That pin is meant to be bent. For those of you that get one of these badges, you might look really close at the microcontroller. You might wonder "hum, that pin doesn't look like its touching anything"... and you'd be right. That's the pin for VDDCore and if it touches the 3.3 volt trace below it, your badge will not have a good time over USB. It still should power up via battery and you can still flash the chip over the debug connector. However, trying to edit the CircuitPython code via the mounted USB drive will either show up as an unknown device or lock up while editing files.
How did we get here? So it seems I copied the initial schematic from our previous CircuitPython badge design we used at the local makerspace. This was before my friend Bill noticed our mistake of tying all the VDD lines together. Apparently VDDCore is not like all the other VDD lines on the SAMD21 and has an internal regulator to drop it to 1.2 volts instead of the 3.3 volts. He fixed this on our previous boards because he's a good person and actually reads the datasheet before ordering PCBs. But this should have been caught in the one run of prototype boards for this project... except with my totally professional hand soldering skills, it turns out I missed soldering just that one pin. Everything worked out fine on the prototype board, so I just made my other changes and ordered the final 100 boards.
Fortunately it seems for what we're doing, everything works fine if that pin is left disconnected. Normally I'd try to cut the trace on the PCB, except for that trace is really short and very close to traces that are needed. It was much faster and more reliable to just bend that pin up so that it doesn't make contact with the pad. Plus if you want to match the datasheet, you could get a 0402 cap in there between the bent pin and the ground pin right next to it... maybe that the solution for Challenge #2... or maybe I just need to RTFM. I'm sensing a trend here.