As also indicated in a comment to the issue; I don't think current implementation can handle the TSMC antenna rules. The rules use only the perimeter of the metal to compute the ratio and does not use the area.
I just did a 0.18µm TSMC tape-out using Europractice as service. I don't have access to Calibre and had to do a few iterations to get antenna problems solved. At first sight I did not understand why there was a mismatch between antenna check results…