One of the main criteria should be to keep it simple!
My thoughts after working my way through the ideas and comments are.
Don't bother with folding wings. Many people have mentioned the reduced drag at that altitude, and it gets less as you go higher. Having the wings and tail permenantly deployed will aid the stability.
Why carry a spaceplane and a balance weight? Might as well make two space planes and increase the chances of a successful flight instead of dead weight. This way you could mount them either side of the balloon via a mesh bag around it, with the bonus that simultanious launches will cancel out the rolling motion caused by firing a single rocket off to the side.
It the mesh bag being loose on the ground when the balloon is small is an issue, small velcro ties or something can be used to gather it up, and will pop open as it expands. The avionics package on the base will keep it upright.
The shape and stability of the balloon shape will be much better at altitude when it has settled and has no other forces on it. Also, the mesh bag with a balloon tight inside it will actually form quite a rigid structure (like an inner tube inside a bike tyre).
There are a lot of valid points about the complexity of a guidence system (even one to fly it bak let alone one that will guide it straight up), and the size and power of the rocket required. I fly RC, but that wont help you much here, but you might want to take a tip from the free flight guys and make it glide in circles. If it flys in a straight line from that height, you might have to catch a ferry to go and get it :-)
I feel like designing one myself now :-)
Good luck chaps!