That is an interesting idea. Give SpaceX an old satellite sitting in the hangar and see if they can handle the launch. IF they fuck up, all you do is release data on a 10+ year old bird and can use that leak as a red herring to hide what you can actually do. And if they launch and it all works? you get an asset in space and save a ton of money on warehousing the expensive hardware. I mean whatever the payload is, it isn't something the NRO gives much of a shit about.
You're clever, dig into that Spaceflight 101 article. They think it's a leftover version of Worldview 3. Guaranteed, a few days from now the geeks will have NROL-76's orbit. Run the numbers on an optical payload similar to that one and figure out what sort of resolution it gets at perigee. 1) Sometimes the most useful images are the oblique ones and an electrooptical bird looking sidelong is a lot harder to predict than an electrooptical bird looking straight down. 2) off-the-shelf hardware, bespoke performance is kind of a SpaceX mantra. The last KH-11 we put up cost something dumb like $8 billion; if SpaceX can demonstrate a tenth the performance at a hundredth the budget they might get paid next time. I mean, it could be an SDS like they're suggesting. It could be NEMESIS like Ted Molczan is suggesting. But we won't really know until someone picks it up.