You should validate the data to ensure you are getting only the number of files you expect.

   #!/Perl -T

   $what_i_am_about_to_type_at_prompt = 'file to remove.gah';

   print "Please enter the file to be removed: \n";
   my @file_I_intend_to_remove =  <STDIN>;

   @file_I_intend_to_remove == 1 or die 'incorrect number of files sup
+plied';

   my $file_i_am_about_to_unlink = shift @file_I_intend_to_remove;

   unless( $file_i_am_about_to_unlink eq $what_i_am_about_to_type_at_p
+rompt ){
      printf "%s\n", $file_i_am_about_to_unlink;
      die 'for the love of ice-cream, don't do an unlink operation jus
+t yet'; 
   }

   # even if it matches you still need to untaint it.
   $file_i_am_about_to_unlink =~ m/\A(file\sto\sremove\.gah)\Z/
    or die 'taint is a pain in the hyperbola, but I enjoy having a wor
+king operating system';
   
   $file_i_am_about_to_unlink = $1;

   # now i can unlink the file as it is the one i am expecting to unli
+nk;
   printf "unlinking %s\n", $file_i_am_about_to_unlink;
   # unlink $file_i_am_about_to_unlink;
[download]

The context you supply to the input operator, determines whether it returns the first field, or all fields. The input operator splits on IFS the Input Field Separator, $/. You do not need to call split on the input operator.

You do need to check what your input is once it is in the array, and before you pass that input to a function. This is data-validation, more commonly known as untainting. For why we do this see perldoc perlsec. To scrub the data, you need to substitute a pattern match back into your scalar.

perlsec is dense. keep it simple and only use straightforward file names while testing.

my $filename = 'filename.fxt';
my $scrubbed = s/\A(filename\.fxt)\Z/$1/
 or die 'need to read up on regex a bit more';
[download]

Common error is to forget to chomp the string, or to try and use filenames with non-printable characters in them, such as spaces. Review chapter 8 to cover what you will need for this.