EUVD-2026-84280
GHSA-h8jc-c8c9-624r
Attack ChainAIDerived
Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain
-
AccessSubmit entity-encoded payload via public translation input
-
Deliveryhtml_entity_decode() bypasses kses on original column
-
ExploitPayload stored in translation memory dictionary
-
ExecutionAdmin opens Translation Memory Suggestion Panel
-
Persistv-html renders suggestion.original unescaped
-
ImpactScript executes in administrator session
Vulnerability Assessment
| Weakness |
CWE
OWASP Top 10:2025
A05:2025 - Injection
ASVS 5.0
V1 - Encoding and Sanitization Output encoding, context-aware escaping
MITRE ATT&CK
2 techniques across 2 tactics: Initial Access, Execution The technique list with ATT&CK links and the kill-chain view is part of the Basic plan and up. |
| Exploitation Conditions | Exploitation requires the TranslatePress translation/rendering pipeline to be active (default for the plugin's core multilingual function) so that html_entity_decode() processes attacker-supplied front-end content before persistence into the translation memory 'original' column, which is intentionally exempt from kses sanitization. No authentication is needed to inject (PR:N), but the payload only executes when a privileged administrator subsequently opens the Translation Memory Suggestion Panel, where the Vue component renders suggestion.original via v-html - that admin view is the necessary triggering condition and the primary limiting factor. Affected: all versions ≤ 3.3.5; fixed in 3.3.6. |
| Impact | The vendor CVSS 3.1 vector (AV:N/AC:L/PR:N/UI:N/S:C/C:L/I:L/A:N = 7.2 High) describes an unauthenticated, network-reachable, low-complexity stored XSS whose scope-change (S:C) reflects that an unauthenticated attacker's payload executes later inside a privileged administrator's browser session. The most serious signal is PR:N combined with an admin-context sink: an anonymous visitor seeds the Translation Memory 'original' column, and the script runs when a WordPress admin opens the Translation Memory Suggestion Panel, so a successful payload can act with the admin's authenticated privileges (account/settings tampering, potential plugin/theme editor abuse toward RCE on WordPress). Tempering that: no EPSS score, no CISA KEV entry, and no public POC are present in the supplied data, so there is no evidence of active or automated exploitation - this is a high-CVSS issue whose real-world urgency currently rests on the discovery report alone. UI:N is slightly generous because exploitation depends on an admin actually viewing the suggestion panel (a form of passive interaction), but the unauthenticated injection vector and admin-session impact keep this a genuine priority to patch rather than a paper-tiger. Note the C:L/I:L rating is conservative for XSS landing in an admin session, where privileged actions could escalate impact beyond 'Low'. |
| Exploit ScenarioDerived | An unauthenticated attacker submits translatable content on the public front end containing an entity-encoded HTML/JavaScript payload; the front-end rendering pipeline runs html_entity_decode() before persistence and the 'original' column is deliberately exempted from kses filtering, so the executable markup is stored verbatim. When a site administrator later opens the Translation Memory Suggestion Panel, the Vue component renders suggestion.original via v-html and the script executes in the admin's authenticated session. No public exploit code is identified in the provided data, but the flaw was reported by Wordfence with specific source-line references. |
Get an e-mail when this CVE is exploited, patched or re-scored. Free account.
Alert meAnalysisAIv1 – Sep 22, 2026
Stored cross-site scripting in the WordPress plugin TranslatePress - Translate Multilingual sites with AI Translation, affecting all versions up to and including 3.3.5, lets unauthenticated attackers plant executable HTML into the translation memory 'original' column through the front-end rendering pipeline, which runs html_entity_decode() on entity-encoded input before the value is persisted from a column that is deliberately exempt from WordPress kses filtering. The injected markup is later rebound into the DOM by the Vue-based Translation Memory Suggestion Panel component (v-html on suggestion.original) and executes in the browser of a privileged administrator who opens that panel, which is why the vendor applies a scope change (S:C) - an unauthenticated payload executes inside an authenticated admin session. No public exploit code and no CISA KEV confirmed active exploitation were identified at time of analysis; the vendor rates this 7.2 (CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:L/I:L/A:N), while independent assessment conservatively applies UI:R because administrator interaction with the suggestion panel is the required trigger. The vendor-released fix is TranslatePress 3.3.6.
RemediationAI
Update the plugin to TranslatePress 3.3.6, the vendor-released fixed version, using the WordPress plugin updater or by downloading the 3.3.6 tag; the patch is documented in the Trac changeset at https://plugins.trac.wordpress.org/changeset?reponame=&new=3697206%40translatepress-multilingual%2Ftags%2F3.3.6&old=3686891%40translatepress-multilingual%2Ftags%2F3.3.5 and the patch that corrects the Vue rendering is at https://plugins.trac.wordpress.org/changeset/3697206/translatepress-multilingual/trunk/assets/src/js/components/translation-memory.vue, with context in the Wordfence advisory https://www.wordfence.com/threat-intel/vulnerabilities/id/717e8479-921b-4f18-8fbe-32e0d8a37590?source=cve. If immediate upgrade is not possible, the most effective compensating control is to stop rendering the vulnerable surface: instruct administrators not to open the Translation Memory Suggestion Panel until patched, since execution requires that admin-side view - the trade-off is loss of translation-memory review workflow and slower bulk translation editing. Because injection is unauthenticated and reaches the database during ordinary front-end translation activity, also review and purge the translation memory table for rows in the 'original' column containing HTML tags or script markers, and consider temporarily disabling the plugin's translation memory feature (which reduces automation and may require manual translation re-entry). Blocking unauthenticated access to the translatepress endpoints or adding a WAF rule that rejects entity-encoded script/HTML patterns in translated page requests limits new injections but will not remove payloads already stored, and cannot protect the admin panel if the plugin remains enabled. Additional hardening: restrict the ability to reach the translation-memory admin panel to a minimal set of trusted accounts and re-check the admin UI in a browser with script execution sandboxed until the upgrade is complete.
Affected VersionsENISA
Affected Packages
Source: OSV (Open Source Vulnerabilities)
Affects your stack
Affects 1 package in euvd
Checking this CVE against your own stacks and servers is part of the Basic plan and up.
Technical ContextAI · show details
The root cause is CWE-79 (Improper Neutralization of Input During Web Page Generation), realized through a double-processing flaw unique to TranslatePress's translation memory subsystem. On the front end, the plugin's rendering pipeline (class-translation-render.php, around lines 996 and 1024 in 3.3.5) passes attacker-controllable page content through html_entity_decode(), which converts entity-encoded characters such as <script> back into literal markup before the string is stored in the translation memory table's 'original' column. That column is intentionally excluded from WordPress kses sanitization in class-translation-memory.php (lines 47 and 60), so no save-time filter neutralizes the payload, and WordPress's standard defense of sanitizing on save never applies. The stored value is subsequently delivered to the admin UI, where the translation-memory.vue component (line 12) renders suggestion.original with Vue's v-html directive, treating the string as trusted HTML rather than text; the browser then parses and executes the embedded script in the administrator's authenticated context. The CPE cpe:2.3:a:cozmoslabs:translatepress_-_translate_multilingual_sites_with_ai_translation confirms a single affected product from vendor Cozmoslabs, with a wildcard version field indicating that every release prior to the 3.3.6 changeset (Trac changeset 3697206) is in scope. The distinction from an ordinary reflected XSS is the two-stage architecture: unauthenticated injection at the front-end decode step, and delayed execution in a higher-privileged session at the admin-side render step.
Know when Wordpress Plugin gets its next vulnerability
- Watch products and get an email when something new is published
- Check this CVE against your own stack and servers
- Exploit and PoC references, threat intelligence from 22 sources
No credit card · 7-day full trial
Severity by source
AV:N/AC:L/PR:N/UI:N/S:C/C:L/I:L/A:N
Unauthenticated remote injection (PR:N, AV:N, AC:L); UI:R because payload fires only when an admin views the suggestion panel; S:C and C:L/I:L for script execution in the admin's session.
Primary rating from Vendor (Wordfence).
CVSS VectorVendor: Wordfence
DescriptionCVE.org
The TranslatePress - Translate Multilingual sites with AI Translation plugin for WordPress is vulnerable to Stored Cross-Site Scripting via Translation Memory Suggestion Panel (v-html on suggestion.original) in all versions up to, and including, 3.3.5 due to insufficient input sanitization and output escaping. This makes it possible for unauthenticated attackers to inject arbitrary web scripts in pages that will execute whenever a user accesses an injected page. Unauthenticated attackers can seed the translation dictionary's original column with executable HTML because the front-end rendering pipeline decodes entity-encoded payloads via html_entity_decode() before persistence, and the original column is deliberately exempt from kses filtering - meaning no save-time sanitizer neutralizes the stored payload before it is later rendered in an administrator's session.
Articles & Coverage 1
Executive ViewAI
TranslatePress (all versions up to and including 3.3.5) carries a stored cross-site scripting flaw that allows an unauthenticated visitor to plant executable markup into a site's translation memory, which a logged-in administrator then unknowingly runs when opening the Translation Memory Suggestion Panel. The vendor has not released a patch at time of analysis, so any WordPress site using the affected plugin for multilingual publishing remains exposed, with the main risk being attacker control of a privileged admin session rather than discrete data loss. Leadership should care because the affected population includes the highest-privilege web accounts, and a single successful injection can pivot into content tampering, credential abuse, or further compromise of brand-facing properties; no public exploit was identified at time of analysis, but the conditions for abuse are fully present.
No vendor-released patch identified at time of analysis. Interim mitigations: deactivate or remove TranslatePress on non-essential sites; restrict and monitor administrator accounts and sessions; re-enable kses filtering on the translation memory storage path; apply CSP and WAF rules that block HTML/script payloads into translation memory fields and admin panel rendering; audit and sanitize existing translation memory rows.
Threat Intelligence Overview
NVD
CVE.org
EPSS
Vulnrichment
How this decision was reached
The engine reads each signal below and applies a fixed list of rules from top to bottom. The first rule that matches sets the recommended action – like a statement, every line is shown.
What every action level means
SSVC AssessmentENISA
Priority Score
Attack Technique
Cross-Site Scripting occurs when an application accepts untrusted data and sends it to a web browser without proper validation or encoding.
More in Wordpress Plugin
View allThe Backup Migration plugin for WordPress is vulnerable to Remote Code Execution in all versions up to, and including, 1
The Hash Form - Drag & Drop Form Builder plugin for WordPress is vulnerable to arbitrary file uploads due to missing fil
Authentication bypass in the LiteSpeed Cache WordPress plugin (versions prior to 6.5.0.1) allows unauthenticated remote
The Simple File List plugin for WordPress through version 4.2.2 contains an unauthenticated remote code execution vulner
The AI Engine WordPress plugin through version 3.1.3 exposes Bearer Token values through the /mcp/v1/ REST API endpoint
The SureTriggers WordPress plugin through version 1.0.82 contains a privilege escalation vulnerability that allows unaut
The Ninja Forms plugin before 2.9.42.1 for WordPress allows remote attackers to conduct PHP object injection attacks via
The Business Directory Plugin - Easy Listing Directories for WordPress plugin for WordPress is vulnerable to time-based
SQL injection in the NotificationX WordPress plugin (versions up to and including 2.8.2) allows unauthenticated remote a
The POST SMTP Mailer - Email log, Delivery Failure Notifications and Best Mail SMTP for WordPress plugin for WordPress i
The MasterStudy LMS WordPress Plugin - for Online Courses and Education plugin for WordPress is vulnerable to union base
The Email Subscribers by Icegram Express plugin for WordPress is vulnerable to SQL Injection via the ‘hash’ parameter in
Exploit-DB