The isValidDuration() regex at objects/video.php:918 uses /^[0-9]{1,2}:[0-9]{1,2}:[0-9]{1,2}/ without a $ end anchor, allowing arbitrary HTML/JavaScript to be appended after a valid duration prefix. The crafted duration is stored in the database and rendered without HTML escaping via echo Video::getCleanDuration() on trending pages, playlist pages, and video gallery thumbnails, resulting in stored cross-site scripting.
Input entry point: objects/aVideoEncoderReceiveImage.json.php:208
// Line 203-211
if (!empty($_REQUEST['duration'])) {
$video->setDuration($_REQUEST['duration']);
}
Insufficient validation: objects/video.php:918
static function isValidDuration($duration) {
// ...
return preg_match('/^[0-9]{1,2}:[0-9]{1,2}:[0-9]{1,2}/', $duration);
// Missing $ anchor here -----------------------------------^
}
The regex matches 00:00:01 at the start of the string but ignores everything after it. A payload like 00:00:01</time><img src=x onerror=alert(1)><time> passes validation.
No sanitization in output function: objects/video.php:3463-3480
public static function getCleanDuration($duration = "") {
$durationParts = explode(".", $duration);
$duration = $durationParts[0];
$durationParts = explode(':', $duration);
if (count($durationParts) == 1) {
return '0:00:' . static::addZero($durationParts[0]);
} elseif (count($durationParts) == 2) {
return '0:' . static::addZero($durationParts[0]) . ':' . static::addZero($durationParts[1]);
}
return $duration; // Returns full string unmodified for 3+ colon parts
}
With the payload 00:00:01</time><img src=x onerror=alert(1)><time>, exploding by : yields 3+ parts, so the full unsanitized string is returned.
Unescaped output sinks:
view/trending.php:72:
<time class="duration"><?php echo Video::getCleanDuration($value['duration']); ?></time>
view/include/playlist.php:159:
<time class="duration"><?php echo Video::getCleanDuration(@$value['duration']); ?></time>
objects/video.php:7200 (gallery thumbnail generation):
$img .= "<time class=\"duration\"...>" . $duration . "</time>";
No Content-Security-Policy headers are set. The application uses raw PHP templates with no auto-escaping framework.
Authenticate as a user with upload permission and obtain a video_id_hash for a video (visible in encoder API responses or via the upload flow).
Send the malicious duration:
curl -X POST "https://target/objects/aVideoEncoderReceiveImage.json.php" \
-d "videos_id=VIDEO_ID" \
-d "video_id_hash=HASH" \
-d 'duration=00:00:01</time><img src=x onerror=alert(document.cookie)><time>'
The isValidDuration() regex matches the 00:00:01 prefix and allows the full string to be stored.
Visit the trending page (/view/trending.php) or any playlist containing the poisoned video. The injected HTML breaks out of the <time> tag and the onerror handler executes JavaScript in the victim's browser context.
The attack requires only upload-level permissions (low privilege) and impacts all users who view any page rendering the poisoned video's duration (high blast radius).
Fix 1 — Anchor the regex (objects/video.php:918):
- return preg_match('/^[0-9]{1,2}:[0-9]{1,2}:[0-9]{1,2}/', $duration);
+ return preg_match('/^[0-9]{1,2}:[0-9]{1,2}:[0-9]{1,2}(\.[0-9]+)?$/', $duration);
Fix 2 — HTML-escape all duration output (defense in depth):
In view/trending.php:72:
- <time class="duration"><?php echo Video::getCleanDuration($value['duration']); ?></time>
+ <time class="duration"><?php echo htmlspecialchars(Video::getCleanDuration($value['duration']), ENT_QUOTES, 'UTF-8'); ?></time>
In view/include/playlist.php:159:
- <time class="duration"><?php echo Video::getCleanDuration(@$value['duration']); ?></time>
+ <time class="duration"><?php echo htmlspecialchars(Video::getCleanDuration(@$value['duration']), ENT_QUOTES, 'UTF-8'); ?></time>
In objects/video.php:7200:
- $img .= "<time class=\"duration\"...>" . $duration . "</time>";
+ $img .= "<time class=\"duration\"...>" . htmlspecialchars($duration, ENT_QUOTES, 'UTF-8') . "</time>";
Both fixes should be applied: the regex fix prevents storage of invalid data, and the output escaping provides defense in depth against any other code path that might store unvalidated durations.
{
"nvd_published_at": "2026-04-21T23:16:21Z",
"cwe_ids": [
"CWE-79"
],
"github_reviewed": true,
"github_reviewed_at": "2026-04-14T23:22:21Z",
"severity": "MODERATE"
}