It seems like you want the brush to not be as efficient as maual reclaim, but then the whole "people don't need to manual reclaim anymore" falls flat, as they will still see the pros do it and think it is important. If it's not supposed to completely replace manual reclaim it's just redundant because it provides a middle ground to two options which both have their role. We don't need another option that is a bit of both. One of the extremes will always be the better option.
Regardless of the above, I feel this feature would just collapse because of the processing power requirements that it will have.
You claim that is possible to implement it in a way that is not laggy, while still providing better pathfinding to the reclaim targets than attack move. Yet when looking at the details all evidence seems to point to the contrary. At this point you need to go into more details how you think you can solve this and why it would still be a good feature even with the limitations you need instead of handwaving the problems away with "we'll limit the number of commands somehow".