NewApp Monitoring is liveSign up free
AgentbeamBeta
AI agent vulnerability · Open WebUI

CVE-2026-44568: Open WebUI has Stored XSS in Pending User Overlay via Incorrect DOMPurify Application Order

mediumCVSS 3.1 4.8GHSA-fq3v-xjjx-95rc

Open WebUI up to 0.8.12 has a stored XSS in the Pending User Overlay. An admin-set message is sanitized before markdown conversion instead of after, so a javascript: link runs in the browser of any pending user who clicks it. Upgrade to 0.9.0. Severity is medium.

Published May 8, 2026 · Updated May 15, 2026 · Analysis Nov 4, 2026

What an attacker can do

AccountPending.svelte renders the admin-configured Pending User Overlay Content with {@html marked.parse(DOMPurify.sanitize(...))}. DOMPurify cleans the raw markdown text, which has no HTML tags, so it removes nothing. Then marked.parse() turns a markdown link into an anchor with a javascript: href, and {@html} inserts it unsanitized. The correct order is DOMPurify.sanitize(marked.parse(src)), which the codebase uses in NotebookView.svelte.

The attacker must be an admin who can edit Admin Settings, General, Pending User Overlay Content. The victims are users with the pending role, who click the link, for example one labelled Contact Support. The script can read their token, show a fake login form or redirect them.

Because only admins can set the text, the risk is lower: a malicious or compromised admin account in a multi-admin deployment attacking users who have not been approved yet. This is why the rating is medium, 4.8, with high privileges required.

Who is affected

Open WebUI up to and including 0.8.12 (pip package open-webui). The reporter reproduced it on v0.8.10. Only instances that set a custom pending overlay and use the pending default role are practically affected.

pip show open-webui

open-webui

Ecosystem
PyPI
Affected
<= 0.8.12
Fixed in
0.9.0
Check
pip show open-webui

How to fix it

Upgrade to Open WebUI 0.9.0:

pip install --upgrade open-webui

The advisory proposes applying DOMPurify after marked.parse(). It names no workaround. Meanwhile, review the Pending User Overlay Content field and remove any link that uses a javascript: URL.

How to tell if you were hit

Open Admin Settings, General, and read the Pending User Overlay Content field for javascript: in markdown links. Check admin audit trails for recent changes to that setting. Review proxy logs for outbound requests from browsers of pending users to unfamiliar hosts.

Original advisory

From the GitHub Advisory Database (CC BY 4.0).

Vulnerability Details

CWE-79: Cross-site Scripting (XSS)

The AccountPending.svelte component renders the admin-configured "Pending User Overlay Content" using marked.parse() inside {@html} with an incorrect DOMPurify application order:

Vulnerable Code

`src/lib/components/layout/Overlay/AccountPending.svelte` (lines 43-48):

{@html marked.parse(
    DOMPurify.sanitize(
        ($config?.ui?.pending_user_overlay_content ?? '').replace(/\n/g, '
')
    )
)}

DOMPurify is applied to the raw Markdown input before marked.parse() processes it. This is the wrong order. DOMPurify sanitizes the Markdown text (which contains no HTML tags), then marked.parse() converts Markdown link syntax into HTML ` tags with javascript: href, and the result is rendered with {@html}` unsanitized.

The correct pattern (used elsewhere in the codebase, e.g., NotebookView.svelte:77) is:

DOMPurify.sanitize(marked.parse(src))  // sanitize AFTER markdown parsing

Steps to Reproduce

Prerequisites

  • Open WebUI v0.8.10
  • Admin account
  • A second user account with "pending" role

Steps

  1. Log in as admin and navigate to Admin Settings → Settings → General.
  1. Set Default User Role to pending.
  1. In the Pending User Overlay Content field, enter:
# Account Pending

Your account is under review.

[Contact Support](javascript:alert(document.domain))
  1. Save the settings.
  1. In a separate browser (or incognito window), create a new account or log in as a pending user.
  1. The pending overlay is displayed. Click the "Contact Support" link.
  1. A JavaScript alert dialog appears showing localhost (the document domain), confirming XSS execution.

Verified Output

The alert(document.domain) executes successfully, displaying "localhost" in a JavaScript dialog box.

Impact

An admin can inject arbitrary JavaScript into the Pending User Overlay Content that executes in the browser context of any pending user who views the overlay page. This could be used to:

  • Session hijacking: Steal pending users' JWT tokens from cookies/localStorage
  • Credential theft: Replace the pending overlay with a fake login form
  • Phishing: Redirect pending users to malicious sites

While this requires admin privileges to set the overlay content, it enables an admin to attack pending users (who have not yet been granted full access). In multi-admin deployments, a compromised admin account could use this to escalate attacks.

Proposed Fix

Apply DOMPurify after marked.parse(), not before:

<!-- Before (vulnerable): -->
{@html marked.parse(
    DOMPurify.sanitize(
        ($config?.ui?.pending_user_overlay_content ?? '').replace(/\n/g, '
')
    )
)}

<!-- After (fixed): -->
{@html DOMPurify.sanitize(
    marked.parse(
        ($config?.ui?.pending_user_overlay_content ?? '').replace(/\n/g, '
'),
        { async: false }
    )
)}

Credited to morimori-dev, Classic298.

Questions

Can a normal user exploit CVE-2026-44568?

No. Only an admin can set the overlay content. The victims are pending users who click the link.

Why is the fix about order of calls?

Sanitizing markdown before parsing does nothing, because the dangerous HTML is created by marked afterward. The fix sanitizes the parsed output.

Which version fixes it?

Open WebUI 0.9.0. Versions up to 0.8.12 are listed as vulnerable.

References

How this page is made: the data panel comes from the GitHub Advisory Database, CISA’s KEV catalog, the NVD and FIRST EPSS, and updates twice a day. The analysis is written by an AI model from the sources listed above and checked by automated rules before it is published. Check versions against the vendor's advisory before you act. Methodology

More Open WebUI vulnerabilities

All Open WebUI entries