Here's a little example using HTML::Tokeparser. It's not generic enough to grab ANY meta tag attributes, but it'll grab all name-content pairs like you have in your examples.

use HTML::TokeParser;

local $/;
my $content = <DATA>;

my $p = HTML::TokeParser->new(\$content);

my %meta;
while (my $token = $p->get_token) {
        next if $token->[1] ne 'meta' && $token->[0] ne 'S';
        $meta{$token->[2]{name}} = $token->[2]{content};
}

print "$_: $meta{$_}\n" foreach (keys %meta);

__DATA__
<meta name="copyright" content="Aaron Anderson">
<meta name="keywords" content="free, cheap, fun">
[download]

gives me:

keywords: free, cheap, fun
copyright: Aaron Anderson
[download]

Update: Took out lc() from around $token->[1], since all tags are ensured to be lowercase -- so says PodMaster! =)

--
Rock is dead. Long live paper and scissors!