Depending on how much footage and how much dialogue, the general workflow is to break down the footage into an Event per scene. Then within each scene break down the lines of dialogue of the script into triplets, which correspond to keyword collections within the Event. So you'd have keyword collections within each scripted scene as L1-3, L4-6, etc. Each range within each take would be assigned to these collections. This requires a lot of assistant editor prep work, but it lets you quickly see all the coverage for each section of dialogue across all takes. That's optional, of course, and might be overkill for a short.
Regarding Baselight later... If Baselight has relinking capabilities as good as Resolve, then you should be fine. You would need to consult with your colorist to confirm that they can take an FCPXML list into Baselight and that it can reconnect to the raw files and their file/folder structure. If Baselight needs an EDL instead, you'll need to invest in EDL-X.
As another option, Baselight can take fcp7 XMLs. A shop that we use asks for the XMLs (for which I use Xto7) and then gives me back an XML that I translate back to fcpxml (7toX). But that is how they prefer to work and maybe EDL would be better for the colorist you're using.