Critical Flaw in AI Engine Plugin Puts 100K Websites at Risk of Remote Code Execution

Published:

spot_img

Significant Security Flaw Discovered in AI Engine Plugin for WordPress

A recently unveiled security vulnerability in the AI Engine plugin for WordPress has raised alarms among site administrators. Affected versions, specifically 2.9.3 and 2.9.4, have a flaw that leaves over 100,000 websites at risk. This security issue, categorized as an arbitrary file upload vulnerability, permits authenticated users, even those with minimal access like subscriber-level accounts, to upload harmful files. The potential consequences of this could include remote code execution (RCE) and complete site takeover.

Understanding the Vulnerability: CVE-2025-7847

The vulnerability, identified as CVE-2025-7847, was disclosed to Wordfence on July 18, 2025. This report came from a security researcher known as ISMAILSHADOW, who reported the issue through the Wordfence Bug Bounty Program. Notably, this vulnerability was introduced a day before it was reported, on July 17, leading to a significant bounty of $1,170 for the discoverer.

Technical Breakdown of the Flaw

The root of this vulnerability lies within the rest_simpleFileUpload() function of the AI Engine plugin. This function failed to implement robust file type validation. When the "Public API" feature is enabled—although it’s off by default—any authenticated user can access the plugin’s REST API endpoint at /mwai/v1/simpleFileUpload. The lack of Bearer Token authentication allows any user to upload files without restrictions, potentially including malicious PHP scripts.

The issue escalates in the upload_file() function within the Meow_MWAI_Modules_Files class. Instead of validating the file’s MIME type or extension, the function indiscriminately uses PHP’s native copy() function for uploaded files. Consequently, attackers can bypass file restrictions and execute their scripts on the server, which is a critical threat stemming from such vulnerabilities.

Conditions for Successful Exploitation

Notably, this vulnerability does not impact all AI Engine users; specific conditions must be met for exploitation:

  • The "Public API" feature must be enabled in the plugin’s settings.
  • There should be no Bearer Token or custom authentication method in place.
  • The user attempting exploitation needs to be authenticated, with even a subscriber-level role being sufficient.

When these criteria align, the REST endpoint becomes an open door for low-level users to execute harmful code directly on the server.

Addressing the Issue: Patches and Updates

In response to this threatening vulnerability, plugin developer Jordy Meow quickly acted. On July 22, 2025, version 2.9.5 was released, incorporating essential fixes to address the security risk. The updates introduced rigorous validation protocols using WordPress’s built-in wp_check_filetype() function in both the simpleFileUpload() and upload_file() methods. This enhancement ensures that only approved file types are accepted, effectively closing the security loophole.

Timeline of Events

  • July 18, 2025: Vulnerability reported to Wordfence.
  • July 18, 2025: Disclosure validated and communicated to the developer through the Wordfence Vulnerability Management Portal.
  • July 21, 2025: A firewall rule was activated for Wordfence Premium, Care, and Response users.
  • July 22, 2025: Updated version 2.9.5 of AI Engine released.
  • August 20, 2025: Plans for protection to be extended to Wordfence Free users.

Wordfence’s prompt actions played a crucial role in mitigating the risk of exploitation. However, due to the critical nature of CVE-2025-7847, all users of the AI Engine plugin are urged to upgrade to version 2.9.5 or above without delay, especially if their site employs the Public API feature.

Importance of Security Measures

The emergence of the CVE-2025-7847 vulnerability serves as a critical reminder about the necessity of proper input validation and access controls, particularly for plugins with public API capabilities. Even minor oversights can lead to severe security threats, including remote code execution. Given the widespread utilization of AI Engine, it’s vital for users to secure their REST API endpoints, ensuring robust authentication is in place to protect their sites from potential attacks.

spot_img

Related articles

Recent articles

Origin Energy Data Breach 2026: Unauthorized Access Exposes PII of 900,000 Customers

On July 28, 2026, Origin Energy confirmed a significant data breach impacting approximately 900,000 current and former customers. This incident involved unauthorized access and...

Mirage Kitten Unveils NightLedger Backdoor and WebSocket Tunnelers for Cyber-Espionage in Middle East and Africa

Recent research has unveiled a new set of malware tools attributed to the advanced persistent threat (APT) group known as Mirage Kitten, which is...

Bank of Baroda Reports Cybersecurity Incident Following Alleged Data Theft Claims

Bank of Baroda, one of India's largest state-owned banks, has reported a cybersecurity incident following claims from a threat actor regarding the theft and...

Fairlife resumes US production after ransomware attack, data breach confirmed

USA – The Coca-Cola Company has announced that its dairy subsidiary Fairlife has resumed most production across its four US facilities following a ransomware...