Recently I had to work a bit more than usual on shell scripting, and had to do a bunch of switches. I used this also in the scripts for this project. A simplified impression shows how we can match user input with supported commands:

#!/bin/sh
CMD=$1
case "$CMD" in
	"run") echo "cmdRun" ;;
	"publish") echo "cmdPublish" ;;
	*) echo "Unknown command $CMD"; exit 1 ;;
esac

This way we can match the provided input with supported cases ("run" and "publish"), and error otherwise (*). The double semicolon (;;) marks the end of a case.