Atomic Edge analysis of CVE-2026-57646:
This vulnerability is an Insecure Direct Object Reference (IDOR) in the Majestic Support plugin for WordPress, affecting versions up to and including 1.1.7. The flaw allows authenticated attackers with Subscriber-level access to perform unauthorized actions by exploiting a missing access control check in the AI-powered reply feature. The CVSS score is 4.3.
Root Cause: The vulnerability resides in the file /modules/ticket/model.php, specifically in the function handling the AI-powered reply feature. In the vulnerable version, the code at line 3209 lacked any permission verification before processing requests. The patch introduces a strict access control block that checks if the user is an admin, an agent with the required AI permission, or denies access otherwise. The diff shows the insertion of a new security check at lines 3212-3236 that explicitly blocks standard users (including Subscribers) from accessing the AI reply functionality.
Exploitation: An authenticated attacker with Subscriber-level privileges can send a crafted AJAX request to the WordPress admin-ajax.php endpoint with the action parameter set to the vulnerable handler (likely ‘majestic_support_ai_reply’ or similar) and a ticketId parameter. The server processes the request without verifying that the user has the necessary permissions, allowing the attacker to invoke the AI reply feature on any ticket. The precise AJAX action name was not provided in the diff, but it is triggered by the code path that calls the vulnerable model method.
Patch Analysis: The patch adds a three-tier access control check before the AI reply logic executes. Admin users retain full access. Agents must have the ‘Use AI Powered Reply Feature’ permission granted via the plugin’s user permissions system. All other users, including Subscribers, are blocked with an error message. This directly addresses the missing authorization by verifying the user’s role and permissions before processing the request.
Impact: Successful exploitation allows an authenticated Subscriber to use the AI-powered reply feature on tickets, potentially viewing and generating replies for tickets they should not have access to. This leads to unauthorized data exposure and could facilitate further social engineering or privilege escalation within the support system.







