Messages and the last read log ID from vmware are stored in the VMware cache for the url, not for the item. This is reasonable, since if you have several items that are filled from one log, then it is correct only once to perform a heavy / expensive operation to get a log from vmware
But:
If you create another item with the skip parameter for the same url, then the proxy / server will still rely on the last log ID saved in the cache. Those. skip does not work for all subsequent (newly) created items for the same URL

At the same time, it is unlikely that someone really needs old (and very old) logs from vmware. The probability of getting performance problems of the monitoring system itself when processing a large log again is also very high
But the situation when the skip was initially missed by mistake is very likely

I propose to change the default logic to the opposite: if a skip is specified, then it is applied for all eventlog items of this URL. => Skip all 'old' logs for his eventlog

Or at least add a clear description it works and clear instructions how to add new item with skip to the documentation