what type of AR rig they're running. It should never go higher than -4, I don't think, because that's a active camo suit's max.



This is a good plan, but I'd rule this would probably take an agent loaded to superuser or admin access level, because it's a mandatory update. I'd still give individuals an analyze test on the update, too: even when I get updates from my servers at work, my local machine runs anti-virus and malware detection on the server updates, for this specific reason. But if you can get the superuser/admin access, most of your battle has been won, and you can probably pull it off.



This seems like a pretty simple abuse of stealthed gateways. Unfortunately, there is a really easy way around this: get more successes on analyze than the node gets on stealthing, while in Node B. Because Node A is slaved to B, if someone rocks the analyze there, they'll still be able to see the hidden gateway; it necessarily must exist for data to get passed between A and B. Rigging a boot to Node C from B, when someone tries to access A... well, that's doable, but it's the result if they fail to hack their way in. If they succeed, they definitionally can't be booted because they've cheated the system.

Yes, but it doesn't make it invisible. It replaces it with a nice set of flowers or a sunset or a pretty picture of a simsense star. Maybe it just makes it a blank splotch. It hides, masks or blots it. It doesn't make it invisible; like the example I gave of The Laughing Man in SAC. It can put that logo over it, but you know something is there.

You cannot create a group of circular slaves. B will foreward the slave attempt (to open a connection) to it's master ©, which goes to D and then E. Thus what you have is E asking to slave to itself, meaning it will foreward all attempts to connect it to -> itself. Note this does not cause a crash because this sort of error is handled via making you read unwired (a painful process).

The invisibility is much more difficult than it looks. The guard is probably slaved to the master node, in whch case you might as well just hack that. I'd probably just SnS the guard or use magic.

Brilliant!

No Doubt, it does not make you invisible, but it can mask the presence of those things you do not want to experience (The scent of the unwashed masses, the presence of all those bums on the streets, etc) with something else, maybe even an overlay of a composite of the background, whatever. Not invisibility, but definitely masked... Unfortunately (or fortunately, dependant on your viewpoint), the level of masking is entirely up to the GM, at least mechanically.

Yes, but if that is the case, implementing this program into the guard's AR unknowingly will surely let him know that something is amiss. Otherwise it's just like recording info from the blank hallway from his eyes and putting it on a loop. The two won't match up.
What Polite Man seems to be talking about here is hacking the guard's comm, loading this program and walking by the guard invisible because the program will make it seem like he's not there.
I'm saying the program will put some sort of overlay over the unpleasantness, but the guard will notice the simsense star portrait or black AR visual in his glasses/cybereyes walking past him and know the jig is up.

Quite Possibly... Yes...

If done correctly, implementing it in his AR won't alert him at all. He will simply continue to see everything was seeing at the time, because he was not looking at the item to be removed. As for the substitution, who says it has to be a simsense star, flowers, or anything like that? If the guard has been staring at the same expanse of wall for an hour, simply have the program overlay that section of wall over everything. Much like fucking up the Z-layer on a CSS call, you cannot see anything behind that nice background image that he expects to see anyway. Yes, it is rather like feeding a loop into a camera, except the camera is his POV, as recorded by his own image link. Where it would get hairy, of course, is if Bob decided to take a walk and stretch his legs only to find himself by all appearances walking on a treadmill until he hits the wall that has been overlaid with the background.

Agreed. So the answer to the question "me and my chummers are basically invisible to this slob, right?" is no. No you are not.
But you are granted a bit of concealment, which is worth it's weight in gold. Combine it with a chameleon suit, an invisibility spell and some thermal smoke and you've got your self a party!

EDIT: Where it would get hairy is when the guard tilts his head slightly to the left, throwing off the background overlay. At this point, his visual stops looking like vision and begins to look like false AR feedback. Your eyes are constantly moving, re-evaluating visual perception. Imagine the guard closed his eyes for a second, and when he opened them, the runners had put a giant painting of the lobby in front of him. As long as he stares at the same spot without moving, he's fine. But the moment his head moves slightly, he's going to know something is amiss. At best it's like the invisibility shield in Predator, which is already a thing in Shadowrun.

In a world where people too often put faith in technology, I don't see why it wouldn't make you near invisible. The software in question, the negator, could be used to replace the only three senses that would allow the guard to detect a team in sight, sound, and scent (overlaid with what he was already sensing moments before). If he actually is not seeing them, but rather seeing the AR input (especially if he's got cyber-senses as opposed to accesories) then I would see no reason a check to perceive through that AR overlay would apply. If the software was able to be perceived through, it would fail it's basic reason for existing.

The way I see it he either has to pull off this hack with every guard's commlink (which takes time where the group is potentially exposed to other guards seeing them, and could still fail and cause problems), or hack the much harder server. A decent firewall, some security minded techs on the opposing side, or Bob deciding to take off his glasses and rub his eyes could all mess up the plan.

I tend to reward such planning, which doesn't mean it would work every time.

1) AR input isn't as vivid as actual visual input unless it's on a UV node. AR provides things like AROs and commercial signs. Your visual input is very distinguishable from your AR input.
2) The things he was seeing moments before change as his perception changes; i.e. his head moves.
3) The reason for the Negator is not to make unpleasant things invisible, but to hide(or block or mask) them from you so you don't have to look at them. So you don't have to look at the fat woman's thong at Wal-Mart. That is worth it's weight in gold.

There is a significant difference between merely censoring an offending object with a black square or blurring it out, and actually replacing the object with digitally recreated realistic 3D illusions of the background behind the object.

The former is reasonable for a 100 program. The latter is not.




-k

You do have a point...

It's not so easy to do with the Negator software, but possible. Before loading it onto the commlink edit it so that the AR that replaces the object in question (the infiltrator) with something believable. Like a sudden spam advertisement, not unbelievable that the occasional viral ad might get through the firewall and past the spam filter. Just make it a boring ad so that they don't give it much thought beyond "Need get on Billy's case about updating the spam filters again. Lazy-ass spider."

The problem though is that you have to pick something that the guards won't pay attention to, or notice is moving down the hall, etc.

Best bet would be to pick a randomized pattern of different tones of grays and do your sneak at night and stick to the shadows.


Anyhow, given the nature of the Negator software I might allow it to give a small camouflage bonus if used cleverly, but not treated like invisibility.

Valid concerns, all, so I'll break down with how I see it.

1. I personally don't see any reason why, given the actual visual information from an image link, it could not pull the old 'loop the camera' trick. The hacker would have to be patient enough to watch for a period of time to count for variations in the POV, but it's not generating anything new, just playing back. I can project a photorealistic image onto any surface now with a projector, it's hard to believe AR, which is just another type of layer, couldn't. If it was generating that background over multiple POVs and extrapolating in realtime, hell no, but this is a static guard we're talking about, his field of view was only so much.
2. I covered this in part in the first section, but I will expand. I've done guard duty before (lord love it), and in a static postion on an eight hour watch a person's ability to process sensory input degrades - they get tired, bored, and lazy. Eventually the routine gets to the point for even the most on the ball person that their mind starts expecting to see certain things, and will forgive some variance. Record field of view for ten or so seconds, and it gives enough for the overlay software to get a general idea of what they should see where. If Bob tilts his head a little different, yes, it is going to be off by a small margin, but if Bob's been on watch for hours on end (any decent infiltration team should know the guard rotation, and hit when they've been on the longest possible time).
3. The reason a gun was made was to shoot something, but that doesn't stop me from hitting people with it (yes, that answer was cheeky, but relevent). This is merely a different application of the software which extends its intended use, but requires patience (to record enough relevant background), skill or toys (to pull off the hack), and the GM not to be cruel (and have Bob sneeze and pop out a contact lense).

I believe 1. also addresses Karma's concern... the method as I see it isn't rendering a realtime 3d image of the background (oh man that would be some sexy good software), but is just using some input hacked from an existing image link to cover the unsavory element (those dirty shadowrunners).

As for the smells and sounds, just take an olfactory sample and a sound recording of the area.

It should be noticed this is the kind of trick I would allow a hacker to pull of sparingly at best... it can be shredded to shit with an aware guard, a good firewall, or plain old back luck. Abuse would result in plenty of sneezy guards.

Made me think of a video I saw for a version of PhotoShop that is out on the market now, and how quickly a person was able to edit pictures to take out certain elements (sorry, I don't have a link to this). Like one was a desert scene with small cacti all over the place and a rode. A click with one tool and the road was outlined, then it was replaced with cacti by selecting another command practically instantly. Once the object outline was gone you couldn't tell there had been a road in the image in the first place. Same video had other similar editing examples.

Considering how fast that program worked now, it makes me wonder why an ARE Negator program, if linked to an Image Recorder, wouldn't be able to do the same in real time to make AROs that hide what you want. Most certainly commlinks of the 2070's can outperform our current supercomputers by an incredible amount. I'm also considering the size of WATSON that was on Jeopardy just earlier this week (IIRC two server rooms), and figuring that Shadowrun agent programs can work just as well on a commlink the size of our current iPods.

And besides, in reference to the GitS:SAC Laughing Man, in one episode he did hack Batou's eyes to make himself invisible but not a big fat paper book he handed to Batou, and in another episode hid himself from a doctor that was messing with the Major when the Major was about to do a body swap.

What about not editing you out, but only removing the bulge in your jacket where the gun is hidden?

1) This is reducing a person to a camera. It is impossible to account for the multitude of angles at which a person views a particular situation. You would have to know exactly how the guard is going to move at every moment without variation. A person has a pattern, but our monkey brains vary that pattern and recognize inconsistencies. We cough, stretch our legs, move about and obverse minor details to keep our minds active, but also to ensure that predators won't be able to detect our patterns with consistancy. It's what let us hunt and keep an eye out for predators back when that was a thing. You think you are watching the same thing over and over, and indeed your perception does waver after a while, but not so much and with such consistency that it can be duplicated with 100% accuracy. Your brain is observing things without you actively knowing it and, when it encounters an inconsistency, it wakes up your conscious mind and has it investigate.
2) This is the painting scenario I was talking about. While you were doing guard duty, could someone have put a painting in front of you of the thing you were watching without you being able to tell the difference?
3) The reason a table was made was for putting drinks on, but that doesn't mean it can't intercept an intercontinental ballistic missile. Wait, yes it does. You are asking a piece of software that is meant to block, hide or mask to perform 3D rendering in real time with UV node accuracy. If a 100 nuyen piece of software can do this, why can't the Edit program? If it can, then why can't you hack people's eyes as the BBB says you can't. If so, the why do cameras exist in the Shadowrun universe? All can be manipulated to make the camera or even the person see whatever the hacker wants all the time. If this is the case, then anyone with cybereyes is essentially blind, as someone with a mere 100 bucks can make their 9k cybereyes worthless. Heck, he can make their audio and olfactory senses worthless on top of that, if they have DNI.
If this is the case, then a decent hacker can drive people to madness and murder by completely manipulating their entire perception with a cheap ARE program and a decent Edit and Exploit program.
That is, essentially, brain hacking.

You might allow "this kind of trick sparingly" but once that gate is open, that gate is open for good. If you let your hacker hack someone's eyes in real time, then you have established a precedent that a hacker can completely manipulate the perceptions of everyone in the world with an AR interface. There is no once and a while. Either they can do it, or they can't. If they can, then no one would have AR any more, because it would be impossible to be trustworthy. That breaks the Shadowrun universe.

Its called Context Sensitive Erasing, and it's been around for years. Got a plugin for GIMP that does it.

And it's actually not that smart. If you don't have anything to "steal" from (ie. lets say you're editing out a guy standing in front of the only window on a building) the algorithm is going to fail pretty badly.

Now, of course, we are talking about moving images which gets both easier and harder again. I've also done this "edit that car out" in Nuke (a weapon of mass creation) and it looks "alright" but not "great" depending on how well the software can track the object-to-remove and if the frames on either side contain "background" and not more moving objects (editing out one person in a crowd is a bitch). And even then, it's not fast. For a 3 second clip it took the machine (which was reasonably beefy) a good 15 seconds to do all the computation.

That algorithm also fails spectacularly when the object-to-delete moves behind something, because as it vanishes (or god forbid, changes shape in some manner, say a guy walking suddenly jumps) and the computer doesn't recognize it as the same object (or drags in some of the other things in the picture.

Sure, now. What about after sixty years of computer evolution? In SR4 we're talking about a level of technology that allows one wearing contact lenses to see 3D holographic images, where you can have a trid call with someone and see their disembodied head floating in your field of view in true 3D as they're talking to you.

Granted the ARE Negator program is a paltry 100 compared to even the basic of hacking and editing programs, but I cannot believe that after sixty years that Context Sensitive Erasing is going to be just as bad then as it is now.

Love this. Almost as good as invisibility and tons more cutomization.
A couple questions and clarifications:
#1 This isn't intended primarily as a solo hack, more of "I've already hacked their system, how many fun things can I do in here?" There's a couple different things I'm considering, like a bunch of AR "shadowrunners" hitting different areas during the run, getting their headsets to play heavy metal at max volume during a run, etc. Sure, they can dump their tech stuff but then how are they gonna organize a response if they can't AR chat with each other?
#2 Presuming I loaded the AR Negator software, that software has to have a pretty good description of me and my chummers. They can investigate that afterwards to get a general physical description of us unless I corrupt the data.
#3 Presuming that AR images block LOS for any mages, if I had my chummers appear as spam-ads on AR, would that block LOS for any mage targetting us?

Not going to get into the realism argument (I don't know enough and this is SR) but I do want to touch on the game aspects.

There are two major differences between an Edit hack and and AR hack.
#1 An AR hack requires preperation and only works in a couple of preset ways. For example, if I install the AR software and we need to bring another person, or the Sam just bought a shiny new gun, or it's an extraction and we have to carry somebody out, there's no way to alter the AR hack short of getting back into the system and releasing a new "update".
#2 An AR hack requires installation, which generally speaking is either going to take longer or require admin (maybe security) access.
The colorary to #2 is that only "profesional" hackers/technos will be able to do this. And at that level, they might as well load a crash worm, or take over your car, or steal your sin, or upload a video of you wacking a LS cop, or...
It's not that this kind of hack prevents hackers from screwing non-hackers over, it's allowing them a greater deal of manipulation. For example, if I hack your cybergun, I can eject the clip. If I'm a little bit better, I can make an AR image of a mage casting and watch you unload at it. They're both nasty-(ish) but one's more effective and just plain cooler.

They would ideally have one or several identifying featurea. Inside a corp environment, filtering out subjects broadcasting employee ID in the range of #7400-7450 should be sufficient. Or anything with barcode X displayed somewhere, or any sound in some frequency range. Extensive, specific description is setting yourself to fail.

I kinda like this idea for the ammount of preparation it takes, but wouldn't have it grant anythin near total invisibility. More like applying a perception enalty up to the hits rolled on. The hacker's Edit or Software check to edit the program to do this. Actual penalty depending on the ammount of sensory "background noise", and the ammount by which the subject to be Negated contrasts with its environment.

I think the AR hacking overlay would work, but only to a point. First off, assuming all your initial conditions are in place and that the security guard mook in question IS using an AR visual/whatever else assist, I'd allow the hacking to augment any stealth attempts. My reasoning is that no matter how good your programming is the security guard would still have the chance to notice you, unless you say, completely blacked out his AR-enhanced vision. In which case he would notice something ELSE but that's besides the point.
I'd allow something like a Software+Edit roll to contribute hits or can be in place of a nonexistent stealth skill much like nanopaste/latex mask disguise kits can. There'd have to be some sort of limitation but that's how i'd run it

There's a bit of software available now that runs on a smartphone and edits out an object on a relatively static background. There's a little blur when the camera moves fast, but that's mostly processing. It can do backgrounds with a repeated theme by derezzing the image, editing out the blurred object and then using intelligent up-scaling and sampling to recreate the area that the object was in. Looks interesting, but it's more like what's being described above.

Now fast forward to the 2070 and think of the smartphones they are running. Think of the software they have available and the agents. The idea of blocking something with a sim star or a pot of flowers would be quaint. If we can already block out a persons face in real time with an image today, I'd have to think that you can produce a seamless multi-sense realtime edit by that time. Especially since AR is basically all about that shit. Just load up Adobe Reality and tell it what you don't want to see. If you're guarding something and someone has hacked your commlink and they are remotely running the software without a GUI, that's Shadowrun.