They might as well just pop a PDF of the user manual up on FileBay if they've made it so that the function is easily discernible from the outside *by anyone* with whatever outer shell they put on it.* Except for the cost of strolling up to take a closer look, they've done the equivalent of having it tour the world's main streets unguarded. It's literally in everybody's territory and there's nothing we can do if any of them want to launch a camera satellite just to take a tour of everything in a given orbit. There's probably something in some treaties that gives us some recourse if someone launches a team to pop the hood and dig through the guts, but AFAIK, as long as they don't touch it, they can look all they want.
Sure, it's expensive to go look, but we're not exactly worried about Somalia here. Russia, China and the few others we are concerned about can afford to do something like that, have the people who could figure out the most by looking, and I'd be surprised if they haven't done it before at some point. After all, they already have satellites in various orbits, so how hard would it be to release an extra payload that's essentially just a remote camera?
There's actually a LOT that those with the skills can tell about a spacecraft or satellite just from a picture. And generally speaking, there is ZERO wiggle room to "just put an obfuscating shell on a satellite" every last ounce of it has a mission critical purpose.* Because every ounce more it weighs means more fuel and a larger rocket to launch it, more fuel for it's RCS/position-keeping system to thrust it around, and monkeys with it's thermal managment in the sun and shade of earth, and the shell and the gear and that extra fuel means less actual gear that supports the satellites main function.
And having another satellite take a picture of it with a close pass is certainly something that's been done several times before, but it is not something you can do casually. The satellite taking the picture has to be within a certain distance, or even the angular resolution of even the best cameras and optics won't see anything, then there's the two relative velocities of the satellites, and all you're likely to get is a dash-shaped line of light as they speed by one another at a relative closing speed of several kilometers a second. And then there's the issue of if you'll even be able to see anything if you can control for all those other factors too. Because getting an f-stop exposure that will show you both the sunlit and unlit parts of the target satellite may not be possible within the focal length of optics you need to see the satellite at a distance of a few hundred or maybe a thousand miles. (Which is kissing-cousin close in space. even near-earth orbital terms...) And then there's the issue of if a satellite is matching the orbit of one of ours that closely, NORAD and the space tracking system, or their counterparts in Russia and China
will know about it. And all that has to be done is either tell the suspect country (assuming it hasn't been watched from launch, which it always is...) to GTFO, or, simply use the satellite's RCS to modify it's orbit so the spy satellite trying to get a look-see has to chase it, or even just throw their satellite into an earth descending orbit, or out into a useless orbit that will ruin it's mission, if ruining it's mission is less important than not letting an enemy examine your satellite.
So from earth-bound photos from the satellite in the clean room, or a publicity photo showing the mate-up between the satellite and the first stage, from the size, shape, type, and numbers of antennas, they can get a very good idea of what frequencies it'll operate on. If any camera apertures are visible, they can get a very good idea of the focal length, and magnification that the satellite is capable of.
From it's size, number, and position of solar panels (if any) they can determine it's operating power budget, and get a pretty good ballpark idea of how much RCS delta-V it has to change or modify it's orbit.
The physics of spaceflight, and that of RF communication, and optics are very constrained, and you can get an almost complete picture of what a device does or is capable of if you can plug in a few variables, or even just a reasonable ballpark range of variables gleaned from something as simple as a photograph.
And yes, if I had to place bets between SpaceX and Grumman/.gov screwing up, I know who I'd bet on. And it's not the company that can take a freaking rocket,
get it back, and land it on it's tail within a few feet of accuracy. *one caveat there, is that it's almost certain that both the U.S. and Soviets/Russia have experimented with cone shaped balloons, or an "umbrella" that a satellite can hide behind from at least someone looking up at it below from Earth, either optically, or with radar, but the aforementioned weight penalty is just too great, even if the shield or blind is made of very thin lightweight materials. Because the struts or sticks to attach it to the satellite, have weight, the system that inflates or deploys it has weight, the control system adds complexity to the satellite's computers... and again, every ounce of a "shield" is an ounce of fuel, or an ounce of mission critical equipment on that satellite you're now missing. And further, advances in infra-red photography, like going back to the late 80's or early 90's, God knows what 2000's era tech is capable of now.. have made such a black blind or shield unworkable anyway. Because if we can see the moons of Pluto in infra-red all the way from Earth, and detect .001 degree differences in the 3 degree above absolute zero cosmic background radiation with "civilian" technology... a spy satellite has
zero chance of hiding just 200-300 miles up.