At long range the missiles should be pretty much invisible
Depends. If you're dumping twenty missiles then that's 2 dTons of volume, and a much higher surface area than a single 2 dTon block, so you might get more of a return than you'd expect.
That's why I'd just incorporate the cost-per-volume for 'stealthy hull materials', straight out of the basic ship design rules. Then there's no argument that the pc's should have seen them before the GM says so.
If you have to worry about that, your rocket is insufficient for ship to ship action anyway.
Depends on what you consider the missile launcher to consist of. If it's just a rail that holds the missile whilst it accellerates from a standing start, and it can be programmed with a number of waypoints for the first stage of the flight, you can point the launcher in whatever direction you like, and use short, low-power burns to get yourself clear of the launching vessel and lined up on the target before hitting the gas on your main drive. Yes, it will take time, but not enough to register in a six-minute turn, and with a bit of intelligent Time-on-target work you can ripple-fire launchers in multiple directons so they form up into one big barrage when their drives light up.
However, if the launcher is something more than that (the canister contains a quick-burn drive, or is a small mass driver), then you will actually leave the launcher doing quite a sizeable clip
in the wrong direction. You can't run 'out of range' of a ship to ship missile per se, but if you've got a mass driver launcher, linked up to the power plant and designed to punch you up to - say - twenty to thirty percent of your peak speed, not having that pointed according to an intercept solution is going to seriously impact your missile's performance.