# CVE-2026-75798: AI Engine WordPress plugin authorization bypass in editor assistant

Source: https://junglewise.ai/threats/cve-2026-75798-ai-engine-wordpress-plugin-authorization-bypass-in-editor  
Publisher: Junglewise Threat Intelligence, https://junglewise.ai/threats  
Published: 2026-08-26. Updated: 2026-09-25  
License: CC BY 4.0 with a link back

CVE-2026-75798 · Severity: medium · CVSS 5.3 · Published 2026-08-26

The AI Engine WordPress plugin allows unauthenticated visitors to execute arbitrary AI queries against the site owner's configured AI provider (e.g., OpenAI) account by bypassing authorization checks on an admin-only feature. An attacker can obtain a guest token and use it to run AI requests with custom instructions and model selection, causing unauthorized charges to the site owner's provider account and potentially exposing sensitive AI capabilities to malicious use.

## Quick answers

**Is CVE-2026-75798 exploited in the wild?** No. Junglewise Threat Intelligence has no report of CVE-2026-75798 being exploited in the wild.

**How severe is CVE-2026-75798?** Medium severity, with a CVSS score of 5.3, published 26 August 2026.

**Which products and versions are affected?** AI Engine Contributors AI Engine 3.4.0 to 3.7.1.

**Is there a fix for CVE-2026-75798?** A patch was published on 2026-08-24: Fixed in version 3.7.2 Junglewise finds every place you run it, drafts the upgrade and applies it once you approve.

**How is CVE-2026-75798 attacked?** It is attacked over the network, in a software library.

**Am I affected by CVE-2026-75798?** Check whether anything you run is named in the advisory. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.

## Technical details

The vulnerability is an authorization bypass (CWE-862) in the AI Engine WordPress plugin versions before 3.7.2. The plugin's editor assistant feature accepts REST API requests but fails to verify that the caller has administrator privileges, instead relying solely on a REST nonce token that the plugin freely distributes to anonymous visitors via a public session route. An attacker first calls POST /wp-json/mwai/v1/start_session to obtain a valid restNonce, then uses it to execute arbitrary AI queries by POSTing to /wp-json/mwai-ui/v1/editor/submit with attacker-controlled parameters (system prompt, model, environment ID). The request executes against the site owner's configured provider account and API key, incurring unauthorized charges. The fix, released in version 3.7.2, implements a permission callback (can_use_assistant()) that returns HTTP 401 for unauthenticated callers and strips server parameters before chatbot resolution.

## Affected products

- AI Engine Contributors AI Engine 3.4.0 to 3.7.1

## Timeline

- 2026-08-24: disclosed
- 2026-08-24: patched: Fixed in version 3.7.2

## References

- [https://wpscan.com/vulnerability/a77168df-61b8-40df-95f5-e19cf6540ce5/](https://wpscan.com/vulnerability/a77168df-61b8-40df-95f5-e19cf6540ce5/)