I don't know of any command that will do this directly for you - and awk would probably be the way to go, but I'm no awk guru. This could probably also be done through some very impressive sed magic - but again, I'm no sed guru.
So doing this with a shell script, I would say the first logical step would be to get the first line, and the last line - from there we can manipulate those two strings.
FILE=$1 #read the filename as an argument
FIRST_LINE=`head -1 $FILE`
LAST_LINE=`tail -1 $FILE`
The next easy part will be grabbing the first letter, using cut:
FIRST_CHAR=`echo $FIRST_LINE | cut -c 1`
Of course, we could have skipped assigning the FIRST_LINE variable with this:
FIRST_CHAR=`head -1 $FILE | cut -c 1`
Getting the last character is a little more dificult.
num_chars=`echo $LAST_LINE | wc -c` # number of characters in the last line
num_chars=`expr $num_chars - 1` #subtract 1 because counting starts at 1, not 0
LAST_CHAR=`echo $LAST_LINE | cut -c $num_chars`
All in all, we end up with:
FILE=$1 #read the filename as an argument
FIRST_CHAR=`head -1 $FILE | cut -c 1`
LAST_LINE=`tail -1 $FILE`
num_chars=`echo $LAST_LINE | wc -c` # number of characters in the last line
num_chars=`expr $num_chars - 1` #subtract 1 because counting starts at 1, not 0
LAST_CHAR=`echo $LAST_LINE | cut -c $num_chars`
echo "${FIRST_CHAR}${LAST_CHAR}"
#echo "${FIRST_CHAR} ${LAST_CHAR}" #put a space between
#echo "${FIRST_CHAR}\n${LAST_CHAR}" #return on different lines
voila.
Unsolved 
