CVE-2024-38791: AI Engine ChatGPT Chatbot SSRF Vulnerability

CVE-2024-38791 Overview

CVE-2024-38791 is a Server-Side Request Forgery (SSRF) vulnerability [CWE-918] in the Jordy Meow AI Engine: ChatGPT Chatbot plugin for WordPress. The flaw affects all versions from unspecified initial releases through 2.4.7. Authenticated attackers with low privileges can coerce the vulnerable WordPress instance to issue arbitrary HTTP requests to attacker-chosen destinations. This enables reconnaissance of internal networks, interaction with cloud metadata services, and access to services that trust the WordPress host.

Critical Impact
An authenticated attacker can abuse the plugin to send server-side HTTP requests to internal or restricted endpoints, exposing sensitive data such as cloud instance metadata credentials.

Affected Products

Meowapps AI Engine: ChatGPT Chatbot plugin for WordPress
All versions up to and including 2.4.7
Deployments where the plugin is installed and accessible to authenticated users

Discovery Timeline

2024-08-01 - CVE-2024-38791 published to NVD
2026-06-17 - Last updated in NVD database

Technical Details for CVE-2024-38791

Vulnerability Analysis

The AI Engine plugin integrates ChatGPT-style chatbot functionality into WordPress and issues outbound HTTP requests to external services. The vulnerability arises because the plugin accepts a URL or resource identifier from an authenticated user and fetches it server-side without validating the destination. An attacker with low-privilege authenticated access can direct these requests to arbitrary hosts.

SSRF in a WordPress plugin context is impactful because the underlying PHP process typically runs inside a cloud tenant or a corporate network. Attackers can reach internal admin interfaces, databases, and cloud metadata endpoints such as 169.254.169.254 that are unreachable from the internet. The confidentiality impact is high while integrity impact is limited to what the fetched endpoints permit.

Root Cause

The root cause is missing or insufficient validation of user-supplied URLs before they are passed to the server-side HTTP client. The plugin does not enforce an allowlist of destinations, does not restrict URL schemes, and does not block requests to private, loopback, or link-local address ranges. This maps directly to the CWE-918 classification for Server-Side Request Forgery.

Attack Vector

Exploitation requires network access to the WordPress site and authenticated credentials at a low privilege level. The attacker submits a crafted request to a plugin endpoint containing a URL that points to an internal resource. The vulnerable server fetches the URL and returns the response body or observable side effects to the attacker. Refer to the Patchstack SSRF Vulnerability Report for additional technical context.

No verified public proof-of-concept code has been published. The vulnerability is described here in prose because no sanitized exploit sample is available from an authoritative source.

Detection Methods for CVE-2024-38791

Indicators of Compromise

Outbound HTTP requests from the WordPress PHP worker to internal IP ranges such as 10.0.0.0/8, 172.16.0.0/12, 192.168.0.0/16, or loopback addresses.
Requests originating from the web server to cloud metadata endpoints like 169.254.169.254.
Unusual authenticated activity against AI Engine plugin REST or AJAX endpoints followed by outbound connections to unexpected hosts.

Detection Strategies

Inspect web server and PHP-FPM logs for AI Engine plugin request handlers correlated with outbound socket activity.
Enable and review WordPress plugin audit logs for low-privilege accounts invoking AI Engine functionality outside expected patterns.
Deploy egress filtering with alerts on any HTTP request from the web tier to RFC1918 space or 169.254.169.254.

Monitoring Recommendations

Forward web server, PHP, and network flow logs to a centralized platform for correlation across the WordPress host and its network path.
Baseline normal outbound destinations from the WordPress instance and alert on deviations.
Monitor cloud provider audit logs for use of instance metadata credentials in unexpected geographies or API patterns.

How to Mitigate CVE-2024-38791

Immediate Actions Required

Upgrade the AI Engine: ChatGPT Chatbot plugin to a version newer than 2.4.7 that addresses the SSRF issue.
Audit WordPress user accounts and remove or downgrade unused low-privilege accounts that could be leveraged to authenticate.
Rotate any cloud instance credentials that may have been exposed via metadata service access from the WordPress host.

Patch Information

The vulnerability affects AI Engine versions through 2.4.7. Administrators should update to the latest release available from the WordPress plugin repository. Consult the Patchstack SSRF Vulnerability Report for the fixed version reference.

Workarounds

Restrict the WordPress web server's egress traffic so it cannot reach RFC1918 ranges, loopback, or 169.254.169.254 metadata endpoints.
On AWS instances, enforce IMDSv2 with hop limit 1 to prevent metadata credential theft via SSRF.
Temporarily disable the AI Engine plugin if an upgrade cannot be applied immediately.

bash

# Enforce IMDSv2 on an EC2 instance hosting WordPress
aws ec2 modify-instance-metadata-options \
  --instance-id i-0123456789abcdef0 \
  --http-tokens required \
  --http-put-response-hop-limit 1 \
  --http-endpoint enabled