In my opinion, I don't like doing much (if any) EQing using outboard gear on any of the signals as they're being recorded. I say this because if you make a bad decision on your EQ settings, you can't edit it in the software after its been recorded. You have to re-record the track to make any changes.
But if you EQ relatively flat, then you can change and redo the EQ settings in your software all you want without having to re-record the track. Same goes for compression.
The exception to this is when you want to use EQ or any other effect to shape the way your guitar amp reacts to your playing. Especially tube amps, where they really change their distortion character depending on what EQ curve you feed the amp's front end.
Another example is if you use a compressor to make your guitar's signal strength consistently remain in your tube amp's "sweet spot" of breakup/overdrive.
Also with compression, a lot of people will use it while tracking to tame a signal whose wide dynamic range makes it difficult to keep from clipping the inputs of their digital audio interface while also retaining a good signal strength (and therefore a good signal-to-noise ratio).
So that's the main thing to consider here: any signal processing you apply to your signal before it gets sent to your recording software is there for keeps and can't be adjusted afterwards. Any VST effects you apply after recording can be tweaked and retweaked to your heart's content.
