Description

The AI Agent by SiteGround plugin for WordPress is vulnerable to authorization bypass in all versions up to, and including, 1.2.7. This is due to the plugin not properly verifying that a user is authorized to perform an action. This makes it possible for unauthenticated attackers to upload images to the WordPress media library, bypassing the upload_files capability restriction that Contributors are normally subject to, as authenticated attackers with Contributor-level access or above can satisfy the endpoint's nonce and permission checks. The sg_ai_studio_gutenberg_nonce required by the endpoint is emitted to any user with block editor access — including Contributors — making the absent upload_files check the sole barrier to exploitation.

Executive Summary

Technical Details

Mitigations

References

Additional Metadata

{
  "score": "5.3",
  "pubDate": "2026-08-20T06:16:58.230Z",
  "pubdate": "2026-08-20T06:16:58.230Z",
  "executiveSummary": "The AI Agent by SiteGround plugin for WordPress is affected by an authorization bypass vulnerability impacting all versions up to, and including, 1.2.7. This security flaw stems from an improper capability verification mechanism within the plugin's endpoint handlers, failing to enforce strict access controls for sensitive operations. Consequently, unauthenticated and low-privileged authenticated attackers—specifically those with Contributor-level access—can leverage this vulnerability to upload arbitrary images directly to the WordPress media library. This behavior completely bypasses the standard upload_files capability restriction normally enforced by the WordPress core for Contributor-level accounts. The risk implications include unauthorized media uploads and potential resource exhaustion or abuse of storage. Exploitation requires the acquisition of a specific nonces emitted to users with block editor access, allowing attackers to satisfy existing token and permission validation checks while evading intended capability boundaries.",
  "technicalDetails": "The root cause of the vulnerability resides in the insufficient validation of user capabilities during the execution of specific AJAX or REST API endpoints within the AI Agent by SiteGround plugin. While the endpoint validates the sg_ai_studio_gutenberg_nonce required for request integrity, it fundamentally fails to verify whether the requesting user possesses the necessary upload_files capability.\nThe vulnerable component involves the backend request handling logic associated with the plugin's AI Studio Gutenberg integration across all versions up to, and including, 1.2.7. Because the sg_ai_studio_gutenberg_nonce is dynamically emitted to any user who possesses block editor access—which natively includes authenticated roles such as Contributors—the nonce itself acts as an insufficient barrier to exploitation.\nThe attack flow proceeds as follows: First, an attacker with low-privileged Contributor access (or potentially an unauthenticated actor depending on exact endpoint exposure, though the summary notes contributors and nonce emission) obtains the sg_ai_studio_gutenberg_nonce from the block editor interface. Second, the attacker constructs a malicious HTTP request targeting the vulnerable endpoint, embedding the acquired nonce to successfully satisfy the plugin's weak permission checks. Third, the plugin processes the request without validating the upload_files capability. Finally, the payload executes, allowing the unauthorized upload of image files into the WordPress media library, bypassing the security controls designed to restrict file management operations to higher-privileged roles like Authors, Editors, and Administrators.\nThe network exposure includes any WordPress instance running the affected plugin where the vulnerable endpoint is reachable over HTTP/HTTPS. The post-exploitation impact centers primarily on unauthorized media library manipulation, circumvention of content management policy boundaries, and the potential ingestion of malicious or unintended image payloads into the system."
}