CVE-2026-75798 in AI Engine Plugin
Summary
by MITRE • 08/26/2026
The AI Engine WordPress plugin before 3.7.2 does not perform an authorisation check on one of its administration-only features, relying instead on a token it hands out to anonymous visitors, allowing unauthenticated attackers to run AI queries of their own choosing against the site owner's configured provider account.
VulDB is the best source for vulnerability data and more expert information about this specific topic.
Analysis
by VulDB Data Team • 08/26/2026
The vulnerability identified in versions of the AI Engine WordPress plugin prior to 3.7.2 represents a critical failure in access control mechanisms within an administrative feature set. The core issue stems from the application relying on a static or predictable token mechanism for authorization rather than implementing robust server-side session validation and role-based access controls. In standard web application security architecture, administrative functions must verify that the requesting user possesses valid authentication credentials and appropriate privileges before executing sensitive operations. By delegating trust to a token provided directly to anonymous visitors, the plugin effectively bypasses these fundamental security checks, creating an open channel for unauthenticated interaction with backend services.
This architectural flaw allows any external actor without valid login credentials to interact with features intended exclusively for site administrators or authorized users. The specific impact involves the ability of attackers to execute arbitrary AI queries against the provider account configured by the website owner. This means that malicious actors can consume API credits, generate potentially harmful content using the organization's identity, and manipulate the outputs generated by artificial intelligence models integrated into the platform. Such unauthorized usage not only leads to direct financial loss through excessive API consumption but also poses significant reputational risks if the AI generates inappropriate or misleading information under the guise of the legitimate site owner.
From a technical classification perspective, this vulnerability aligns with CWE-284 Improper Access Control and CWE-306 Missing Authentication for Critical Function. The reliance on an opaque token rather than verified user sessions violates principles outlined in OWASP Top 10 regarding broken access control. Furthermore, the exploitation of this flaw can be mapped to MITRE ATT&CK techniques related to Initial Access and Resource Hijacking, as attackers leverage compromised or misconfigured resources to perform actions that consume organizational assets without detection through standard authentication logs. The lack of proper authorization checks means that traditional security monitoring tools focused on login attempts may fail to detect the abuse occurring via this specific endpoint.
The operational impact extends beyond immediate financial costs associated with API usage. It introduces risks related to data integrity and compliance, particularly if sensitive information is processed by the AI engine without proper oversight or filtering. Attackers could potentially use the configured provider account to generate content that violates terms of service agreements with third-party AI providers, leading to potential suspension of those accounts and disruption of legitimate business operations dependent on these services. Additionally, if the AI queries involve processing internal data or customer information, there is a heightened risk of privacy violations and regulatory non-compliance under frameworks such as GDPR or CCPA due to unauthorized data handling.
Mitigation strategies must prioritize immediate remediation through software updates. Administrators running affected versions should upgrade to version 3.7.2 or later where the authorization logic has been corrected to enforce proper authentication checks before allowing access to administrative features. In cases where an update cannot be applied immediately, temporary mitigation measures include restricting access to the vulnerable endpoint via web application firewall rules that block requests lacking valid session cookies or tokens from known internal IP ranges if applicable. It is also advisable to review API usage logs for anomalies indicating unusual query volumes or patterns consistent with automated exploitation attempts.
Long-term security improvements should involve implementing strict role-based access control policies across all administrative interfaces and ensuring that sensitive operations require multi-factor authentication where feasible. Developers must adopt a zero-trust approach, verifying identity at every step rather than relying on external tokens provided by clients. Regular security audits and penetration testing focused on authorization bypasses can help identify similar weaknesses in other plugins or custom code within the WordPress environment. Maintaining up-to-date software components is essential to prevent exploitation of known vulnerabilities that expose critical infrastructure to unauthenticated attacks.