As a counter anecdote, I ordered ("[s]hips from and sold by Amazon.com", not "[f]ulfilled by Amazon"; just double-checked) a branded microSD card a few months ago, and did a full device write / read on it, and it died on the read pass. I wasn't even thinking that it could be counterfeit at the time: this is just a basic smoke test I do on all new storage devices. But now that I think of it it does seem awfully suspicious.
Yeah, it's very common that rejected cards will find their way back into the supply chain, or that employees will run "ghost shifts" with rejected sub-standard wafers. Bunnie Huang did a fascinating teardown on some fake cards.
There's no official Win32 distribution but it compiles and runs without issue on Cygwin. When I get home, I can send you a Dropbox link with a build and enough Cygwin DLLs to make it run in a standard Windows command prompt.
As mentioned on the f3write site - one of the very typical hacks is to make the controller report a larger capacity than it actually has - or a larger capacity than is actually functional/reliable. So this is the obvious test to perform - write it full and see if you can read it out.
Amazon will co-mingle their stock with third parties as long as the stocking unit (typically UPC) matches. So if some retailer sends them fake memory cards you get a fake memory card, "Sold And Shipped By Amazon.com". And of course eBay is full of this junk.
At this point I prefer to buy from camera stores. My theory is that since their customers will be making their livelihood off the memory cards that they have an additional incentive to keep their supply chain clean.
