CVE-2026-13731: WPBot unauthenticated stored XSS

CISO Take

WPBot, an AI chatbot plugin used on WordPress sites for live support and lead generation, fails to sanitize or escape the 'conversation' parameter, letting anyone inject persistent JavaScript into chat session data without authenticating. The AJAX nonce meant to gate that save request is broadcast in plaintext on every public page via wp_localize_script, so the one control that would have taken some effort to bypass is handed to attackers for free — pushing exploitation to script-kiddie difficulty despite the CVSS score of 7.2 not fully reflecting the ease of abuse. There's no CISA KEV listing, no public PoC, and no Nuclei template yet, and EPSS sits at 0.24% in absolute terms, though that still ranks in the 85th percentile of all scored CVEs, so this isn't negligible — it looks like a target of opportunity rather than something under active mass exploitation today. The real risk is who views the payload: an admin or support agent opening the WPBot chat session report will execute the injected script inside their authenticated session, opening the door to cookie/session theft and admin account takeover. Patch to the fixed version as soon as it's released, and until then restrict or disable public access to the WPBot widget and review stored conversations in the plugin's chat-sessions tables for injected script or event-handler content.

Sources: NVD EPSS wordfence.com plugins.trac.wordpress.org

What is the risk?

High practical risk despite the moderate CVSS (7.2) and low absolute EPSS (0.24%, 85th percentile). Attack complexity is effectively trivial: no authentication, no user interaction, and the nonce that should act as a barrier is publicly exposed on every frontend page load, so any anonymous visitor can forge a valid save request. The vulnerability is confined to a single WordPress plugin (not a widely-depended-upon AI framework or library), which limits blast radius relative to a supply-chain-style CVE, but any site running WPBot up to 8.4.9 with the chatbot enabled is exposed with zero prerequisites. Impact is scoped to stored XSS (confidentiality/integrity low, no availability impact per the CVSS vector), but stored XSS that executes in an administrator's session is a well-established path to full site takeover.

How does the attack unfold?

Reconnaissance

Attacker loads any public page of a WordPress site running WPBot and extracts the AJAX nonce, which is broadcast in cleartext to all visitors via wp_localize_script.

AML.T0049

Initial injection

Using the harvested nonce, the attacker sends an unauthenticated AJAX request with a malicious script embedded in the 'conversation' parameter, which is stored without sanitization or output escaping.

AML.T0049

Trigger

A site administrator or support agent opens the WPBot chat session report page to review conversations, and the stored payload executes in their authenticated browser session.

Impact

The executed script steals the admin's session cookie or performs privileged actions on their behalf, enabling account takeover and full site compromise.

Reconnaissance

Attacker loads any public page of a WordPress site running WPBot and extracts the AJAX nonce, which is broadcast in cleartext to all visitors via wp_localize_script.

AML.T0049

Initial injection

Using the harvested nonce, the attacker sends an unauthenticated AJAX request with a malicious script embedded in the 'conversation' parameter, which is stored without sanitization or output escaping.

AML.T0049

Trigger

A site administrator or support agent opens the WPBot chat session report page to review conversations, and the stored payload executes in their authenticated browser session.

Impact

The executed script steals the admin's session cookie or performs privileged actions on their behalf, enabling account takeover and full site compromise.

What systems are affected?

Package	Ecosystem	Vulnerable Range	Patched
WPBot	pip	—	No patch

Do you use WPBot? You're affected.

How severe is it?

CVSS 3.1

7.2 / 10

EPSS

0.7%

chance of exploitation in 30 days

Higher than 49% of all CVEs

Source: EPSS v3 — FIRST.org

Exploitation Status

Exploit Possible

Exploitation: LOW

Sophistication

Trivial

Exploitation Confidence

low

✓ VulnCheck KEV (exploitation reported — broader/earlier than CISA) — Jul 2026

○ Nuclei detection template available

Composite signal derived from CISA KEV, VulnCheck KEV, CISA SSVC, EPSS, Metasploit, Exploit-DB, trickest/cve, Nuclei templates, and inthewild.io exploitation reports.

What is the attack surface?

AV Network

AC Low

PR None

UI None

S Changed

C Low

I Low

A None

What should I do?

1 step

1) Update WPBot to the version that fixes this issue as soon as the vendor releases a patched build beyond 8.4.9 — check the WordPress.org plugin changelog and Wordfence advisory for the fixed version number. 2) Until patched, consider disabling the WPBot chat widget on public pages or placing it behind a WAF rule that blocks script tags/event handlers in AJAX POST bodies to the WPBot save-conversation endpoint. 3) Audit existing stored chat sessions in the plugin's database tables (wpbot-chat-sessions.php handlers) for <script>, onerror=, javascript:, or other injected markup and purge/sanitize any matches. 4) Restrict who can view chat session reports to the minimum necessary admin roles, since the exploit only fires when a privileged user opens the report view. 5) Rotate admin session cookies/credentials if any anomalous chat content is found, since a successful hit indicates possible session compromise. 6) Detection: monitor AJAX requests to the WPBot conversation-save endpoint for anomalous payload patterns from unauthenticated sources, and flag any nonce values observed being reused across many distinct IPs.

What does CISA's SSVC say?

Decision Track

Exploitation none

Automatable Yes

Technical Impact partial

Source: CISA Vulnrichment (SSVC v2.0). Decision based on the CISA Coordinator decision tree.

How is it classified?

Code Execution Auth Bypass Plugin Agent AML.T0049 - Exploit Public-Facing Application

Which compliance frameworks are affected?

This CVE is relevant to:

ISO 42001

A.6.2.6 - Third-party and customer relationships / AI system security

NIST AI RMF

MANAGE-4.1 - Post-deployment monitoring for AI system risks

OWASP LLM Top 10

LLM05 - Improper Output Handling

Frequently Asked Questions

What is CVE-2026-13731?

WPBot, an AI chatbot plugin used on WordPress sites for live support and lead generation, fails to sanitize or escape the 'conversation' parameter, letting anyone inject persistent JavaScript into chat session data without authenticating. The AJAX nonce meant to gate that save request is broadcast in plaintext on every public page via wp_localize_script, so the one control that would have taken some effort to bypass is handed to attackers for free — pushing exploitation to script-kiddie difficulty despite the CVSS score of 7.2 not fully reflecting the ease of abuse. There's no CISA KEV listing, no public PoC, and no Nuclei template yet, and EPSS sits at 0.24% in absolute terms, though that still ranks in the 85th percentile of all scored CVEs, so this isn't negligible — it looks like a target of opportunity rather than something under active mass exploitation today. The real risk is who views the payload: an admin or support agent opening the WPBot chat session report will execute the injected script inside their authenticated session, opening the door to cookie/session theft and admin account takeover. Patch to the fixed version as soon as it's released, and until then restrict or disable public access to the WPBot widget and review stored conversations in the plugin's chat-sessions tables for injected script or event-handler content.

Is CVE-2026-13731 actively exploited?

No confirmed active exploitation of CVE-2026-13731 has been reported, but organizations should still patch proactively.

How to fix CVE-2026-13731?

1) Update WPBot to the version that fixes this issue as soon as the vendor releases a patched build beyond 8.4.9 — check the WordPress.org plugin changelog and Wordfence advisory for the fixed version number. 2) Until patched, consider disabling the WPBot chat widget on public pages or placing it behind a WAF rule that blocks script tags/event handlers in AJAX POST bodies to the WPBot save-conversation endpoint. 3) Audit existing stored chat sessions in the plugin's database tables (`wpbot-chat-sessions.php` handlers) for `<script>`, `onerror=`, `javascript:`, or other injected markup and purge/sanitize any matches. 4) Restrict who can view chat session reports to the minimum necessary admin roles, since the exploit only fires when a privileged user opens the report view. 5) Rotate admin session cookies/credentials if any anomalous chat content is found, since a successful hit indicates possible session compromise. 6) Detection: monitor AJAX requests to the WPBot conversation-save endpoint for anomalous payload patterns from unauthenticated sources, and flag any nonce values observed being reused across many distinct IPs.

What systems are affected by CVE-2026-13731?

This vulnerability affects the following AI/ML architecture patterns: chatbot / conversational support widgets, plugin-based AI integrations, lead generation forms embedded in AI chat UI.

What is the CVSS score for CVE-2026-13731?

CVE-2026-13731 has a CVSS v3.1 base score of 7.2 (HIGH). The EPSS exploitation probability is 0.66%.

What is the AI security impact?

Affected AI Architectures

chatbot / conversational support widgetsplugin-based AI integrationslead generation forms embedded in AI chat UI

MITRE ATLAS Techniques

AML.T0049 Exploit Public-Facing Application

Compliance Controls Affected

ISO 42001: A.6.2.6

NIST AI RMF: MANAGE-4.1

OWASP LLM Top 10: LLM05

What are the technical details?

Original Advisory

The WPBot – AI ChatBot for Live Support, Lead Generation, AI Services plugin for WordPress is vulnerable to Stored Cross-Site Scripting via the 'conversation' parameter in all versions up to, and including, 8.4.9 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. The AJAX nonce required to authenticate the save request is publicly emitted on every frontend page via wp_localize_script, making it freely obtainable by any anonymous visitor and removing any practical barrier to exploitation.

Exploitation Scenario

An attacker loads the public-facing page of any WordPress site running WPBot and inspects the page source or network requests to extract the AJAX nonce, which is emitted in cleartext via wp_localize_script for every visitor. Using that nonce, the attacker sends a forged AJAX request to the WPBot conversation-save endpoint with a 'conversation' parameter containing a malicious script payload (e.g., a hidden `<script>` tag that exfiltrates cookies to an attacker-controlled server) — no login or CAPTCHA required. The payload is stored unescaped in the site's chat-session database. Some time later, a site administrator or support agent opens the WPBot 'view chat sessions' report in wp-admin to review recent conversations; the browser renders the stored payload and executes the attacker's script in the admin's authenticated context, exfiltrating their session cookie or performing actions (e.g., creating a new admin user, installing a malicious plugin) on their behalf — resulting in full site compromise from what started as an anonymous, unauthenticated chatbot message.

Weaknesses (CWE)

CWE-79 Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting') Primary CWE-79 Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting')

CWE-79 — Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting'): The product does not neutralize or incorrectly neutralizes user-controllable input before it is placed in output that is used as a web page that is served to other users.

[Architecture and Design] Use a vetted library or framework that does not allow this weakness to occur or provides constructs that make this weakness easier to avoid [REF-1482]. Examples of libraries and frameworks that make it easier to generate properly encoded output include Microsoft's Anti-XSS library, the OWASP ESAPI Encoding module, and Apache Wicket.
[Implementation, Architecture and Design] Understand the context in which your data will be used and the encoding that will be expected. This is especially important when transmitting data between different components, or when generating outputs that can contain multiple encodings at the same time, such as web pages or multi-part mail messages. Study all expected communication protocols and data representations to determine the required encoding strategies. For any data that will be output to another web page, especially any data that was received from external inputs, use the appropriate encoding on all non-alphanumeric characters. Parts of the same output document may require different encodings, which will vary depending on whether the output is in the: etc. Note that HTML Entity Encoding is only appropriate for the HTML body. Consult the XSS Prevention Cheat Sheet [REF-724] for more details on the types of encoding and escaping that are needed. HTML body Element attributes (such as src="XYZ") URIs JavaScript sections Casca

Source: MITRE CWE corpus.