If you have C6-3 in the side boosters and C6-0 in the central, then the upper stage, if it ignites at all, will do so before the side boosters detach. So the whole assembly of central booster plus both side boosters will detach as the upper stage ignites, then the side booster ejection charges will fire and the side boosters will separate while the whole booster assembly is on its way down.
I'd put B6-0s into the side boosters. Although it has no official ejection charge, the forward pressure at burnout is enough to blow off a loose nose cone and eject a streamer (or to blow off an entire upper stage, which is why you need to vent the gases or tape the booster and sustainer engines together to prevent the upper stage from separating before it can ignite). I have a parallel-staged model of my own which relies on this fact to detach its boosters and eject streamers while the sustainer continues on its way, though the sustainer is not staged.
As for recovery of the central booster, here's what I did for Rheinbote, the four-stage rocket which you see in my profile picture:
View attachment 364901 View attachment 364902
At the top of two of the boosters is a thinner tube through which the burning particles can pass to ignite the next stage. Venting is through the rear; the motor mount is centred by strips of balsa, not centring rings, so the gas is not confined. Wrapped round the thin tube at the top is a small Nomex streamer. This has to be fireproof because any wadding wrapped round the streamer will disappear the moment the next stage separates, leaving the streamer right in the exhaust of the next stage. (I tried it once with a mylar streamer wrapped with wadding, and switched to Nomex after seeing the crispy remains of the mylar.
) So the streamer isn't long enough to be a proper recovery system but is enough to destabilise the booster, making it tumble rather than dive down balllistically. (You have room for a longer streamer because your outer body tube is BT-60, while mine is BT-55.)