AI Engine <= 3.8.0 - Unauthenticated Stored Cross-Site Scripting via 'model_' Parameter → PHP Error-Log Injection → Advisor Indirect Prompt Injection
Strategic Overview
- Status
- Patched in 3.8.1
- Affected Plugin
- AI Engine – The Chatbot, AI Framework & MCP for WordPress
- Affected Version
<= 3.8.0- CVSS
- 7.2High
- Weakness type
- CWE-79 · Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting')
- CVE
CVE-2026-96561
At a glance
CVE-2026-96561 is a high-severity Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting') vulnerability in the AI Engine WordPress plugin, affecting versions <= 3.8.0. It carries a CVSS score of 7.2 (reachable over the network; low attack complexity). Exploitation requires no authentication. The issue is fixed in version 3.8.1; sites on affected versions should update now. Disclosed September 2026, reported by whale120.
Vulnerability Overview
The AI Engine – The Chatbot, AI Framework & MCP for WordPress plugin for WordPress is vulnerable to Stored Cross-Site Scripting in versions up to, and including, 3.8.0 This is due to a chain of missing input neutralization and output escaping across the /mwai-ui/v1/chats/submit REST endpoint, the PHP error-log parser (MeowKit_MWAI_Helpers::php_error_logs), the Advisor task (Meow_MWAI_Modules_Advisor::run_advisor), and the Advisor dashboard widget (advisor_metabox): the server-parameter denylist in chat_submit strips only exact key names such as 'model' while convert_keys() later canonicalizes 'model_' back to 'model', allowing an unauthenticated caller to place an attacker-controlled string (including CR/LF) into $query->model; final_checks() throws an Exception whose message embeds that raw string, and the non-streaming, non-admin catch branch writes it to the PHP error log unmodified — creating a forged log line that the plugin's own parser subsequently returns as recent PHP-error content; run_advisor() then appends that content verbatim to the AI prompt (indirect prompt injection — CWE-1427), the returned JSON is stored in the mwai_advisor_data option with no schema validation or HTML sanitization, and advisor_metabox() concatenates the resulting 'title' and 'description' values directly into the WordPress dashboard widget without esc_html(), wp_kses(), or equivalent escaping. This makes it possible for unauthenticated attackers to inject arbitrary web scripts in pages that will execute whenever an administrator accesses the WordPress dashboard.
Technical Analysis
The vector marks this flaw as remotely reachable over the network, with low attack complexity — no special timing or configuration is needed, and no privileges on the target site, and no interaction from a victim user.
CWE-79: Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting')
AI Engine <= 3.8.0 carries this weakness at model, and reaching it takes no account at all. Cross-site scripting happens when input from a request is written into a page without being escaped for the context it lands in, so the browser parses attacker-supplied text as markup or script.
Injected script runs with the privileges of whoever views the affected page, which is how these flaws turn into administrator session theft, silent account creation or persistent backdoors in page content. For AI Engine the fix is 3.8.1: builds <= 3.8.0 are affected, anything from 3.8.1 onward is not.
Remediation
Update to version 3.8.1, or a newer patched version
How does WordSec protect against this?
Requests carrying this pattern arrive without a session, so nothing upstream of the firewall gets a chance to reject them: WordSec's web application firewall inspects request payloads before WordPress loads them. None of that substitutes for the fix: AI Engine 3.8.1 closes this, and updating the plugin is the step that ends it.
- Firewall
- Login Security
- Alerts
External References
Related records
Same weakness class
Other vulnerabilities in AI Engine – The Chatbot, AI Framework & MCP for WordPress
- 9.8CVE-2025-11749: AI Engine Info Exposure to Privilege Escalation
CVE-2025-11749 - 9.8CVE-2023-51409: AI Engine: ChatGPT Chatbot Arbitrary File Upload
CVE-2023-51409 - 9.1CVE-2024-34440: AI Engine: ChatGPT Chatbot Arbitrary File Upload
CVE-2024-34440 - 8.8CVE-2026-15988: AI Engine <= 3.6.5 CSRF to Privilege Escalation
CVE-2026-15988 - 8.8CVE-2026-8719: AI Engine 3.4.9 Privilege Escalation
CVE-2026-8719 - 8.8CVE-2025-7847: AI Engine 2.9.3 - 2.9.4 Arbitrary File Upload
CVE-2025-7847 - 8.8CVE-2025-5071: AI Engine 2.8.0 - 2.8.3 Insufficient Authorization
CVE-2025-5071 - 8.0CVE-2025-6238: AI Engine 2.8.4 Insecure OAuth Implementation
CVE-2025-6238
Vulnerability data © Defiant, Inc., provided under the Wordfence Intelligence T&C