Improve documentation of `--killSignal`

I discovered, almost by accident, something rather unintuitive: the --killSignal=... CLI parameter doesn't have any effect with the forever stop command, as one might expect.

Instead, you have to pass this CLI parameter to the previous forever start command. Then, a subsequent forever stop ... command can be issued (without --killSignal), and it will use the kill-signal as specified in the forever start command.

While that design is unusual and counterintuitive, I suspect maybe there's some deeper reason for it. So all I'm suggesting here is that the documentation -- specifically the CLI --help output -- include this important detail. That might save other folks the 30 minutes I lost.

Happy to PR if this is acceptable.