Team beta version

Note: This plugin requires HTTPS to function (due to Web Crypto API usage).
The Story: The Forging of the Ultimate Defense
This project began with a simple goal: to create a lightweight honeypot trap to catch basic bots. But during development, we faced a harsh reality: Simple traps are obsolete. Modern AI bots and headless browsers can easily step over traditional defenses. A basic trap was no longer enough; we needed a fortress.
So, we forged a completely new architecture. We added Proof of Work, Behavioral Analysis, and Rate Limiting. What started as a simple honeypot evolved into a 15-layer invisible firewall. Like a samurai’s blade, it operates with absolute precision—completely invisible to your real customers, yet ruthlessly executing a “Silent Kill” on spam bots before they ever reach your inbox or database.
Samurai Honeypot protects every Contact Form 7 and WPForms form with fifteen independent defense layers. 
Each layer contributes a score, and blocked submissions are handled by a 3-Tier Triage System:
<h4>3-Tier Triage System</h4>
<ul>
<li>Tier 1 (Pass): Score below the threshold (default 50) — email is sent normally.</li>
<li>Tier 2 (Quarantine): Score between the threshold and 99 — email is silently suppressed (Silent Kill) and the submission is saved to the built-in Quarantine Log for admin review.</li>
<li>Tier 3 (Drop / Instant Kill): Score of 100 or higher — email is silently suppressed and the submission is permanently dropped without logging. This protects your database from bloat during DDoS or mass bot attacks.</li>
</ul>
Warning — False Positives: Depending on your environment, legitimate emails may occasionally be flagged as spam. Enterprise proxies, strict corporate firewalls, outdated browsers, VPNs, and unusual network configurations can trigger detection layers. You MUST check the Quarantine Log periodically to identify and recover any false positives. The plugin cannot distinguish all edge cases automatically.
<h4>Defense Layers</h4>
<ol>
<li>JS Injection Gate — Blocks bots that cannot execute JavaScript. Tokens are fetched via REST API for full page-cache compatibility.</li>
<li>Polymorphic Honeypot — Decoy field name is cryptographically derived per token (not exposed in the API response), hidden from humans via CSS.</li>
<li>Token Signature — HMAC-SHA256 signed stateless token with IP and Form ID binding.</li>
<li>Time Trap — Detects impossibly fast submissions. Browser autofill is automatically exempt.</li>
<li>Proof of Work — SHA-256 computational challenge via Web Crypto API that forces CPU cost on bots.</li>
<li>Behavioral Entropy — Hash-verified human-like event counters: mouse, keyboard, touch, scroll. Uniqueness tracking detects script reuse.</li>
<li>Headless UA Block — Server-side User-Agent check instantly blocks known headless browsers and automated tools (Headless Chrome, Puppeteer, PhantomJS, Selenium, Playwright, Nightmare, Electron). Toggleable in settings for E2E testing compatibility.</li>
<li>Headless Detection — Detects automated browser environments (navigator.webdriver, plugin count, window.chrome, language count).</li>
<li>UA Age Detection — Scores based on Chrome version age. 2+ years: +10, 3+ years: +20, 4+ years: +30. Bots often use hardcoded old User-Agent strings that never update.</li>
<li>Rate Limiting — Per-IP submission rate limiting with IPv6 /64 normalization.</li>
<li>Replay Protection — Atomic token consumption (INSERT IGNORE) + TTL expiry enforcement.</li>
<li>IP Blacklist — Manually configured IP/CIDR blacklist for known bad actors.</li>
<li>Content: URL Limit — Flags messages containing more URLs than the configured threshold.</li>
<li>Content: BBCode — Detects BBCode link syntax (<code>[url=...]</code>) that never appears in legitimate form submissions.</li>
<li>Content: Denylist — Matches against WordPress Disallowed Comment Keys (Settings > Discussion).</li>
</ol>
<h4>Key Features</h4>
<ul>
<li>3-Tier Triage — Pass, Quarantine (with local log), or Drop. No legitimate message is lost without a trace — quarantined submissions are saved for admin review.</li>
<li>Built-in Quarantine Log — Blocked Tier 2 submissions are saved to a local database table (up to 1,000 entries, FIFO). View date, score, trigger reasons, and full form data from the admin panel. No external plugin required.</li>
<li>DDoS-Resilient Tier 3 Drop — Submissions scoring 100+ are immediately dropped from memory without any database write. This prevents database exhaustion during mass bot attacks.</li>
<li>GDPR Compliant — No cookies, no external service calls, no plugin-specific PII stored. IP addresses are one-way hashed with a site-specific salt before any storage — raw IPs never touch the database. No consent banner required.</li>
<li>Stateless Tokens — No database writes for token generation; prevents DoS via DB bloat.</li>
<li>Zero Trust Client — All client-submitted data is verified server-side with HMAC signatures and hash integrity checks.</li>
<li>Zero Configuration — Activate and all Contact Form 7 / WPForms forms are protected automatically.</li>
<li>Cache Compatible — Tokens are fetched via REST API, so page caching works fine.</li>
<li>Multiple Forms — Works correctly with multiple forms on the same page.</li>
<li>IP Whitelist / Blacklist — Whitelist trusted IPs or CIDR ranges to skip all scoring. Blacklist known bad IPs to add +100 score instantly. Optionally whitelist all logged-in WordPress users.</li>
<li>Headless Browser Blocking — Server-side User-Agent check instantly blocks Headless Chrome, Puppeteer, Selenium, and other automated browsers (+100 score). Enabled by default; can be toggled off for E2E testing.</li>
<li>Content Rules — Detect spam patterns in form content: excessive URLs, BBCode link syntax, and WordPress Disallowed Comment Keys matching.</li>
<li>Trusted Proxy Support — Optional mode for Cloudflare and reverse proxy environments with IP range validation.</li>
<li>Lightweight — Three PHP files, no external dependencies, no jQuery.</li>
</ul>

Samurai Honeypot for Forms

Invisible 15-layer anti-spam for Contact Form 7 & WPForms. Score-based 3-Tier Triage silently blocks bots — no CAPTCHA, no user friction.

Erik

gardenboi

Are you unsatisfied with your current antispam solution for Contact Form 7? It might be using an ineffective method to combat the specific type of bot attacks you’re facing. Fortunately, I have a solution for you! 
Antispam for Contact Form 7 is a simple yet highly effective plugin that protects your mailbox from bot flooding. Say goodbye to tedious configurations and captchas, which often lead to reduced conversions and inconvenience for genuine users. Our plugin utilizes a combination of on-page and off-page bot traps, along with an auto-learning mechanism powered by a statistical “Bayesian” spam filter called B8. 
CF7-AntiSpam seamlessly integrates with <a href="https://wordpress.org/plugins/flamingo/" rel="ugc">Flamingo</a> and enhances its functionality. When both plugins are installed, Flamingo gains additional controls, and an extra dashboard widget is enabled.
<h3>SETUP</h3>
Basic – Install and go! No configuration, keys, or registrations are required to activate the antispam protection. In this case, some protections, such as fingerprinting, language checks, and honeypots, will be enabled. 
Advanced – For CF7A to properly analyze the email content using its dictionary, it needs to parse the input message field of your form. To notify the antispam to check this field, you’ll need to add a “marker” to each contact form on your website. Simply add ‘flamingo_message: “[your-message]”‘ in the additional settings panel of each contact form you want to secure. This process follows the same method used with Flamingo. While this step may seem tedious, it is required for advanced text statistical analysis. Without it, the B8 filter cannot be enabled. 
GeoIP – (Optional) If you need to restrict which countries or languages can email you, you can enable this functionality. To enable GeoIP, you’ll need to agree to the GeoLite2 End User License Agreement and sign up for GeoLite2 Downloadable Databases. This will provide you with the required key to download the database. For detailed instructions, please refer to the dedicated section in the cf7-antispam plugin settings.
<h3>Antispam Available Tests</h3>
✅ Browser Fingerprinting 
✅ Language checks (Geo-ip, http headers and browser) 
✅ Honeypot 
️🆕 Honeyform* 
✅ Domain Name System Blackhole List (aka DNSBL) 
✅ blocklists (with automatic ban after N failed attempts, user defined ip exclusion list) 
✅ Hidden fields with encrypted unique hash 
✅ Time elapsed (with min/max values) 
✅ Prohibited words in message/email and user agent 
✅ B8 statistical “Bayesian” spam filter 
✅ Identity protection 
✅ Webmail protection
<h3>Extends Flamingo and turns it into a spam manager!</h3>
With this plugin, you can now review emails and train B8 to identify spam and legitimate messages. This feature proves useful, especially during the initial stages when some spam emails may slip through. 
Already using Flamingo? Even better! Just remember to add ‘flamingo_message: “[your-message]”‘ to the advanced settings (similar to other Flamingo labels) before activating the plugin. Alternatively, you can explore the advanced options and select “rebuild dictionary.” 
Upon activating CF7A, all previously collected emails will be parsed, and B8 will learn and develop its vocabulary. This pre-trained algorithm gives you a head start. How cool is that? 
Additional Notes: 
– A new column has been added to the right side of the Flamingo inbound page, displaying the level of spaminess for each email. 
– If you unban an email on the Flamingo “inbound” page, the corresponding IP will be removed from the blocklist. However, marking an email as spam will not blocklist the IP again. 
– Before activating this plugin, please make sure to mark all spam emails as spam in the Flamingo inbound section. This auto-training process will help the B8 algorithm. 
– If you receive a spam message, please avoid deleting it from the “ham” section. Instead, place it in the spam section to teach B8 how to differentiate between spam and legitimate messages.
<h3>B8 statistical “Bayesian” Filter</h3>
Originally created by <a href="https://en.wikipedia.org/wiki/Gary_Robinson" rel="nofollow ugc">Gary Robinson</a> <a href="https://www.linuxjournal.com/article/6467" rel="nofollow ugc">b8 is a statistical “Bayesian”</a> spam filter implemented in PHP. 
The B8 filter is a foundational example of Machine Learning (ML) for text classification, representing an early, yet powerful, statistical approach in Natural Language Processing (NLP). This approach precedes feature-weighting methods like TF-IDF, which in turn paved the way for modern deep learning architectures, such as Transformers and GPT. 
The filter tells you whether a text is spam or not, using statistical text analysis. What it does is: you give b8 a text and it returns a value between 0 and 1, saying it’s ham when it’s near 0 and saying it’s spam when it’s near 1. See <a href="https://nasauber.de/opensource/b8/readme.html#how-does-it-work" rel="nofollow ugc">How does it work?</a> for details about this. 
To be able to distinguish spam and ham (non-spam), b8 first has to learn some spam and some ham texts. If it makes mistakes when classifying unknown texts or the result is not distinct enough, b8 can be told what the text actually is, getting better with each learned text. 
This takes place on your own server without relying on third-party services. 
More info: <a href="https://nasauber.de/opensource/b8/" rel="nofollow ugc">nasauber.de</a>
<h3>Identity protection</h3>
To fully protect the forms, it may be necessary to enable a couple of additional controls, because bots use the public data of the website to spam on it. 
– The first is user related and denies those who are not logged in the possibility of asking (sensitive) information about the user via wp-api and the protection for the xmlrpc exploit wordpress. 
– The second one is the WordPress protection that will obfuscate sensitive WordPress and server data, adding some headers in order to enhance security against xss and so on. 
Will be hidden the WordPress and WooCommerce version (wp_generator, woo_version), pingback (X-Pingback), server (nginx|apache|…) and php version (X-Powered-By), enabled xss protection headers (X-XSS-Protection), removes rest api link from header (but it will only continue to work if the link is not made public).
<h3>Mailbox Protection (Multiple Send)</h3>
Enhance email security by enabling the “Multiple Send” feature, which prevents consecutive email submissions to the user’s mailbox. This measure is effective in thwarting automated spam attempts and ensures a secure communication environment.
<h3>Security & Privacy: A Local-First Approach</h3>
AntiSpam for Contact Form 7 is built with your security and privacy as the top priority. Unlike many modern anti-spam solutions that rely on external cloud services or third-party subscriptions, our plugin is designed to run entirely on your own WordPress installation.
<ul>
<li>100% Local Processing: All anti-spam logic, checks, and data processing are performed directly on your server. No data is ever sent to, or stored by, any external third-party service (including ours).</li>
<li>Not a Software as a Service (SaaS): This plugin is a standalone, self-contained software solution, not an interface to a paid or subscription-based external service. Once installed, it works autonomously.</li>
<li>Enhanced Security: Since there is no central server or external API endpoint to communicate with, your website is immune to potential risks associated with centralized services, such as Single Point of Failure or data breach risks.</li>
</ul>
You retain complete control and ownership over the security of your Contact Form 7 submissions.
<h3>Privacy Notices</h3>
AntiSpam for Contact Form 7 only processes the IP but doesn’t store any personal data directly from the user input. However, it creates a dictionary of spam and ham (non-spam) words in the WordPress database. 
This dictionary is built from words found in the submitted messages, meaning it may contain words that were part of the user’s e-mail message or personal data. This data is “degenerated,” which means the words might be normalized or altered before being stored. 
The sole purpose of this word collecting is to build a dictionary used for local, decentralized spam detection.
<h3>Support</h3>
Community support: via the <a href="https://wordpress.org/support/plugin/contact-form-7-antispam/" rel="ugc">support forums</a> on wordpress.org 
Bug reporting (preferred): file an issue on <a href="https://github.com/erikyo/contact-form-7-antispam" rel="nofollow ugc">GitHub</a>
<h4>Contribute</h4>
We love your input! We want to make contributing to this project as easy and transparent as possible, whether it’s:
<ul>
<li>Reporting a bug</li>
<li>Testing the plugin with different user agent and report fingerprinting failures</li>
<li>Discussing the current state, features, improvements</li>
<li>Submitting a fix or a new feature</li>
</ul>
We use GitHub to host code, to track issues and feature requests, as well as accept pull requests. 
By contributing, you agree that your contributions will be licensed under its GPLv2 License.
My goal is to create an antispam that protects cf7 definitively without relying on external services. And free for everyone. 
if you want to help me, <a href="https://github.com/erikyo/contact-form-7-antispam" rel="nofollow ugc">GitHub</a> is the right place 😉
<h3>copyright</h3>
AntiSpam for Contact Form 7, Copyright 2021 Codekraft Studio 
AntiSpam for Contact Form 7 is distributed under the terms of the GNU GPL
This program is free software: you can redistribute it and/or modify 
it under the terms of the GNU General Public License as published by 
the Free Software Foundation, either version 3 of the License, or 
(at your option) any later version.
This program is distributed in the hope that it will be useful, 
but WITHOUT ANY WARRANTY; without even the implied warranty of 
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. 
See the LICENSE file for more details.
<h4>Resources</h4>
<ul>
<li>Contact Form 7 and Flamingo © 2021 Takayuki Miyoshi,<a href="https://it.wordpress.org/plugins/contact-form-7/" rel="nofollow ugc">LGPLv3 or later</a></li>
<li>B8 https://nasauber.de/opensource/b8/, © 2021 Tobias Leupold, <a href="https://gitlab.com/l3u/b8/-/tree/ab26daa6b293e6aa059d24ce7cf77af6c8b9b052/LICENSES" rel="nofollow ugc">LGPLv3 or later</a></li>
<li>GeoLite2 <a href="https://www.maxmind.com/en/geolite2/eula" rel="nofollow ugc">license</a></li>
<li>GeoIP2 PHP API <a href="https://github.com/maxmind/GeoIP2-php" rel="nofollow ugc">GeoIP2-php</a></li>
<li>chart.js https://www.chartjs.org/, © 2021 Chart.js <a href="https://github.com/chartjs/Chart.js/graphs/contributors" rel="nofollow ugc">contributors</a>, <a href="https://github.com/chartjs/Chart.js/blob/master/LICENSE.md" rel="nofollow ugc">MIT</a></li>
<li>Sudden Shower in the Summer, Public domain, Wikimedia Commons https://commons.wikimedia.org/wiki/File:Sudden_Shower_in_the_Summer_(5759500422).jpg</li>
</ul>
<h3>Contibutions</h3>
Mirek Długosz – <a href="https://github.com/erikyo/cf7-antispam/pull/30" rel="nofollow ugc">#30</a> fixes a crash that occurred when analysing flamingo metadata 
MeliEve – <a href="https://github.com/wp-blocks/cf7-antispam/pull/42" rel="nofollow ugc">#42</a> Fix “internal_server_error” when message is empty 
MeliEve – <a href="https://github.com/wp-blocks/cf7-antispam/pull/61" rel="nofollow ugc">#61</a> Handle deferrer script loading 
Zodiac1978 – <a href="https://github.com/wp-blocks/cf7-antispam/pull/67" rel="nofollow ugc">#67</a> Remove warning for unsafe email configuration w/o protection 
JohnHooks – <a href="https://github.com/wp-blocks/cf7-antispam/pull/61" rel="nofollow ugc">#66</a> Readme + plugin env 
sdellenb – <a href="https://github.com/wp-blocks/cf7-antispam/pull/163" rel="nofollow ugc">#66</a> Fix $reason parameter for calling cf7a_ban_by_ip
<h3>Special thanks</h3>
This project is tested with BrowserStack. <a href="https://www.browserstack.com/" rel="nofollow ugc">Browserstack</a>
<h3>MaxMind GeoIP2</h3>
This plugin on demand can enable GeoLite2 created by MaxMind, available from <a href="https://www.maxmind.com" rel="nofollow ugc">https://www.maxmind.com</a> 
While enabled you may have to mention it in the privacy policy of your site, depending on the law regulating privacy in your state! 
* GeoIP2 databases <a href="https://www.maxmind.com/en/accounts/current/geoip/downloads" rel="nofollow ugc">GeoLite2 Country</a>
<h3>DNSBL servers privacy policies</h3>
<ul>
<li>dnsbl-1.uceprotect.net <a href="http://www.uceprotect.net/en/index.php?m=13&s=0" rel="nofollow ugc">www.uceprotect.net license</a></li>
<li>dnsbl-2.uceprotect.net <a href="http://www.uceprotect.net/en/index.php?m=13&s=0" rel="nofollow ugc">www.uceprotect.net license</a></li>
<li>dnsbl-3.uceprotect.net <a href="http://www.uceprotect.net/en/index.php?m=13&s=0" rel="nofollow ugc">www.uceprotect.net license</a></li>
<li>dnsbl.sorbs.net <a href="http://www.sorbs.net/information/faq/" rel="nofollow ugc">sorbs.net license</a></li>
<li>zen.spamhaus.org <a href="https://www.spamhaus.org/organization/dnsblusage/" rel="nofollow ugc">spamhaus.org license</a></li>
<li>bl.spamcop.net <a href="https://www.spamcop.net/fom-serve/cache/297.html" rel="nofollow ugc">spamcop.net license</a></li>
<li>b.barracudacentral.org <a href="https://www.barracuda.com/company/legal/trust-center/data-privacy/privacy-policy" rel="nofollow ugc">barracudacentral.org privacy-policy</a></li>
<li>dnsbl.dronebl.org <a href="https://dronebl.org/docs/faq" rel="nofollow ugc">dronebl.org</a></li>
<li>all.spamrats.com <a href="https://spamrats.com/tos.php" rel="nofollow ugc">spamrats.com tos</a></li>
<li>bl.ipv6.spameatingmonkey.net <a href="https://spameatingmonkey.com/faq" rel="nofollow ugc">spameatingmonkey.net</a></li>
</ul>
<h3>Inspirations, links</h3>
<ul>
<li>Nikolai Tschacher <a href="https://incolumitas.com/pages/BotOrNot/" rel="nofollow ugc">incolumitas.com</a></li>
<li>Antoine Vastel <a href="https://github.com/antoinevastel/fpscanner" rel="nofollow ugc">fp-scanner</a>/<a href="https://github.com/antoinevastel/fp-collect" rel="nofollow ugc">fp-collect</a></li>
<li>Niespodd <a href="https://github.com/niespodd/browser-fingerprinting" rel="nofollow ugc">niespodd</a></li>
<li>Thomas Breuss <a href="https://gist.github.com/tbreuss/74da96ff5f976ce770e6628badbd7dfc" rel="nofollow ugc">tbreuss</a></li>
<li>Domain Name System-based blackhole list <a href="https://en.wikipedia.org/wiki/Domain_Name_System-based_blackhole_list" rel="nofollow ugc">wiki</a></li>
<li>dnsbl list <a href="https://en.wikipedia.org/wiki/Comparison_of_DNS_blacklists" rel="nofollow ugc">wiki</a></li>
</ul>

AntiSpam for Contact Form 7

A trustworthy antispam plugin for Contact Form 7. Wave goodbye to spam and keep your inbox clean!

Samurai Honeypot for Forms vs AntiSpam for Contact Form 7