I don't think you're doing anything wrong, but a few thoughts.
The main purpose of phantom power is to support condenser mics, so, right, you shouldn't have to think about it when using dynamics like the Shure MV7.
I haven't used an MV7, but I've used the Shure SM58 and SM7b a lot in the past. They're some of the quietest dynamic mics out there, so I'm not surprised you're seeing waveforms like that.
There are basic questions of technique like whether you're speaking at an ideal distance (I assume somewhere between making contact with the mic's foam and twelve inches away), and how loudly you're speaking; no doubt you've considered that.
It's a problem to "clip," or produce waveforms that are so tall that they're forced to flatten at the top, because that produces irreversible distortion. It's not such a problem to have unusually short waveforms like you're getting (with a lot of "headroom"), because you can always use post-processing to raise the levels afterward – so it's preferable to have waveforms that are too short than too tall. You might lose fidelity if you try to use post-processing on extremely quiet recordings, but the regions in your screen shot look fine to me.
For quieter dynamic mics, there are devices called mic "activators," "boosters," or "preamplifiers," whose function is to convert phantom power to a boost in gain before the signal reaches the XLR interface. (The Cloudlifter CL-1 and the Fethead are examples.) With the aforementioned plug-ins available like those in GarageBand, I don't think you'll need a device like this, but you can look these up as a point of interest, as affirmation that some of these mics are so quiet that such devices were commoner in the pre-digital era.
Good luck with the production.