Live › vulnerability › CVE-2024-6706
Go to dashboardOpen WebUI, a web front end for language models, can be coaxed by a crafted prompt into outputting JavaScript that executes in the page. An attacker who gets a malicious prompt in front of another user can run script in that user's browser session.
Summary written from the sources below. Check the originals before acting.
First seen by InfraStack 2026-10-26 15:12 UTC, 1h before NVD. All times on this page are UTC.
Deployments of Open WebUI older than 0.3.14 where users view model output, including content produced by other users or shared prompts. Installations limited to a single trusted user with no shared content have a much smaller exposure. The sources do not say whether any specific configuration is required.
| Product | Ecosystem | Affected | Fixed in |
|---|---|---|---|
| open-webui | pypi | < 0.3.14 | 0.3.14 |
| openwebui open webui | vendor | = 0.1.105 | – |
pip show open-webui; versions below 0.3.14 are affected.0.3.14 or later.Each row is the first time that source carried this signal. The analyst’s rows are its own findings.
Attackers can craft a malicious prompt that coerces the language model into executing arbitrary JavaScript in the context of the web page.
This page as .md · .json · in your agent: get_signal("CVE-2024-6706")
GitHub Advisory Database, CC-BY 4.0 · NVD data is provided by NIST and is public domain. Posts quoted belong to their authors.