AI Engine EUVD-2026-43626

Exploitation requires an authenticated WordPress account with editor-level (or higher) privileges on a site running AI Engine before 3.5.5, and the attacker must reach the plugin feature that downloads a file and writes it using a caller-supplied filename - the traversal lives in that filename parameter. No user interaction and no non-default configuration is indicated by the CVSS vector (UI:N, AC:L). Limiting factors: the PR:L requirement means fully unauthenticated attackers cannot trigger it directly, the write destination is bounded by the web server user's permissions, and escalation to code execution depends on being able to write into an executable/webroot path.

The CVSS 3.1 vector (AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:N, score 8.1) describes low-complexity, network-reachable exploitation requiring only editor-level authentication and no user interaction, with high confidentiality and integrity impact but no availability impact - consistent with an arbitrary file write that can escalate to code execution. Publicly available exploit code raises practical risk and lowers the barrier to attack, but the vulnerability is NOT in CISA KEV, so there is no confirmation of active in-the-wild exploitation; POC availability and active exploitation are distinct signals and should not be conflated. No EPSS score is provided in the input, so probabilistic exploitation likelihood cannot be quantified here, and no SSVC decision is supplied. The main mitigating factor is the PR:L requirement: an attacker must already hold or steal an editor account, which limits fully anonymous mass exploitation but is well within reach of compromised credentials, malicious insiders, or privilege escalation on multi-author sites.

An attacker who has obtained editor-level credentials (via phishing, credential stuffing, or a compromised contributor) invokes the plugin's download-and-save function with a crafted filename containing traversal sequences, causing a PHP web shell to be written into the WordPress webroot. They then request that file over HTTP to execute arbitrary code and take over the site. Publicly available exploit code lowers the effort required to reproduce this chain.