When you say that they are derived from the same clock and that the positive edges are guaranteed to be aligned, do you mean that they arrive at the FPGA like that, or that they are guaranteed to arrive at the sampling flip flops like that?
If these clocks start out aligned, but are distributed by different clock trees, there is no guarantee that they will stay aligned.
So, if you know for a fact that these clocks arrive at the sampling block aligned, you could probably dispense with metastability hardening. If you can afford the extra timing hit though, it is better to be safe than sorry and sample them.
r.b.