the answer hopefully shouldn't shock anyone, but thankfully, buffer size is completely irrelevant at mix time.
a little thought experiment is what i'll demonstrate it with.
suppose buffer size did affect the mix. noticeably or not, as skeksis268 pointed out, phase would be adversely affected. anyone who has mic'd up a guitar cab, with more than 1 mic, knows that mic position affects the overall sound (more so than moving a single mic - moving one mic makes a HUGE difference, as you move them in and out of phase) through waveform cancellation and addition. if buffer size affected this, then you'd have a very noticeably different sound for every buffer size.
in reality, buffer size has no effect at all. since the advent of automatic delay compensation, it's all sample accurate. a mix down at 256 samples is sample for sample identical to any other setting (given everything else is exactly the same).
as for the OP's questions, relating triggering, again, cubase (and most other VST hosts with auto delay compensation) will compensate for the delay of any plugin. even if the plugin runs at a delay larger than the buffer size (for example the SIR impulse loader plugin - runs a fixed delay of several thousand samples), it will compensate for it. so, triggering will always be sample accurate, and what you hear in the live play, is what you'll hear in the mix down. (imagine if it didn't!)
the buffer which you can control the size of, is the final stage that a sample goes, before it is played through the speakers, ie, it's already "mixed down", in a sense. i don't know the internal workings of cubase, but i wouldn't be suprised if the mix-down entirely bypasses the buffer.