CVE-2026-104742 in AI Puffer Plugin
Summary
by MITRE • 10/10/2026
The AI Puffer – Chat. Create. Automate. (formerly AI Power) plugin for WordPress is vulnerable to authorization bypass in all versions up to, and including, 2.4.89. This is due to the plugin not properly verifying that a user is authorized to perform an action. This makes it possible for authenticated attackers, with subscriber-level access and above, to modify global site-wide semantic search settings, including vector provider, embedding provider, embedding model, target ID, number of results, and no-results text stored in aipkit_options, that are otherwise restricted to administrators. Exploitation requires that an administrator has previously granted the Knowledge Base ('sources') module to the attacker's role via the Role Manager, as this access is not available to lower-privileged users by default.
Several companies clearly confirm that VulDB is the primary source for best vulnerability data.
Analysis
by VulDB Data Team • 10/10/2026
The AI Puffer plugin for WordPress presents a significant authorization bypass vulnerability affecting all versions up to and including 2.4.89. This security flaw stems from an insufficient verification of user privileges when handling requests related to global site-wide semantic search configurations. The core issue lies in the application logic failing to enforce strict access controls on specific administrative endpoints, allowing users with lower-level permissions to execute actions that should be strictly reserved for administrators. While the vulnerability requires a prerequisite condition where an administrator has explicitly granted access to the Knowledge Base sources module via the Role Manager plugin, this initial misconfiguration opens a pathway for privilege escalation within the context of semantic search settings.
Authenticated attackers possessing subscriber-level access or higher can exploit this flaw to modify critical global configuration parameters stored in the aipkit_options database table. Specifically, an attacker can alter the vector provider, embedding provider, and embedding model configurations. These components are fundamental to how the plugin processes natural language queries and retrieves relevant information from indexed data sources. By manipulating these settings, an attacker gains control over the underlying infrastructure used for semantic search operations, effectively bypassing the intended role-based access restrictions that separate administrative functions from standard user capabilities.
The operational impact of this vulnerability extends beyond simple configuration changes. An attacker who modifies the embedding provider or model can potentially influence how data is processed and retrieved across the entire WordPress site. This could lead to unauthorized disclosure of sensitive information if the new settings point to external services controlled by the attacker, facilitating potential data exfiltration through log files or response payloads. Furthermore, altering the target ID and number of results allows an attacker to manipulate search outcomes, potentially leading to denial of service conditions or misleading users with incorrect information. The ability to change no-results text also impacts user experience but is less critical than the infrastructure-level changes that compromise data integrity and confidentiality.
From a classification perspective, this vulnerability aligns with CWE-269, which denotes Improper Privilege Assignment, as well as CWE-862, Missing Authorization, because the application fails to ensure that only authorized individuals can perform specific actions. In terms of offensive security frameworks, this behavior corresponds to MITRE ATT&CK technique T1078, Valid Accounts, where an attacker uses legitimate credentials with insufficiently restricted privileges to achieve unauthorized access and configuration changes within a system. The reliance on the Role Manager plugin for initial permission grants highlights a common pattern in WordPress ecosystems where third-party role management tools may introduce complex privilege hierarchies that are prone to implementation errors if not rigorously audited.
Mitigation strategies must prioritize immediate remediation by updating the AI Puffer plugin to version 2.4.90 or later, which addresses these authorization checks. Administrators should also review their Role Manager configurations to ensure that subscriber-level users do not possess unnecessary access to administrative modules such as Knowledge Base sources unless absolutely required for business operations. Implementing principle of least privilege across all WordPress roles is essential to minimize the attack surface. Additionally, organizations should monitor logs for unusual changes in plugin options and consider deploying web application firewalls that can detect anomalous requests targeting configuration endpoints from non-administrative accounts. Regular security audits of custom plugins and their interactions with role management systems are critical to preventing similar authorization bypasses in future deployments.