It merely doesn't prove causation. But it does imply it (as in: hints there might be one), and it's necessary for causation to exist.
Any causation mechanism will also have a correlation in its base.
True - in logic/philosophy, implies means a sufficient basis for so it is just a semantic problem. I wouldn't say that a result can be necessary for a cause though, that is dangerously circular.