#pointing does what ACP normally does before a target (unless its last pointing error is "small enough"). If you have pointing updates / auto-center turned off then #pointing forces it to auto-center anyway.

I assume that when this directive is used, ACP slews the scope to the specified RA and Dec. It then takes an image and plate solves. It determines the center of the image, and calculates the correction to get to the exact RA and Dec, and then does a small mount movement to the exact coordinates.
Yes, except if the mount accepts a SYNC, then it does a SYNC and re-slews. In this case the indicated scope coordinates will be close. Of it can't be synced (TCS has a pointing model and doesn't want to be synced) then yes it jogs to center.

If the scope has consistant approach on, does the correction trip that threshold for an L350? Should Consistant Approach be turned off?
For a Planewave L350, never use Consistent Approach. This is an ancient feature for horribly awful mounts that have so much slop in their gears... well... you get the idea. If I remove it someone will howl, but it should be removed.

Can the plate solve image be from the autoguider camera/camera2?