GPTranslate CVE-2026-103309

Exploitation requires the plugin's 'server-side translations' feature to be enabled - this is the explicit precondition named in the description and is not necessarily the default, so sites using client-side translation are not affected. The attacker needs no authentication (PR:N) and can submit the malicious translation remotely because the plugin fails to restrict who may store translations. Key limiting factors: a victim must subsequently load the poisoned translated page (UI:R), the vulnerable version must be below 2.34.14, and the attack complexity is High (AC:H), indicating conditions outside the attacker's direct control (e.g., reliance on the translation-storage path and specific configuration).

This is a stored XSS flaw (CWE-79) in the GPTranslate WordPress plugin. The provided CVSS vector (CVSS:3.1/AV:N/AC:H/PR:N/UI:R/S:U/C:H/I:H/A:H, base 7.5) signals an unauthenticated, network-reachable issue (PR:N) that nonetheless requires high attack complexity (AC:H) and victim interaction (UI:R). The CIA impacts are rated High across the board with scope Unchanged, which appears inflated for a reflected-in-page stored XSS - typical client-side XSS yields Low impacts with a scope change (code runs in the victim's browser, a different security authority), so the real severity is likely moderate rather than the 7.5 shown. No EPSS score, no CISA KEV listing, and no public POC are present in the input, so there is no public exploit identified at time of analysis; combined with AC:H and the non-default 'server-side translations' requirement, this reads as a genuine but conditional, lower-urgency issue rather than a mass-exploited priority. The headline limiting factor is the AC:H / non-default-configuration dependency, which materially lowers real-world risk despite the unauthenticated vector.

On a site running GPTranslate before 2.34.14 with server-side translations enabled, an unauthenticated attacker submits a translation containing a malicious script payload, which the plugin stores without sanitization. When a visitor or administrator later loads the affected translated page, the unescaped payload executes in their browser, enabling session/cookie theft or actions in their authenticated context. No public exploit code was identified in the provided data.