I don’t think anyone is making any strong accusations, but the most plausible cause is accidental user input.
Here is the GCS message timeline showing autotune completion and saved gains after landing, indicating that autotune parameter saving completed at time 11:40:07:
2023-08-16 11:39:31.291 AutoTune: Success
2023-08-16 11:39:31.291 Event: DATA_AUTOTUNE_SUCCESS
2023-08-16 11:40:04.926 Event: DATA_LAND_COMPLETE_MAYBE
2023-08-16 11:40:05.728 Event: DATA_LAND_COMPLETE
2023-08-16 11:40:07.661 Event: DATA_DISARMED
2023-08-16 11:40:07.661 AutoTune: Saved gains for Roll
2023-08-16 11:40:07.661 Event: DATA_AUTOTUNE_SAVEDGAINS
2023-08-16 11:40:07.662 Event: DATA_MOTORS_INTERLOCK_DISABLED
2023-08-16 11:41:09.161 PreArm: Motors: SPIN_ARM higher than SPIN_MIN
2023-08-16 11:41:40.161 PreArm: Motors: SPIN_ARM higher than SPIN_MIN
Here is the timeline for changes to MOT_SPIN_ARM at time 11:41:04, a full minute after autotune completed its parameter changes:
2023-08-16 11:41:04.678 MOT_SPIN_ARM 0.100000 -> 0.230000
2023-08-16 11:41:04.759 MOT_SPIN_ARM 0.230000 -> 0.220000
2023-08-16 11:41:05.069 MOT_SPIN_ARM 0.220000 -> 0.230000
A brief search through the Copter-4.3 branch shows no automatic routine that appears to change MOT_SPIN_ARM, and no mention of that parameter in the autotune libraries. Admittedly, my search was only cursory.
Is it possible you accidentally changed MOT_SPIN_ARM (maybe intending to change MOT_THST_HOVER, whose value is close to 0.22-0.23) or clicked on something like the motor test page and mistakenly made an input?