The first reply should work fine, unless your script generates other warnings unrelated to character encoding, and if you are simply testing for utf8 vs. some different (single-byte) encoding, then Encode::Guess should do nicely as well. Here's another way, which involves specifically testing that a given file is encoded correctly as utf8 data (with no errors, corruptions, or use of non-utf8 characters):

use Encode;

my $filename = "whatever";

eval {
    open my $file, "<:raw", $filename or die $!;
    local $/;
    local $_ = <$file>;
    decode( "utf8", $_, Encode::FB_CROAK );
}
die "$filename is invalid utf8: $@\n" if $@;
[download]

For more info on that, check the Encode man page, esp. the section titled "Handling Malformed Data".