Well if it's under the solder of course it was during manufacture, but it could be earlier in the process as well. Right now I assume when you say tags you mean pads, the copper surface on the board right ?
It may have well been the photo resist. I worked for a PCB manufacturer for a short time when I was younger and I know cleanliness is esential. Somebody screwed up, and now I would wonder if the problem is not just a one time occurance. If it is on the same part of the board in more than one unit, the process was faulty. It's more likely though that it was just a one time thing, or maybe just a few. When they run these things through a bath maybe the fluid got low or something like that.
In this case I guess the photoresist was conductive enough to let the unit work, and the thermal mass of the solder during the wave soldering was enough to make some kind of bond. It obviously didn't last forever but a soldering iron simply cannot do what a pool of molten solder can.
I would be VERY interested to know if the same problem is found again on different specimens of the same model.
J