Your RSA-2048 keys break in 2030. Find every one of them before attackers do.
🐘
🐘 Packagist
Not in CISA KEV
MEDIUM severity

CVE-2026-43876 wwbn/avideo

MEDIUMFix: WWBN/AVideo@078c434

CVE-2026-43876 is a medium-severity (CVSS 6.4) Cross-site Scripting (XSS) vulnerability in wwbn/avideo. No vendor fix is recorded yet; mitigation options are listed below.

WWBN AVideo: HTML Injection in notifySubscribers.json.php Enables Platform-Branded Phishing Emails to Channel Subscribers

Also known asGHSA-g9cm-rxp7-6gv5
Published
May 11, 2026
Updated
Aug 12, 2026
Affected
1 pkg
Patched
See advisory
Exploits
None indexed
Exploitation data as of Sep 21, 2026 · OSV.dev, NVD, FIRST.org (EPSS)

Exploitation Status

Proof-of-concept exploit code exists

  • CISA’s SSVC triage found public proof-of-concept exploit code for this CVE, though no confirmed active exploitation.

Exploitation and automatability from CISA’s SSVC triage for CVE-2026-43876.

EPSS Exploitation Probability

via FIRST.org ↗
0.2%probability of exploitation in next 30 days
Lower Risk0.00%
Lower risk than most CVEs5th percentile — riskier than 5% of all scored CVEsHighest risk

EPSS (Exploit Prediction Scoring System) is a daily probability model maintained by FIRST.org. It estimates the likelihood a CVE will be exploited in production environments within the next 30 days, derived from real-world threat intelligence signals.

How urgent is this, really

CVE-2026-43876 plotted by exploitation likelihood (EPSS) against impact (CVSS). The shaded corner — EPSS 50%+ and CVSS 7.0+ — is where this CVE doesn't sit, though severity or exploitability alone can still warrant action.

Where this sits among everything scored

Of 377,636 CVEs with a current EPSS score, this one falls in the < 10% band (highlighted). Real counts from FIRST.org, not a sample — log-scaled since the landscape is heavily right-skewed.

Real-World Exposure

1 pkg affected
🐘wwbn/avideo

Real-time download stats are indexed for npm and PyPI packages. This vulnerability affects Packagist packages — download data is not available via public APIs for these ecosystems.

Description

Summary

objects/notifySubscribers.json.php takes the raw message POST parameter and passes it into sendSiteEmail(), which substitutes it directly into an HTML email template (via str_replace on the {message} placeholder) and renders it with PHPMailer::msgHTML(). There is no HTML sanitization, character escaping, or output encoding on the attacker-controlled message between $_POST['message'] and the rendered email. Any authenticated user with upload permission can therefore broadcast arbitrary HTML — phishing links, tracking pixels, CSS/UI spoofing — to every subscriber on their channel (up to 10,000 recipients per invocation). The email is sent From: the platform's configured contact address and wrapped in the site's official logo and title, so attacker-supplied HTML arrives with the appearance of an official platform communication.

Details

File: objects/notifySubscribers.json.php

10: if (!User::canUpload()) {
11:     forbiddenPage('You can not notify');
12: }
13: forbidIfIsUntrustedRequest('notifySubscribers');
14: $user_id = User::getId();
15: // if admin bring all subscribers
16: if (User::isAdmin()) {
17:     $user_id = '';
18: }
19:
20: require_once 'subscribe.php';
21: setRowCount(10000);
...
23: $Subscribes = Subscribe::getAllSubscribes($user_id);
...
34: $subject = 'Message From Site ' . $config->getWebSiteTitle();
35: $message = $_POST['message'];
36:
37: $resp = sendSiteEmail($to, $subject, $message);

Controls present at the entry point:

  • User::canUpload() — gates access to any account that can upload (a baseline authenticated uploader role; in typical AVideo configurations where authCanUploadVideos is enabled, this is any logged-in user with a verified email).
  • forbidIfIsUntrustedRequest('notifySubscribers') — in objects/functionsSecurity.php:138-165, this delegates to isUntrustedRequest() which only validates same-origin via requestComesFromSameDomainAsMyAVideo() (objects/functionsAVideo.php:199-206), i.e. a Referer/Origin header check. It is not a CSRF token. An attacker acting on their own authenticated browser session trivially satisfies the Referer check.

There is no CAPTCHA, no rate limit, no per-recipient quota, and no unsubscribe link. setRowCount(10000) allows up to 10,000 subscriber rows to be pulled and mailed in a single request. For admin callers (User::isAdmin()$user_id = ''), Subscribe::getAllSubscribes('') returns the entire subscriber set for the platform rather than the caller's channel.

File: objects/functionsMail.php

 59: function sendSiteEmail($to, $subject, $message, $fromEmail = '', $fromName = '')
 60: {
 ...
 78:     $subject = UTF8encode($subject);
 79:     $message = UTF8encode($message);            // UTF-8 normalization, no HTML handling
 80:     $message = createEmailMessageFromTemplate($message);
 ...
119:             $mail = new \PHPMailer\PHPMailer\PHPMailer();
120:             setSiteSendMessage($mail);
...
125:             $systemEmail = $config->getContactEmail();
126:             $systemName  = $config->getWebSiteTitle();
...
136:             $mail->setFrom($systemEmail, !empty($fromName) ? $fromName : $systemName);
...
143:             $mail->msgHTML($message);           // renders as HTML
...
162:             $resp = $mail->send();
266: function createEmailMessageFromTemplate($message)
267: {
268:     if (preg_match("/html>/i", $message)) {
269:         return $message;                       // attacker-supplied full-HTML is returned verbatim
270:     }
...
274:     $text = file_get_contents("{$global['systemRootPath']}view/include/emailTemplate.html");
...
279:     $words   = [$logo, $message, $siteTitle];
280:     $replace = ['{logo}', '{message}', '{siteTitle}'];
281:
282:     return str_replace($replace, $words, $text);   // raw substitution into HTML template
283: }

Execution flow from attacker input to sink:

  1. $_POST['message']objects/notifySubscribers.json.php:35 (raw, no validation).
  2. sendSiteEmail($to, $subject, $message) at objects/notifySubscribers.json.php:37.
  3. UTF8encode($message) at objects/functionsMail.php:79 (encoding only; does not strip or escape HTML).
  4. createEmailMessageFromTemplate($message) at objects/functionsMail.php:80str_replace('{message}', $message, $text) at objects/functionsMail.php:282, substituting attacker HTML directly into the {message} placeholder in view/include/emailTemplate.html.
  5. $mail->msgHTML($message) at objects/functionsMail.php:143. PHPMailer renders the combined template (containing the attacker's unsanitized HTML) as an HTML email.
  6. The From: header is $config->getContactEmail() / $config->getWebSiteTitle() (objects/functionsMail.php:125-136). The template contains the platform's logo via getURL($config->getLogo()). The result is an attacker-controlled HTML body delivered from the platform's trusted sender address, officially branded.

Note that the preg_match("/html>/i", $message) at line 268 actively helps the attacker: any payload containing <html> short-circuits template substitution and is sent as-is, allowing the attacker to control the full email body including DOCTYPE, head, and body.

PoC

  1. Obtain an account with upload permission (on AVideo installations where authCanUploadVideos is enabled, any registered and email-verified user qualifies). An admin account broadens the recipient set to the entire platform rather than just the attacker's own subscribers.
  2. Ensure the attacker's channel has at least one subscriber (via Subscribe::getAllSubscribes($user_id)), or use an admin account to target all platform subscribers.
  3. Submit the request. The Referer header must match the platform origin to pass forbidIfIsUntrustedRequest (trivial when running from the attacker's own authenticated browser):
curl -b 'PHPSESSID=<uploader_session>' -X POST \
  -H 'Referer: https://target.example/' \
  'https://target.example/objects/notifySubscribers.json.php' \
  --data-urlencode 'message=<h1 style="color:#c00">Action Required: Verify Your Account</h1>
<p>Dear Subscriber,</p>
<p>We detected unusual activity on your account. Please
<a href="https://attacker.example/phish">click here to verify your identity within 24 hours</a>
or your account will be suspended.</p>
<p>Thank you,<br>The Support Team</p>
<img src="https://attacker.example/track.png" width="1" height="1">'
  1. Expected response:
{"error": false, "msg": ""}
  1. Every subscriber in the target set receives an HTML email:

    • From: <[email protected]> (the platform's configured contact email — not the attacker's address).
    • Subject: Message From Site <SiteTitle> - <SiteTitle> (built at objects/notifySubscribers.json.php:34 + objects/functionsMail.php:138-141).
    • Body: the view/include/emailTemplate.html template with the platform's real logo substituted at {logo} and the attacker's unsanitized HTML substituted at {message}, including the phishing anchor and tracking pixel.
  2. Delivery is batched via partition($to, $size) at objects/functionsMail.php:114-118 over up to 10,000 subscribers in a single request. There is no rate limit, CAPTCHA, confirmation step, or unsubscribe header.

Impact

  • Any authenticated uploader can weaponize the platform's own email infrastructure and brand (contact email, logo, site title) to deliver phishing content to their channel subscribers.
  • Because the From: address is the platform's canonical contact email and the template wraps the attacker content in the official logo and site title, recipients have no visible indication that the content originated from an uploader rather than the operator. Recipients who have previously received legitimate notifications from the same address are especially likely to trust the email.
  • Phishing payloads can include credential-stealing links mimicking password reset / account verification flows, tracking pixels that enumerate subscriber IPs and mail-client metadata, and CSS-based UI spoofing over the template.
  • An admin account (User::isAdmin()$user_id = '') expands the blast radius to every subscriber record on the platform, not just the attacker's own subscribers.
  • Up to 10,000 recipients per request with no rate limiting, CAPTCHA, or unsubscribe link, so a compromised or malicious uploader can sustain large phishing campaigns at minimal cost, while the sending IP reputation is borne by the platform operator.
  • A stolen uploader session (e.g., via an unrelated XSS or token leak) is sufficient to mount the attack; no additional credentials or admin access are required.

Recommended Fix

Sanitize or encode $_POST['message'] before it reaches PHPMailer::msgHTML(). Options, in order of preference:

  1. Reject HTML outright and force plain text. In objects/notifySubscribers.json.php:

    $message = $_POST['message'] ?? '';
    // Strip all HTML; allow only newlines / plain text.
    $message = strip_tags($message);
    $message = nl2br(htmlspecialchars($message, ENT_QUOTES | ENT_HTML5, 'UTF-8'));
    
  2. Or allow a very restricted subset using a proven HTML sanitizer (e.g. HTMLPurifier with a minimal whitelist: p, br, strong, em, a[href|title], ul, ol, li), and forbid <script>, <style>, inline event handlers, <img>, <iframe>, data:/javascript: URIs, and framework-style template tokens.

  3. Additionally remove the preg_match("/html>/i", $message) short-circuit at objects/functionsMail.php:268-270, which lets a caller replace the entire email body by including a <html> tag. The template should always be applied.

  4. Defense in depth:

    • Require a real anti-CSRF token on this endpoint (e.g. validateCSRF() with a per-session token in a header or POST field), and drop the Referer-only forbidIfIsUntrustedRequest as the sole protection.
    • Require User::isAdmin() to notify subscribers from accounts not scoped to a channel; for non-admin uploaders, make the From display name clearly attribute the message to the uploader ("{uploaderName} via {siteTitle} <contact@site>" already works for non-system senders in objects/functionsMail.php:130-134 — apply the same attribution to subscriber notifications).
    • Enforce per-account and per-IP rate limits on notifySubscribers.json.php (e.g. one broadcast per account per N hours, max M recipients per day).
    • Include a List-Unsubscribe header and a per-recipient unsubscribe link.
    • Add a preview + confirmation step before dispatch.

Affected Packages

1 total
EcosystemPackageVulnerable rangeFix
🐘Packagistwwbn/avideoall versionsNo fix

Detection & mitigation playbook

Open-source dependency
  1. Detect

    Scan your dependency tree (package-lock.json, pnpm-lock.yaml, requirements.txt, go.sum, etc.) for wwbn/avideo, including transitive dependencies — a direct dependency you never call can still pull in a vulnerable version.

  2. Remediation status

    No patched version of wwbn/avideo has shipped for CVE-2026-43876 yet. Where your build allows, override or pin the dependency away from the vulnerable range, and apply any maintainer-recommended mitigation.

  3. Mitigate without a patch

    If you can't upgrade right away: gate or disable the affected feature, validate untrusted input at the boundary, and avoid passing attacker-controlled data into the vulnerable path. O3's runtime protection blocks exploitation in production as an interim safeguard until the upgrade lands.

  4. How O3 protects you

    O3 Security's impact-aware SCA analyses which vulnerable code paths your application actually calls, so a match like CVE-2026-43876 can be triaged on real exposure rather than presence alone.

Tailored to CVE-2026-43876. Runtime protection reduces exposure until a permanent patch is applied and verified — it complements patching, it doesn't replace it.

Frequently Asked Questions

## Summary `objects/notifySubscribers.json.php` takes the raw `message` POST parameter and passes it into `sendSiteEmail()`, which substitutes it directly into an HTML email template (via `str_replace` on the `{message}` placeholder) and renders it with `PHPMailer::msgHTML()`. There is no HTML sanitization, character escaping, or output encoding on the attacker-controlled `message` between `$_POST['message']` and the rendered email. Any authenticated user with upload permission can therefore broadcast arbitrary HTML — phishing links, tracking pixels, CSS/UI spoofing — to every subscriber on t
O3 Security · Impact-Aware SCA

Is CVE-2026-43876 in your dependencies?

O3 Security finds CVE-2026-43876 across Packagist dependencies, including transitive ones, and its impact-aware SCA ranks findings by whether your code actually calls the vulnerable path.

CVE-2026-43876: wwbn/avideo (Medium 6.4) | O3 Security