CVE-2026-54009
Undergoing Analysis Undergoing Analysis - In Progress
File Content Disclosure in Open WebUI

Publication date: 2026-06-23

Last updated on: 2026-06-23

Assigner: GitHub, Inc.

Description
Open WebUI is a self-hosted artificial intelligence platform designed to operate entirely offline. Prior to 0.9.6, POST /api/chat/completions accepts an image_url.url value that, when it does NOT start with http://, https://, or data:image/, is interpreted as a file id and resolved against the global file table with no ownership check. an authenticated user can therefore set image_url.url to another user's file id, the server reads that file from disk, base64-encodes it, and injects the data URI into the LLM request. the user then prompts the LLM to describe / OCR the file and reads the content back. This vulnerability is fixed in 0.9.6.
CVSS Scores
EPSS Scores
Probability:
Percentile:
Meta Information
Published
2026-06-23
Last Modified
2026-06-23
Generated
2026-06-24
AI Q&A
2026-06-23
EPSS Evaluated
N/A
NVD
EUVD
Affected Vendors & Products
Showing 1 associated CPE
Vendor Product Version / Range
open_webui open_webui to 0.9.6 (exc)
Helpful Resources
Exploitability
CWE
CWE Icon
KEV
KEV Icon
CWE ID Description
CWE-639 The system's authorization functionality does not prevent one user from gaining access to another user's data or record by modifying the key value identifying the data.
Attack-Flow Graph
AI Quick Actions
Instant insights powered by AI
Executive Summary

CVE-2026-54009 is a cross-user file disclosure vulnerability in Open WebUI versions 0.9.5 and earlier. It occurs in the /api/chat/completions endpoint where the image_url.url field can be manipulated. If the URL does not start with http://, https://, or data:image/, the server treats it as a file ID and retrieves the file from disk without checking if the requesting user owns the file.

An authenticated user can exploit this by specifying another user's file ID, causing the server to read that file, encode it in base64, and inject it into a large language model (LLM) request. The attacker can then prompt the LLM to describe or perform OCR on the file, effectively disclosing its contents.

The root cause is missing ownership checks in the function that retrieves files by ID, allowing unauthorized access to other users' files.

Compliance Impact

This vulnerability allows authenticated users to access files belonging to other users without proper authorization checks, leading to unauthorized disclosure of potentially sensitive information.

Such unauthorized access and disclosure of user files can result in violations of data protection regulations and standards like GDPR and HIPAA, which require strict controls on access to personal and sensitive data.

Impact Analysis

This vulnerability allows any authenticated user to read files uploaded by other users without permission. These files can include images and documents with image-like MIME types.

An attacker can access sensitive or private information contained in these files by leveraging the LLM to describe or extract text from them.

The impact is significant because it compromises confidentiality (high confidentiality impact) but does not affect integrity or availability.

Detection Guidance

This vulnerability involves the manipulation of the `image_url.url` field in POST requests to the `/api/chat/completions` endpoint, where an authenticated user can specify a file ID to access files belonging to other users without proper ownership checks.

To detect exploitation attempts on your system or network, you can monitor POST requests to `/api/chat/completions` that contain `image_url.url` values not starting with `http://`, `https://`, or `data:image/`, as these are interpreted as file IDs.

Suggested commands to detect such activity include:

  • Using grep on server logs to find suspicious POST requests: `grep -i 'POST /api/chat/completions' /path/to/access.log | grep -E '"image_url.url":"(?!http://|https://|data:image/)'`
  • Using network monitoring tools like tcpdump or Wireshark to filter HTTP POST requests to `/api/chat/completions` and inspect payloads for `image_url.url` fields with unexpected values.
  • Implementing application-level logging to record the `image_url.url` parameter values and the authenticated user IDs making the requests for further analysis.
Mitigation Strategies

The primary mitigation is to upgrade Open WebUI to version 0.9.6 or later, where this vulnerability has been fixed by enforcing ownership checks on file access.

If upgrading immediately is not possible, consider the following temporary measures:

  • Restrict access to the `/api/chat/completions` endpoint to trusted users only, minimizing the risk of exploitation.
  • Implement additional access control checks in your deployment to verify that file IDs requested via `image_url.url` belong to the authenticated user.
  • Monitor logs and network traffic for suspicious requests as described in the detection section to identify and respond to potential exploitation attempts.
Chat Assistant
Ask questions about this CVE
Hi! I’m here to help you understand CVE-2026-54009. Ask me anything about the vulnerability, its impact, or mitigation strategies.
0/70
EPSS Chart