@FramedArchtiect
Since you seem to be interested in feedback, i'll give some general ideas, which i think would be useful to keep in mind:
-ingame descriptions should be long and clear enough in respect to game effects that testing is minimized
e.g. evocation bloodbriars in game description says something like "Cast a spell that allows cities to claim nearby resources faster." That contains practically no information about what the game effect would be. Faster border expansion for cities (e.g. like Krepost special effect from normal game)? One time border boost? Reduced gold costs for buying squares? One would have to test to find out, although the effect is in principle known from mod code.
e.g. evocation wall of fire in game description neither in policy selection screen nor in help menu says only 3 can be built. And the building description of wall of fire does not say clearly whether 3 at the same time existing is the limit or whether its in total only 3 may ever be built (which makes a big difference and again would have to be tested).
e.e. abjuration sanctuary (at least i think so) says something like "Cast a spell to boost strength of all units near capital". That means one should gather units then select it and then use the upped units to conquer some neighbors, as its a one time strength boost for the units currently gathered round capital? Or is it a permanent ability of the capital, which increases strength for all units as long as they are in certain radius? Which radius? What boost exactly, relative or absolute (e.g. +10% or +1 strength)?
There are further examples, these are listed just as an example what the idea is, a lot of special content is nice, but frustrating if one has to test to know the game effects.
-try to avoid many one time boost, especially those, that are stronlgy dependent on timing
It is not fun, if for getting a decent benefit one has to micro a lot.
E.g. evocation command plants, if it is selected when all cities are just 1 turn away from growing its effect is a one time boost of a few food per city. If it is selected just after all cities have grown it can be a bonus worth several thousand points of food. Same if some settlers are moving aroung, just before they settle its a non-boost, directly afterwards its a great boost. Without changing much on the other hand this could be avoided, for example instead of filling granaries up, command plants provides 3/4*the amount of food the full granary would have. Then its less relevant what the exact granary filling of each city currently is. As a further modification, if possible, it also provides a similar food bonus whenever a new city is founded, then even the timing of settlers is less relevant and actually its never a bad time to pick the ability. Yet the overall game effect is similar.
-keep AI in mind or more precisely try to model buildings, units, magical paths and so such way, that the AI is not total failure in using them
e.g. just the above, i suspect that the AI does not "know" the bonus command plants give, so if he will ever get it, it will be completely random whether its a boost or not.
Another example is the wall of fire, i suspect the AI cannot handle the concept of a defensive building that can only be build a limited number of times, as standard game does not contain such buildings. So rather likely the AI will build them anywhere, instead of where they are needed. Unless its absolutely desired a lesser bonus in all cities would be easier for the AI to handle and therefore preferable.
Obviously with some things, its just bad luck for the AI, but if for something variations are possible, the one best usable for the AI should be taken.
-whenever possible magical and mundane should be distinct or in other words they should feel differently
This depends a bit on the design goal, but to explain the idea, its not fascinating to play a fantasy mod, in which a late game unit is an uber dragon, which can operate only from cities, has 10 tiles range, 95 bombard attack, cannot be intercepted and requires the late game resource alamantium, because then its just a renamed stealth bomber. Or when can build a swordsman with strength 14 and nothing else and then later can build a demon with strength 21 and nothing else. There is no magic, just that instead of longswordman its called demon.
If you know the games Masters of Magic and Age of Wonders i think the point would be even clearer. In Master of Magic there are mundane units, for which the advanced versions mostly have improved statistics, but normally nothing special, and then there are magic creatures which can have abilities like stone gaze, flying, regeneration, teleport, weapon immunity, which are mostly completly absent from mundane units. In Age of Wonders the first unit of a certain race and of a certain magic path are regenerating zombies and net throwing posisinous wall climbing spiders and the high end units are black breath bone dragons and a lifestealing grim reapers - it does not feel in anyway special to have a magical unit and if i didnt remeber i would not know which of regenerating zombies, poisonous wall climbing spiders, bone dragons and grim reapers are the "mundane" units and which require "magic".
For this mod, this means that unique promotions should be only avaiable to units that are tied to some of the magic or alignment choice, e.g. no unit that is simply avaiable in the general tech tree should as standard have teleport, summon and such. The units from the general tech tree should without specific magical path choices just have promotions similar to the ancient to middle age units of standard civ. (though as far as i can see, this is currently so, just keep it)
(Ok, enough wall of text.)