That's a cool setup, I did something similar with a Pegasus Pocket Powerbox under my saddle.
I doubt this is the issue, but for clarity I am talking about the wires that go through the mount. Those wires will not snag in the conventional sense, but do need enough slack that they smoothly turns inside the axis as the axis turns. If it is tight, as it turns, it will stick and release a bit on insides of the axis. So all I was suggesting is make sure where it comes out from the mount on the saddle and on the bottom if you tug a bit on it with your fingers there's just a bit of slack.
Unrelated: In looking at the calibration, both axis are a bit more off the lines than mine. I have encoders, so that might be the reason, I have no other experience as have not tried turning them off. Below is my last calibration run.
I'm clueless what it would suggest if yours is abnormal; it is not backlash since it is on RA also. I'll let someone smarter comment more fully on the guide logs, but the RA axis drift looks larger than I would have expected (which could be calibration especially since it is similar to DEC), but again I could be jaded by encoders.