This is a rather strange procedure. Or is it the sparse description of what's being done that's causing some confusion?
If you're using "a custom crosscut sled" to make 6mm wide fingers/slots in the ends of a plank intended as a box side then you should have the workpiece firmly located in position by means of a tab-for-the-slot that's a fixed part of the crosscut sled ..... is that right? You move the plank sideways as each slot is cut, to cut the next one, leaving a finger of the same width between the slots as you do so ..... ? And you're pushing the workpiece on the sled just past the crown of the blade then pulling it back, by moving the sled forward then back, so it never reaches the riving knife .... ?
If so, the width of the riving knife is irrelevant as the workpiece never reaches it. If the workpiece is thick, it may reach the riving knife - but the knife is made redundant by the workpiece being kept aligned by the crosscut sled and its tab (and perhaps a clamp).
The sled with its tab will keep the workpiece moving aright as you clasp the workpiece against the fence of your crosscut sled - although you could increase safety by clamping the workpiece to the sled-fence as well, for each push through the blade, so it can't come loose from the sled fence and its locating tab then skew into the blade and kick back.
How do push sticks come into it?
Apologies if these guesses are entirely wrong. A more comprehensive description of your process might help, particularly what you mean exactly by "a through cut". This could mean that you don't go right past the blade or it could mean that the workpiece is thicker than the blade height so only cuts a groove in the underside of the workpiece when passed right over the blade. I presume its the former, if you're making finger joints - what you call "box joints" ...... ?