At the moment I use windows speech recognition which is quite good out of the box. Specifically, I find it quite good for the following:
Does the in built speech recognition support all of the above?
Speech does support dictation, menu item and button selection, and some programs have "speakable items" for their own features.
It doesn't natively support the rest, but you can write scripts or Automator workflows to do what you want, and make them "speakable". When you've finished writing the scripts, use Speech to "Make this speakable" while each file is selected in the Finder, one at a time. This will put an alias in the Speakable Items folder, so make sure you put them somewhere where you won't then rename, move, or delete them.
