CVE-2024-8060
| CVE ID | CVE-2024-8060 |
| Vendor Disposition | Rejected, already fixed before publication |
| Published | 2025-03-20 |
| Issuing CNA | huntr / Protect AI (from a bounty report) |
| Claimed Severity | High (CVSS 8.1, CVSS:3.0/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:H/A:H) |
The record lists v0.3.0 as affected, together with every later version up to the latest. v0.3.0 is the first version in which the transcription upload no longer keeps the uploaded file name: it stores the file under a server-generated UUID in a fixed directory. The traversal the report describes existed only in v0.2.5 and earlier.
So the record was wrong about every version it named, and the vulnerable versions had been fixed for more than nine months when it was published on 2025-03-20 as a High-severity path traversal leading to remote code execution as root, scored 8.1.
The issuing CNA has withdrawn the record, so the official CVE entry now carries the REJECTED state and NVD mirrors it. The public record now matches Open WebUI's vendor disposition below. There is no advisory.
Timeline
This CVE was withdrawn by its issuing CNA after the dispute was escalated to the CVE Program. Open WebUI's assessment below is the basis on which the record was rejected.
| Date | Event |
|---|---|
| 2025-03-20 | huntr / Protect AI publishes the CVE. |
| 2026-07-22 | Open WebUI publishes this disposition, rejecting the report as already fixed before publication. |
| 2026-07-22 | Open WebUI contacts huntr / Protect AI directly to dispute the record. |
| 2026-08-03 | Open WebUI follows up with huntr / Protect AI. No response is received and the record is unchanged. |
| 2026-08-08 | With no response from huntr / Protect AI, Open WebUI escalates this record to the CVE Program Root. |
| 2026-08-12 | The CVE Program opens a formal dispute with huntr / Protect AI under CVE Program Rule 4.1 (Vulnerability Determination) and requests the CNA's response. |
| 2026-08-12 | huntr / Protect AI replies to the CVE Program and agrees to withdraw the record. |
| 2026-08-13 | Officially rejected The record is updated on cve.org. The CVE record is now officially REJECTED. |
| Still open | The huntr report page has not followed the withdrawal. The report is still shown as Valid with a green check and its status as "Awaiting fix", while the CVE field on the same page reads Rejected. |
| 2026-08-14 | With the report status unchanged, Open WebUI raises the matter with the CVE Program, asking that records withdrawn by the CNA no longer be presented as valid and awaiting a fix. CNA Rule 4.5.2.3 states that a CNA's published vulnerability information "MUST generally support and MUST NOT contradict information published by the CNA in corresponding CVE Records". |
The record is now in the REJECTED state on cve.org, and NVD mirrors that state. No action is required from users, and the CVE should not be treated as an Open WebUI vulnerability.
Disclosure Pathway
A report describing the same endpoint and the same sink, unrestricted file upload via /audio/api/v1/transcriptions, was filed through our GHSA channel on 2024-07-23, eight months before this CVE was published. That report did not describe path traversal or remote code execution; the escalation asserted in this record is not what was reported to us.
What the CVE Claims
In Open WebUI v0.3.0, the audio transcription endpoint /audio/api/v1/transcriptions is claimed to perform insufficient validation on file.content_type and to allow user-controlled filenames, producing a path traversal by which an authenticated user could overwrite critical files in the container and achieve remote code execution as root. The record classifies this as CWE-22 (Path Traversal) and scores it CVSS 8.1 (High).
Why This Is Not a Vulnerability
The record names v0.3.0 and every later version. From v0.3.0 on, the handler generates the destination name itself:
ext = file.filename.split(".")[-1]
id = uuid.uuid4()
filename = f"{id}.{ext}"
file_dir = f"{CACHE_DIR}/audio/transcriptions"
file_path = f"{file_dir}/{filename}"The on-disk name is a server-generated UUID under a fixed directory. The only attacker-influenced component is ext.
The traversal itself was real in earlier versions. Up to v0.2.5, the same endpoint saved each upload as {UPLOAD_DIR}/{file.filename} for any signed-in user, so a crafted file name could write outside the upload directory. A refactor of the handler, first released in v0.3.0 on 2024-06-09, replaced that with the UUID name above.
The Attacker-Controlled Part Cannot Traverse
In v0.3.0 and later, ext is file.filename.split(".")[-1], the substring after the final dot. Because it is a split-on-"." segment, it cannot contain a dot at all, and therefore cannot contain the .. sequence a traversal requires. The UUID is also a literal prefix on the final path component ({id}.{ext}), so ext cannot begin the component either. At most ext could contain a forward slash, which would place the file in a subdirectory beneath the transcriptions directory, never above it.
In every version the record names there is consequently no traversal primitive, no way for a write to escape the transcriptions cache directory, and no arbitrary-file overwrite.
Applicable Security Policy Rules
- Supported Versions: An issue already fixed at the time of filing is not accepted. The traversal was fixed in v0.3.0, before this record and the report behind it.
- Rule 1: A vulnerability must be a real, exploitable weakness. In every version the record names, the reported traversal does not exist: the on-disk name is a server-generated UUID, and the only user-influenced component cannot contain a traversal sequence.
- Rule 3: Reports must reflect an understanding of the affected components. The record names v0.3.0, the first version that no longer keeps the sender's file name, as affected.
Impact to Users
No action required on v0.3.0 or later: uploaded transcription files are stored under a server-generated UUID name in a fixed cache directory. If you still run v0.2.5 or earlier, upgrade.