The first method is in Awk itself.

#!/usr/bin/awk -f
#
# USAGE EXAMPLE:
# echo the input of some var
# $ foo -v var=8
#

BEGIN {
    if (length(var) == 0) {
        printf "%s %s\n", ENVIRON["_"], "is a proof-of-concept help message";
        printf "%s\n", "Usage:";
        printf "%s\n", "------";
        printf "%s %s %s\n", "$", ENVIRON["_"], "-v var=NUM";
        printf "%s\n", "substitute NUM with the number you want echoed";
        exit
    }
    else {
        printf "%s %s\n", "You have entered ", var;
    }
}

The disadvantage to this is that it only provides a help message if no option is provided. If you actually type --help, then you get Awk's help message, which is not useful in this context.

The shell script wrapper method uses the shell to parse options, which are then passed to an embedded Awk script:

#!/bin/sh

if [ "${1}" = "--help" -o "${1}" = "-h" -o "${1}" = "" ]; then
    echo "This is a help message."
    exit
fi

/usr/bin/awk -v var="${1}" '

BEGIN {
printf "%s %s\n", "You provided", var;
}'

The disadvantage here is only that you're not just writing an Awk script, you're writing a shell script with embedded Awk. I can't think of a reason not to do it this way (even though in the script that served as the inspiration for this episode, I don't use this method).