I think there are a lot of variables that would effect what he saw. If the sphere is huge, the light would probably be so diffused that he might see a faint glow in some directions and probably blackness in others. If the sphere was pretty small (and we assume a regular flashlight, not a giant spotlight or anything) I think he would see kind of a blurry reflection of himself in some directions and a reflection of bright light (not any brighter than the bulb) in others. I think he would absorb enough of the light so that it wouldn't really be blinding, unless he lined up the flashlight so that it reflected right into his eye after only a couple reflections without being blocked by his body at all. Mirrors aren't perfectly reflective, so with each reflection some of the light would be lost (actually converted to heat).
Now if we consider a "point astronaut"

then (I think) more of the energy coming from the flashlight would preserved as light instead of converting to heat in which case it might get to a blinding level ... if point astronauts have eyes

.
I don't think the thing would heat up enough to kill the astronaut. Eventually the sphere is going to reach an equilibrium where it's emitting the same amount of energy as is coming out of the flashlight, it's just a matter of how hot it get's before that equilibrium is reached. This is dependent on how much energy is coming out of the flashlight and how well the sphere can release heat. Fortunately for the astronaut, I don't think the flashlight would really be strong enough to kill him. However, if we make the sphere sufficiently insulated ... well, it would be curtains for the astronaut. Of course the astronaut would be creating his own heat as well. Hmm, there must be some point of insulation at which if the astronaut doesn't turn on his flashlight, he won't reach a temperature to kill him, but if he does use his flashlight, he'll overheat though I'm sure it's a very fine line.
Going back to the "point astronaut", when the inside of the sphere reached a point where there was enough light hitting the mirrored surface so that it's reflective inefficiency caused the sphere to lose enough heat to it's surroundings to reach equilibrium, it would no longer get any brighter inside. Adding an astronaut with physical size just gives the system another way to convert light to heat. And (I think), since there would be heat energy inside the sphere that could heat up the sphere shell, helping it to release heat, less light would have to strike it to be converted to heat to reach equilibrium.
So, the smaller the astronaut the brighter the inside of the thing would be.
Also, there are many people on this forum that have much better understanding of physics than I do. So please, if anything I have said isn't accurate, don't have any reservations about letting me know.
edit: Missed the part about it being perfectly reflective. In which case, the mirror would not convert the light to heat. The point astronaut would result in the inside of the sphere becoming increasingly more bright until the flashlight stopped emitting light. The astronaut with physical size would be the sole converter of light to heat (assuming the inside of the sphere is a vacuum, which I think we're supposed to assume because it's in space) and his best bet would be to get as much of himself as possible pressed against the surface of the sphere so that conduction can occur. And turning off the flashlight might not be such a bad idea.