Custom Action Points System - grid move cost callculation

I'm working on a turn based battle system (on hex cells), where every combatant action (such as move, rotation, changing stance such as prone, using abilities, using item from quickslot, using item from inventory/bag, etc) will cost certain amount of Action Points. (Might you ever played UFO: Enemy Unknown or UFO: Terror from the deep, or it's remake Xenonauts, you know what I'm tallking about.)

Most of the cases can be covered using already existing ORK system of ATPs. But since there is no direct access to the ATP value, some interresting options are limited. Those options are mostly comming from idea, that I would like to allow player characters to interrupt their turn - doing some actions, then waiting for other combatants to play and then finishing the moves. And I want to be able to influence the combatant current APs on the fly.

For example, imagine such situation:
1) Player Character A is first to play in the turn. It has 10 APs available at start of each turn. Player decides to apply defence mode for cost of 4 AP, thus boosting it's defence values. Having 6 APs still left, player postpones the rest of the Character A's actions for later in this turn.
2) NPC is next to play. Making some attacks on Character A, using all AP's.
3) Then Player Character B is next to play, who is a support class. He casts "Haste" spell on the Character A, thus doubling it's current APs. Since Character A has already spent 4 APs this turn, it still has 6 APs left. By being boosted by "Haste", he now receives additional 6, thus having 12 APs to spend this turn. Then he use it (to move somewhere, or to attack, or cast a spell, or whatever)... and finishes the turn.

You can think about several similar situations using this concept, such as immediate buffs, debuffs etc.

I think that this cannot be done using current system, since I'm not able able to read and modify actual character ATPs on the fly. What could be done currently is, that when "Haste" is casted on a Character, this character will receive ATP bonus in the next turn(s). But not immediatelly. Or, I can do it, but I will loose the autocallculation of the move cost (I like the curent functionality, that moving through the cells is consuming the ATPs).

In general, IMHO it would be enough simply to have some variable where ORK is keeping the curent ATP value reacheable, in order to be able to read it and to change it when needed - assuming ATP check is done just before every action. And then all this could be done :-)