I distinctly regretted not updating the computers of the fortress immediately when I had the time. Not that my reasoning was unsound. I chanced a calculated risk. The likelihood of a cyber attack was rather low, and I decided not to replace the obsolete and obsolescent equipment with modern systems which would experience the same fate in a couple of months.

I had to design the new computer system first, and then have the NADA make them, but they would be orders of magnitude more powerful than anything on the market today.

So I took the risk, and it seemed like it bit me in the behind. You sadly can’t win every time.

The first thing I did was check up on the status. The computer systems were, unfortunately seriously compromised. I would have to go over every single file before I could trust the system again. Nothing I could not manage, and most of it I could let the cluster do, but for now, I had to survive first.

Normally anybody in this situation would be neck-deep in the excrement. As soon as the system was compromised the other side controlled the battlefield. Normally, that is.

I on the other hand had a couple of trumps in reserve, almost all of them depending on the Q-link to work. The first one was that I switched over the control to the cluster. The security system spun back up, and I got a threat analysis, while the door by the garage closed again.

Unfortunately, this would not keep long. As soon as the assailant found the nodes I hooked the cluster into he could simply isolate them. But for now, I have won a couple of minutes to react.

Next, I began to switch the viron to something I normally used. While thematic virons are nice and feed the ego of the owner or the creator, they usually make it easy for a jack to adapt and assisted their stealth.

Personally, I preferred an ever-shifting illogical viron. I heard them described as Escheresque, whatever that means. What it did though was prevent the enemy from using an easily deciphered theme to make his subconscious fit in.

In the swirling chaos of impossible shapes I found not one, but two jacks, unable to adapt their stealth fast enough. I activated my own stealth, synchronized with the viron, and moved closer to them.

I was not particularly surprised that I was significantly faster than them. I was not completely convinced that I had, at this moment, the fastest combination of board and implant, as some of the top 10 held their cards close to their chests, but the likelihood was high.

My superficial analysis showed them to be using Kawamoto boards, probably late Yurei. That placed them clearly as corporation hackers. And that made it a bit easier for me. Whoever had sent them had not splurged on the much more expensive, but much more capable as well, Shinobi line.

The Yurei were a well-known variable in my world, and I was confident that I could take out both of them.

While still being mostly invisible I loaded up Excalibur and Aegis, two of my most powerful combat utilities.

Aegis was essentially the standard defensive utility, namely an adaptive input buffer, writ large. The twist here was that I had the cluster manage the adaptive part. I was able to do that, contrary to standard operations, because I had linked my supercomputer via Q-link, which made it better than an optical link.

Others were reliant on the resources their board could offer, as including a bigger server or cluster over the matrix introduced a certain amount of lag, and lag was deadly in our game.

Excalibur on the other hand was essentially the opposite. Any attack utility was essentially based on the same basic principle to go through the defenses. They tried to break the encryption and then overwhelm the input buffer so that the actual payload could be delivered to the board, or the jack if one was so inclined.

And again I was able to incorporate the Cluster into the equation. It was able to brute force the encryption faster than others could finesse out the encryption, not that I resorted to brute force, not alone that is. I had the experience that I got the best results when I combined the methods and used finesse to steer the brute force.

Naturally, most of my attack utilities used the cluster. Excalibur was my solution if I wanted to absolutely ruin my opponent's day. Most attack utilities were designed to either damage software, hardware, or wetware. It was surprisingly hard to create something that could do equal damage to programs, the computers they run on, and the brains of the hacker using them.

Excalibur had been my 76^th attempt, and the one that finally worked. I tested it on cloned brains, any type of hardware I could get my hands on, and every software I could find.

It was not the best in all of them or even any of them, but it was devastating for all. It was nothing I would use against an abyss-dweller, but against corp-drones, especially ones with second-tier boards, it was a one-shot kill utility.

Just as I was closing to the first one, they opened a port to let in a veritable swarm of hunter-killers. Eight HKs changed the matchup considerably. I was still certain that I could take them with a bit of planning and a hunk of luck, but instead of a curb stomp, it would be work and hard work at that.

Depending on what HKs they summoned the hackers were actually the lesser threat.

Accordingly, the next thing I did was analyze the HKs, to harvest any piece of information I could. The results were… strange. The programs were a hodgepodge of different commercial and civilian bots. The whole was unprofessional to an extent that I could not understand.

I mean, obviously whoever had sent them was a corporation and one that was not averse to using Kawamoto-tech. And Kawa had a whole lineup of hunter-killer lines from nearly legal to blackest of operations, and from cheap and barely usable to run when you see them. So why these kludged bots?

Not that I was complaining. It was just the hunk of luck I needed to make my chances pretty good.

This new situation demanded a revision of my battle plan though. I would no longer be able to take the intruders with utilities alone, and so I had the cluster spin up a few Banshees, my premier anti-bot combat bot, and preload one of my own hunter-killers.

The Banshees had the advantage that I incorporated every stealth utility I could my hands on or create and the strongest anti-bot utility I had.

Each of them would be more than enough to play with two or three of these strange HKs, so the four of them should be enough.

They were almost completely useless against a hacker or for causing damage to any computer hardware, but against HKs they were essentially the best I had.

On one hand, I was glad it was happening in my system, with me controlling the viron. On the other hand, if it had happened in a system I did not care about, I would just nuke the joint and go home.

Then it was time to put the plan into action. The intruders naturally knew that something was happening. The shift in the viron alone has to have told them that. But at least for one of them that would not be enough.

With a silent sigh, I rammed Excalibur through the avatar on the right of me. For us visitors in the Matrix it looked as if a sword made of light pierced through him, and he slowly burned from the inside.

What happened in the background was that the cluster already had optimized the ice-breaker for KVS, or Kawamoto Virtual System, as they called the OS of their boards.

That by the way was why I wrote Precious OS myself, and even the version I sold with the Mark IV was slightly different in the way it handled buffer encryption, DEP, and other defensive measures.

Nobody but me could prime an attack utility to the idiosyncrasies of my board.

With the primed ice breaker Excalibur, or closer to the point, the cluster, made short work of the encryption of the buffer and quickly filled it up with random junk. Without the decryption, I would have needed to fire attacks non-stop, wasting time, hoping to fill up the buffer, and having the next attack go through.

The next step was overwhelming the DEP by sending attacks to several known weak points in KVS, managing to embed and run the trojan that made the rest of the attack possible. The ASLR was defeated by elevating the trojan to kernel-level permissions.

Then it got interesting. At least for me.

One, the trojan infected the jack with a sub-routine, loading a cascade into it, which would, over the course of a couple of milliseconds, overload the autonomic nervous system with conflicting signals, essentially shutting it down, and with it the person.

At the same time, the jack fired into every sensory cortex, inducing something like a fever dream, that would, if my tests and simulations were reliable, knock the hacker out from one moment to the next.

All that would kill the hacker.

Two, the trojan would activate every writing laser, and overwrite the data storage with several passes of random numbers, while shooting down every other process and overwriting the memory with, you guessed it, random numbers.

Three, that was what I was especially proud of. The Cirrium processor line had an interesting design vulnerability. To be fair, it was almost impossible to exclude this vulnerability from any processor.

The data lines inside the chip, these delicate strings of silver formed an intricate network. An intricate network with an interesting resonance frequency. If it was stimulated in the right way, at the right frequency it became hot. Extremely hot. And thermal throttling would do nothing to stop it, especially as the trojan was suppressing the necessary functions.

In short, a few seconds after the hacker fell down dead and the OS crashed, the processor would melt. If one considered that it was composed mostly of carbon, which has a melting point of a bit over 3500°, celsius that is, it is pretty obvious that the whole board caught fire almost immediately.

By the way, my personal Cirriums had a slightly modified layout, changing the resonance frequency, and included a few bridges out of zinc, which would melt easily and break the network up. It would still ruin the processor, but the rest of the board was safe.

All that though had happened in a fraction of a second, and the hacker was screaming his digital death cry.

The other one reacted faster than I thought he had in him and opened up space between us.

At the same time, the Banshees assaulted the HKs, taking down four of them in the initial attack.

Naturally, that was the end of the good news for me. Almost immediately my buffer screamed in alarm, as I was attacked by an unknown assailant. I quickly moved some distance away and observed the situation. The kludged HKs had been joined by two new ones.

Even at the first glance, it was obvious that these were professionals. From the skin they sported, they seemed to be Kawamoto Basan class HKs. Usually, nothing I could not master. But two of them were hard on the limit of my abilities. Two, plus one hacker, plus the surviving Frankensteins would be hard. Very hard.

I had no other choice than to use my own HK, something I had wanted to avoid. Something like that has the tendency to become less effective once it has been used, and I wanted to keep my trumps hidden as long as I could. Still, it was better than kicking the bucket. And so I ordered the cluster to spawn not one, but two instances of the HK I had it preload.

I have to say, the visuals were impressive. I patterned them after the appearance of the monster I named the HK for in one of the ancient movies. Not one of the remakes could ever top that.

And so the figure made out of black smoke, red fire, and darkness, the Balrog appeared. Twice.

Unlike the Banshees, the Balrog was more versatile. It had a varied arsenal, going from non-lethal anti-organic up to tactical nuke. It was equally capable to go against a bot as it was to go against a human. And in the current situation, I had them go for the kill.

The problem for me was that the sneak attack had nearly overwhelmed my buffer. It was slowly reinitializing, but for the next few seconds in real-time, I would have to be very careful.