Heroes 5.5 WarGame Edition - Tactical Edition.

Well, for now there are vacant skills — Barbarian dampening of Light, Darkness, Chaos, and Summoning.
The dampening effect itself definitely won't be present (there are specialized perks for this — corruption, hold, weakness — 3 perks are certainly enough to bring down any school).
As for what will actually happen, that is a question (currently only stats are being proposed).

Barbarian education is also free (currently Barbarians do not have access to it); theoretically, its effect can be completely blocked and rewritten with one's own.
So ideas are welcome.
The only requirement is that the skill must actually play out in combat, such as the new logistics (which will additionally give a starting bonus of 0.1/0.2/0.3 ATP to creatures — and this bonus is especially strong for low-initiative creatures, like undead).

Added after 33 minutes
I once liked the role-playing model from Heroes 4 (compared to Heroes 3). We could consider options from there:

1. Tactics: increase in creature speed and initiative. There is a direct analog — logistics (starting ATP bonuses in the skill, speed and initiative bonuses/penalties in abilities).
2. Defense. Has an analog.
3. Attack. Has an analog.
4. Leadership: increase in luck and morale. There are analogs — Leadership and Luck. In Heroes 4 there is defensive luck (to reduce damage), and there is an analog for this as well (deadly misfortune).
5-8. Combat, Melee, Ranged, Resists — useless, since the hero is not a unit. But there are analogs for boosting hero damage (double damage and stunning strike in the Attack branch).
9. Scouting: increase in vision radius and information. There are analogs in the logistics branch (combat penalties to speed and initiative have also been added via scripts against enemy heroes).
10. Navigation: not needed, as it is too situational.
11. Pathfinding: not needed, as penalties for non-native terrain will be removed (there's no other way with 24 classes).
12. Stealth: cannot be implemented directly, but artifact theft bypassing guards could be added by tying it to corresponding abilities in logistics.
13. Nobility: increase in city growth. For this, the hero needs to be tied to a city; not sure if that's necessary. There is an indirect analog — army gathering, +10% to the size of each squad daily (rounded down). Naturally, level restrictions for the hero should also be added; for example, the total cost of such bonus creatures cannot exceed 100-200 gold per hero level.
14-15. Economy and Mines. Just money and resources. Not sure if this is needed. But as an additional effect, resources could be tied to the dwarf race.
16. Diplomacy: a certain percentage of neutral creatures can be bought out before battle. A direct analog ability is planned (not sure if it'll be possible to implement).
17-26. Skills that grant access to spells of corresponding schools of magic. Skills that strengthen spells of corresponding schools of magic. All of this has a direct analog in the same schools of magic as in Heroes 5.
27-31. Skills that increase mana pool and mana regeneration speed. Essentially, just boosting knowledge and mana; there are analogs.
32. Charisma: same as diplomacy, but in smaller quantities and for free. Uninteresting because diplomacy exists.
33. Resurrection: post-battle losses are partially compensated. Interesting as part of some ability.
34. Necromancy: has a direct analog.
35. Summoning: summons certain creatures into the army. By design, it has analogs for orcs (wolves, manticoras).
36. Sorcery: increase in direct damage from spells. Successor — irresistible magic.

Added after 28 minutes

Yes, I've looked into MMH5.5. In terms of classes/skills, IMHO it's trash — everything is given to everyone, all heroes are samey, linear progression scheme, which is strange because in MMH5.5 there are 8 skill slots.
With such potential, ideally, new skills should be introduced rather than handing out racial ones to the crowd.
There are, for example, wonderful perks from H4 like Nobility and Summoning. A perfectly implementable theme via scripts. And as an option: Ranger starts with Avenger, having a chance to get Summoning (not Summoning magic), Druid — with Summoning, having a chance to get Avenger.

MMH5.5 were pioneers in terms of classes; I'm not sure that without such an example, I would have managed to do better the first time.
Therefore, while criticizing their version, I am glad it exists and allows one to see how things can be done better without repeating their mistakes.

Added after 40 minutes
By the way, I played a game (War and Magic) where 95% of losses are restored after every battle; very interesting idea. It allows for intense PvE battles without counting every penny. This includes auto-battles (where about 20 battles are conducted per cycle).

I think I'll introduce something similar here someday. For example, resurrecting 50% of losses after each battle. This is realistic — battles usually end not when one side is completely destroyed, but when the enemy army is incapacitated. Even in a rout, most of the army survives.

What consequences this could have:

1. PvE battles will feel closer to PvP, where the goal is to win an intense fight rather than reduce losses to zero at any cost. I've noted for myself that I enjoy playing heavy fights where real resistance is felt and losses reach 20-40%. Such losses are naturally unacceptable in PvE, but 10-20% are quite acceptable (leveling up the hero one extra level can compensate for them).

2. Many abilities and abuses that are extremely effective in PvE (shooters, non-retaliators, "ones", resurrections, vampirism) will have their effectiveness halved (in terms of saving losses). This is good for balance because it will allow fighting with full combat compositions rather than pivoting to skeleton archers padded with single ghosts (because if you balance archers and warriors in PvP, then in PvE archers are definitely preferable since they are ranged).

3. All sorts of "cheese" strategies are halved in strength (since 50% of losses return). A boosted lucky Armageddon from a Warlock, for example, could easily wipe out 50% of an army (while consuming all mana); this is normal for full-scale battles, but too much for simple cheese (whereas 25% losses are another matter).

4. Players' ability to restore armies after defeats increases (losing 50% of a force vs losing 100% is a big difference). However, specifically regarding the restoration of losses upon defeat, I'm not sure if this can be elegantly implemented via scripts — for example, returning the hero with 50% of their troops to the tavern, or if they will have to be returned directly to the garrison.

Added after 3 hours 6 minutes
The spell loadout system in duels will be as follows:

1. Each race will have 3 different loadouts, depending on the magic profile. For example, for humans:
- Skilled Light + Basic Darkness
- Skilled Darkness + Basic Light
- Developed Light + Developed Darkness
2. The basic set includes:
- 2 t1 spells (all)
- 1 t2 spell
3. The developed set includes:
- 2 t1 spells (all)
- 1 t2 spell
- 1 t3 spell
- 1 extra spell
- 1 t4 spell (if its cost exceeds 75% of the hero's mana, then another extra spell instead)
4. The skilled set includes:
- 2 t1 spells (all)
- 1 t2 spell
- 1 t3 spell
- 1 t4 spell (if its cost exceeds 75% of the hero's mana, then 1 extra spell instead)
- 1 t5 spell (if its cost exceeds 75% of the hero's mana, then another extra spell instead)

Why is there a maximum cost limit of 75% mana? So that a spell doesn't sit as dead weight in the book. It is not beneficial for a hero to cast spells that immediately consume all their mana (because a combination of cheaper spells will provide more utility).

Thus, in total, there will be 6+3 spells in the book for heroes with a skilled school, and 6+6 spells for heroes with a double developed school.
With this distribution, the chance to encounter a spell in the book is:
t1 - 40-50% (if magic is profile-specific for the race, such heroes will have all t1 spells)
t2 - 15-25%
extra - 15-20%
t3 - 12-15%
t4 - 8-12%
t5 - 4-8%

Extra level spells are special spells that are not in the guild, but are granted when leveling up certain abilities; their power and cost vary from t2 to t4.

Why is t1 more common than t2? In the future, when playing on the map, most heroes will have 1 t1 spell of each profile school in their book, except for specialists whose profile spell tier may be higher. That is, at the start, heroes will have 2 spells (archmages — 4), all of which are t1. Why should there be starting spells? Heroes have magical stats and they should play out immediately without waiting for the guild to be built (just as a hero can immediately go fight with combat stats and a starting army). Why exactly 2? This is the minimum number where there is already a choice, and it's not a chore to open the book for it :) In general, it's clear that t1 should be more common than t2, and this is achieved precisely in this way.