A way to search for "cute" domains, without blowing through WHOIS usage limits

cute-domain-search.sh

#!/bin/sh
# "Cute" domain search -- < oogali AT gmail.com >
#
# Example:
#  [oogali@illusion ~]$ ./cute-domain-search am 2
#  IL.AM IS AVAILABLE
#  LE.AM IS AVAILABLE
#  LY.AM IS AVAILABLE
#  OG.AM IS AVAILABLE
#  OL.AM IS AVAILABLE
#  QT.AM IS AVAILABLE
#  ST.AM IS AVAILABLE
#  TI.AM IS AVAILABLE
#
# And yes, it pains me that the documentation is more lines than the
# actual script itself

if [ $# -ne 2 ]; then
  echo "$0 <tld> <# of letters>"
  exit 1
fi

tld=${1}
letters=${2}

rs=`dig ${tld} @a.root-servers.net ns +noadditional +noquestion +nostats +nocomments | grep "^${tld}" | awk '{ print $5 }' | head -1`
for domain in `egrep "^[A-Za-z]{${letters}}${tld}\$" /usr/share/dict/words | tr '[A-Z]' '[a-z]' | sort | uniq | sed "s/${tld}\$//"` ; do
  dig ${domain}.${tld} ns @${rs} | grep -q 'NXDOMAIN' && echo "${domain}.${tld} is available" | tr '[a-z]' '[A-Z]'
done