Out of the list I had made, I had settled on two finalists.

  Most of the options I had found available just didn’t speak to me. Most of what I had found seemed to be physically unwieldy. There was a reason why most characters of fantasy stories were human-like–they were different but similar enough that we, the readers, could resonate with them. I was no exception to this feeling.

  The first finalist was ‘Dark Elf.’

  Not by my choice, but my dungeon’s theme was very much related to necromancy. First, there was the castle ruins and literal underground chambers that likely was for that very magic. Second, my monsters were all undead. dark elf plus necromancy was a fantastic thematic match. On the other hand, if the ‘dark’ in dark elf had similar characteristics to ‘dark’ in dark goblin, that was going to be a problem. I did not want my companion to be corrupted by the black magic. Plus, while the aesthetics may be cool, I didn’t want an evil companion.

  The second finalist was ‘Kitsune.’

  There was no argument that foxkin was the best variant of beastkin, and Kitsune was simply an upgrade over that. Kitsune were cool, and I didn’t think that just because I had been a super big weeaboo. Okay, admittedly, I was a little bit of a weeaboo. But I contend that every modern man had an inner geek that liked anime and manga!

  But that wasn’t the main reason anyways.

  My skeletons could have a maximum of four skills, one for each cardinal axis of the sigil array. My Skeleton Warriors or Light proved this: Above level 5, the system only reinforced or improved existing skills.

  The dwarves broke this rule by having “blessings,” which modified two existing skills and added two extra slots. The dark goblins broke this rule by having a black skill that consumed other skills and added them in a horrible, eldritch way.

  I didn’t currently have the means to create my own blessing. Worse, even if it was possible, I might not have the time or resources to create it. So, I couldn’t copy that method. The dark goblins skill was black eldritch magic. That magic had been useful, but it also had a kind of will of its own. I don’t know if it was Evil (with a capital ‘E’) and even if I didn’t have a black skill myself, I was hesitant to expand that type of magic more, let alone give it to my future companion. So, I didn’t want to copy that skill either. However, the evidence was clear–there were ways to get extra skills.

  There was something else, too. The fourth skill the dark goblins had, [Sapiophagia], allowed them to consume the essence of others–steal their skills. The fourth skill of the dwarves was [Ferophagia], which allowed them to ingest different metals for various temporary buffs. Whereas the lower level skills for both races provided simple, passive, benefits, these two were active, adaptive. That flexibility was valuable.

  Adaptive, flexible, growth skills–that was what I wanted.

  Back to Kitsune.

  The defining trait of the Kistune, as a creature of legend, was its nine tails. Some more modern stories described how Kitsune started with only one tail and gained more tails as they gained more power. Flexibility, growth–the tails could have it.

  Making the decision turned out to be easy. Turning theory into reality–that was another matter entirely.

  The first step, I felt, was getting a better handle on sigils. Yes, the system was pretty good in taking my intent and forming it into a result that roughly conformed to my imagination. The skeleton warriors were a good example of this. Even though they were high level–and thus more complicated–the system imprinted them with a “flashbang” skill that was about ninety percent what I wanted. What I was looking to achieve this time was orders of magnitude more difficult, however. I needed to better align my vision, understanding, and will to make sure that I could bend the system properly.

  So, it was study time.

  I made a new empty note screen, but this one wasn’t for writing words. Instead, I focused on the Skeleton Warrior so I could see its sigil array while still keeping the empty note screen in my mind. Then I grabbed the latter and attached it to the former, intending it to show the sigil. That only gave me the names again, so I wiped the note and started again.

  This time, I focused on getting the empty screen to take a snapshot of what I was seeing through my regular dungeon senses. That was difficult though, since my senses were not like eyes–I didn’t have a single point of perspective. So, I went back to the screen and made it mostly transparent. Then I stretched it “around the sigil” like a reverse panoramic camera looking in rather than out. Finally, I willed the sigil to imprint onto the screen, like I was making a negative similar to a daguerreotype picture. This time it worked–giving me a 3D representation of the skeleton warrior’s sigil array.

  I went and did the same with every other sigil array I could. For those that I couldn’t directly perceive, I was able to use my own memory and Sigilmancy to recreate them. I also made separate versions for each individual sigil within a node (for example, [Bone Bash]).

  In this way, I began my study by looking at skill sigils first. I wanted to ascertain whether there were any commonalities across sigils (i.e. same or similar mana patterns within the sigil). I spent a few hours fruitlessly searching through before I realized there was an easier and faster way. I again used Interface to overlay two skill sigils and create a screen that merged the two and compared them. My first version was a failure as I had thought of it like an officer worker–showing differences–so I remade it to highlight only the sameness. After a few tweaks to include a small amount of deviation, I was getting useful information.

  First, the modified sigils of the Dwarves with classes showed the most similarities. This was obviously because they all had the same base racial traits. For example, the healer’s skill [Health Recovery] shared a few small pieces in common with the fighter’s skill [Bulwark] (as they were both in the first, level 2, node). In addition, some patterns had not been altered or only very slightly altered as they matched with the base trait, [Hale]. Going further, I also compared [Health Recovery] with Mr. Crazypant’s consumed skill [Heal Wound]. These two skills also showed pattern convergence.

  Unlawfully taken from Royal Road, this story should be reported if seen on Amazon.

  I continued to work through the skills, building up a database of similar patterns in the skills while noting what concepts I believed they might represent. It was a lot of guesswork.

  When dusk came, a goblin minion crossed the doors and scouted the room and then left. The forced disruption to my concentration and work annoyed me greatly. I deleted the demesne zone at the front entrance so I wouldn’t be bothered and got back to work.

  I pulled up the schematic for the Skeleton Warrior of Lights’s [Bone Dense] skill. This skill was a passive ability that improved the resilience, strength, and defense of the warrior’s bones. [Bone Dense] only had a very small section that was similar to another section in [Bone Bash]. My guess was that this section related to the common word ‘bone.’ I copied that section into an empty schematic screen then copied over the common sections from [Health Recovery] and [Heal Wound.]

  Nothing happened, of course–it was just a drawing at this point.

  I added another connected screen to interpret the drawing into descriptive words, but the screen returned a string of letters that was only partially comprehensible. I didn’t expect to get everything right, but even getting some sort of feedback was progress.

  I spent a full day manipulating the schematic and moved parts around, morphed others, and added different new pieces. All the while, my interface provided feedback by updating the descriptor screen. I felt like I was a monkey at a typewriter, as most of the time I was just making more gibberish. Sometimes, though, I got closer to seeing something more concrete, in which case I saved the version.

  I realized that I was missing an important variable in the process–my intent. The system had “filled” my intent when creating monsters. But now I was ignoring that part. The system could still help me to fill in the gaps.

  I had generated enough mana that I could create a new level 2 monster, allow the system to automatically fill in the gaps, and get more useful information.

  With yet another empty status screen, I started to push my willpower to create a new monster, envisioning the skill I wanted using the other skill pieces as templates. However, instead of simply pushing my will and spending the mana, I carefully and as slowly as possible pushed mana into the new creation. At the same time, I focused my intent (and even the mana) through the status screen, commanding it to show me the process of development.

  It was working.

  Keeping all the parts of the process together was a significant strain, but I watched as mana swirls came into being, moved around, and positioned into the various parts of the sigil array and skill sigil. Mana kept pouring in and the sigil array was close to completion. But I didn’t let it finish. When it was as close as I thought possible, I pulled as hard as I could on the mana, stopping the last bits from leaving. I then forcefully cancelled the entire thing. The sigil array collapsed then exploded in a beautiful starburst. The mana rushed back into my core.

  No, most of my mana did. The intentionally failed creation had cost me a single point of lost mana. The small amount of waste mana was worth it, though, as the status screen showed me a schematic of the nearly complete sigil array and skill sigil.

  The system liked what I was doing too, as Sigilmancy and Interface both increased a level.

  I studied the skill sigil and did another comparative analysis with the skills I had based it on. The skill sigil wasn’t entirely finished, but I felt that, with 19 of 20 mana having gone into it, it was within the margin of error. My higher-level Sigilmancy was also giving me an intuitive sense of what was going on, itself filling in the gaps of my knowledge.

  Although I wanted to see a proof of concept resulting from my hard work, I did not want another gimpy level 2 skeleton.

  I would make the skill into an item.

  I had only seen one enchanted item so far–the coin. However, when I thought about mimicking the coin to create an enchanted item, Sigilmancy warned me off. When I focused my intent on what I wanted but started thinking of an enchantment like the coin, the skill gave a sense of resistance. Instead, it felt like I needed to put the entire skill sigil into the item.

  So, using Interface, I planned out a full schematic for the item I wanted as well as the associated sigil that would imbue it with magic. When I felt good enough with what I had, I willed the item into creation and to match the schematic. I pushed mana into the item’s creation.

  On the top of the helmet of one of the Skeleton Warriors, a coronet made of bone and lined with Silverium appeared. The fleur-de-lis symbol was etched around the coronet’s band, and small balls of highly-reflected crystal were also inlaid into it. I inspected it.

  It had cost a staggering seventeen mana.

  That–that was exactly the skill I wanted to create!

  Skeletons didn't need healing--I didn't think the dead could be healed by, well, healing magic. However, they were made of bone and a skill that could regenerate bone seemed very useful. It was an appropriate mirror to not only Mr. Crazypants, but other delvers who would likely have someone with healing skills.

  I wondered if the skeleton warrior would actually use the coronet, though. The warrior wasn’t exactly sentient, sapient even, so it may not be aware of the external skill available to it. I used Interface to try to see if I could confirm a connection between the two. Amazingly, my skill did actually show a connection–it wasn’t mana, per se, but something much like what the dwarf commander’s aura skill did. I had Interface update the status screens of the warrior and coronet to more accurately reflect the connection.

  ‘Bound’ was a bit of a misnomer–the coronet was not permanently attached the warrior, as far as I could tell–but I thought it was still the most apt term to use.

  Overall, I was very pleased with my progress and hard work. On the other hand, almost two days had elapsed and I still didn’t have a companion.

  It was time to put my nose to the grindstone and figure out how to get nine into four.