I’ve been following this forum for several years, and the behavior outlined by chrywo is new to me. While it is theoretically possible that this is a firmware issue, since the 513 version of the firmware has been out for nine months and used by hundreds (thousands?) of people, it is highly unlikely the issue is firmware. My best guess is that the pins are the board are bad (not responding to a pullup setting) either due to a manufacturer error, or there was a wiring issue at some point that damaged the pins.
I suggest your first step be verifying this is the identical problem. Carefully check the voltage of each of the signal pins with the board booted. It should be 3.3V. It was around 2.0V for chrywo. Verify (M119) with Dupont wires and direct wire connections if the problem still occurs. This second test takes all of your wiring and switches out of the equation.
If the following two test follow chrywo’s results, a third test would be to wire a pullup resistor between VCC and the signal pin and retest. If the problem is just a pullup resistor issue, adding an external pullup resistor should fix the issue.
There is a fourth thing that can be tried. It is possible in the firmware to reassign endstops to alternate pins. This is something I’ve seen done once for the Rambo board with a bad pin, but I’ve not seen it done on an SKR Pro, but it should work.