How do I check for spam using the async SpamAssassin and Rspamd interfaces?

Two async spam-checking variants are available, both used in filter_end:

md_async_spamc_check - sends the message to spamd using the raw SPAMC wire protocol over an async socket. Does not require Mail::SpamAssassin::Client.
md_async_rspamd_check - POSTs the message to the Rspamd HTTP API over an async TCP socket. Does not require LWP::UserAgent.
md_async_spam_assassin_check - runs SpamAssassin in-process (no spamd), reading ./INPUTMSG. Use when spamd is not available.

Each returns the same values as its synchronous counterpart.

use Mail::MIMEDefang::Async;

sub filter_end {
    my($entity) = @_;

    # --- Option A: async spamc (needs a running spamd) ---
    my ($score, $threshold, $report, $isspam) = md_async_spamc_check(
        host    => '127.0.0.1',
        port    => 783,
        user    => 'defang',
        timeout => 30,
    );

    if (defined $score) {
        action_change_header('X-Spam-Flag',  $isspam  ? 'YES' : 'NO');
        action_change_header('X-Spam-Score', "$score / $threshold");
        if ($isspam && $score >= 10) {
            return action_bounce("Message rejected as spam (score $score)");
        }
    }

    # --- Option B: async Rspamd ---
    my ($hits, $req, $tests, $rpt, $action, $is_spam) =
        md_async_rspamd_check('http://127.0.0.1:11333');

    md_syslog('Info',
        "Rspamd: score=$hits/$req action=$action spam=$is_spam tests=$tests");

    if ($action eq 'reject') {
        return action_bounce("Message rejected by Rspamd (score $hits)");
    }
    if ($action eq 'add header' || $is_spam eq 'true') {
        action_change_header('X-Spam-Score', "$hits/$req $tests");
    }

    # --- Option C: in-process SpamAssassin (no spamd required) ---
    # my ($hits, $required, $names, $full_report) =
    #     md_async_spam_assassin_check();
}