AI Engine CVE-2026-111041

Requires an authenticated account with the Editor role or above on a target WordPress site running AI Engine before 3.7.9; the plugin's AI media management actions fail to verify that the acting user owns or is permitted to edit the targeted object. This is the core precondition - no special plugin configuration is named in the description, so the flaw applies to default installs, but the PR:H barrier is the key limiting factor: subscriber/contributor/author-level and unauthenticated users cannot trigger it. Realistic risk is therefore concentrated in multi-author, editorial, agency, or multisite deployments where Editor-role users exist and are intentionally not trusted to modify site-wide content; single-administrator sites face little practical exposure.

This is a medium-severity (CVSS 4.9) broken access control / missing authorization issue (CWE-862), and the signals consistently point to a moderate, non-urgent priority rather than a critical one. The CVSS vector CVSS:3.1/AV:N/AC:L/PR:H/UI:N/S:U/C:N/I:H/A:N shows network reach and low complexity, but crucially requires PR:H - a pre-authenticated account with Editor role or above - and impacts integrity only (C:N/A:N). CISA's SSVC framework reinforces a low-urgency read: Exploitation=none (no observed active exploitation), Automatable=no, and Technical Impact=partial. A public proof-of-concept exists (WPScan), so exploit code is available, but this is NOT the same as active exploitation - the CVE is not in CISA KEV. No EPSS score was provided in the input, so probabilistic exploitation likelihood cannot be factored in and should be verified separately. The main real-world risk is confined to multi-author, editorial, or multisite WordPress deployments where Editor-role users exist and are deliberately NOT trusted with site-wide content control; in single-admin or small sites the PR:H prerequisite makes this largely a privilege-boundary/insider concern rather than a broadly exploitable remote threat.

On a multi-author WordPress site running AI Engine before 3.7.9, a user who has been granted the Editor role invokes an AI media management action against a post ID they do not own or are not permitted to edit. Because the plugin never verifies ownership or edit capability on the targeted object, the Editor can modify or change the type of arbitrary posts across the whole site, tampering with content outside their assigned scope. A public proof-of-concept for this flow has been published by WPScan.