One word spam and possible honeypot bug

Hi Benjamin,

It is always a pleasure to hear from you! Thank you so much for your continued support and for taking the time to report this; users like you really help me improve the plugin.

Yes, there is actually a quick way to solve this immediately.

I believe the reason the honeypot seems to “pass” is likely because this specific form has very few text or number inputs (fewer than 3). Currently, the validation logic relies on a cumulative “spam score.” If there are very few fields, the total score might not reach the threshold required to trigger the spam flag.

Here is how to fix it:

1. Adjust the Spam Score: You should adjust the spam test value settings. Try setting it to 0.5 if you have two text inputs, or even 1 if you only have one. As a reminder, an email is flagged as spam if the score is greater than or equal to 1.
2. Filter “Gibberish” (One-word messages): To stop the random one-word strings, we can use a custom filter in your functions.php file. We can set a rule where if the message body contains fewer than X words (e.g., 3 words), it is marked as invalid.

Here is a snippet you can use:

// Example logic for your functions.php
add_filter( 'cf7a_spam_check_chain', 'filter_short_messages', 20 );

function filter_short_messages( $spam_data ) {
    // If explicitly whitelisted by previous filters, skip.
    if ( $spam_data['is_whitelisted'] ) {
        return $spam_data;
    }

    // Count words in the message body
    $word_count = str_word_count( $spam_data['message'] );
    
    // If less than 3 words, treat as spam
    if ( $word_count < 3 ) {
        $spam_data['spam_score'] += 5; // Add points to ensure it fails
        $spam_data['is_spam'] = true;  // Flag as spam
        $spam_data['reasons']['short_message'] = "Message too short (less than 3 words).";
    }

    return $spam_data;
}

For additional information please refer to: https://modul-r.codekraft.it/2025/11/developer-guide-customizing-cf7-antispam-filters/

I am currently planning to implement a “proportional spam rank” in a future update to handle cases with few inputs automatically, so manual adjustments won’t be necessary down the road.

Let me know if this solves the issue for your client! (and sorry for the late reply)

Best regards, Erik! 🙂