My immediate thought was to mask off the unaffected area and use some silver conductive paint but I'm not a EE, there may be a better solution.

Flux the hell out of it (liquid flux works better for this) and just run the soldering iron tip with a really, really tiny amount of solder on over it. The contacts are so small that it will definitely bridge and the flux will prevent joining them all together.

That doesn't look like the pad is broken and the two pads aren't shorting either. You would still want to push the gap between the two pads clean though, a knife would work there but be careful, this is a multilayer board and if you manage to dig into the board you'll likely expose one of the inner layers and then things can get pretty iffy. If the pad actually is broken don't bother trying to fix it with solder or paints unless you want to ruin that pin on your memory socket.

Well I gave it a shot but I guess it's just a junk module. Massaged the suspect pin with a knife and tried it again in a Toshiba Portege M200 (with 855PM rev. B1 chipset). This laptop has two slots and with a known-good DIMM in the other slot the system would at least POST. It recognized the additional RAM, but would not boot Windows, only DOS, and there was garbage on the screen in text mode.

So I went a step further and taped-off the adjacent pins, applied a tiny amount of flux, and managed to cover the "hole" with solder. Amazed that it worked on the first try, and since the solder was only on the upper half of the pin I thought it was still safe to insert as long as it didn't go in too deep. But this mod didn't yield any improvement, the Toshiba still acted the same as before, only booting to DOS and with screen corruption in text mode. I didn't have any program handy that would test the memory from DOS, but I'm assuming it would have failed.

Ultimately I wasn't able to verify whether the 855PM rev. A3 chipset in the NEC would recognize a 1GB stick.