BaseFortify by Axxemble
| Product | BaseFortify, a product of Axxemble (basefortify.eu, contact at axxemble.nl) |
| Problem | Every entry carries the rejection notice, and most still show a CVSS score, an EPSS probability and an AI-generated assessment that contradicts it |
| Records still shown as active | 0 of 26 (checked 2026-10-30). All 26 carry a Rejected badge; 21 still show a CVSS score, all 26 an EPSS probability and 19 an AI summary of the original vulnerability. |
| First contacted | 2026-08-08 |
| Channels tried | [email protected] |
| Status | Awaiting response |
Records
We checked all 26 identifiers rejected against Open WebUI on 2026-10-30. Every one has a BaseFortify entry with a "Rejected - Do Not Use" badge and the rejection reason as its description. The rest of most entries still presents a finding. 21 of them show a CVSS base score and severity, and 19 show an AI Executive Summary describing the original vulnerability. All 26 show an EPSS probability and percentile. The scores and EPSS values appear once the page's JavaScript runs. Each identifier below is in the REJECTED state at cve.org and at NVD:
| Identifier | Withdrawn by | Rejected since | BaseFortify page | Still shown beside the Rejected badge | Our assessment |
|---|---|---|---|---|---|
| CVE-2024-7033 | huntr / Protect AI | 2026-07-16 | basefortify.eu | CVSS 7.2 HIGH, EPSS | Disposition |
| CVE-2024-7034 | huntr / Protect AI | 2026-07-16 | basefortify.eu | CVSS 7.2 HIGH, EPSS, AI summary | Disposition |
| CVE-2024-7036 | huntr / Protect AI | 2026-08-13 | basefortify.eu | CVSS 7.5 HIGH, EPSS, AI summary | Disposition |
| CVE-2024-7037 | huntr / Protect AI | 2026-08-13 | basefortify.eu | CVSS 7.2 HIGH, EPSS | Disposition |
| CVE-2024-7038 | huntr / Protect AI | 2026-07-16 | basefortify.eu | CVSS 2.7 LOW, EPSS, AI summary | Disposition |
| CVE-2024-7039 | huntr / Protect AI | 2026-07-16 | basefortify.eu | CVSS 6.7 MEDIUM, EPSS, AI summary | Disposition |
| CVE-2024-7040 | huntr / Protect AI | 2026-07-16 | basefortify.eu | CVSS 4.9 MEDIUM, EPSS, AI summary | Disposition |
| CVE-2024-7045 | huntr / Protect AI | 2026-08-13 | basefortify.eu | CVSS 4.3 MEDIUM, EPSS, AI summary | Disposition |
| CVE-2024-7046 | huntr / Protect AI | 2026-08-13 | basefortify.eu | CVSS 4.3 MEDIUM, EPSS, AI summary | Disposition |
| CVE-2024-7049 | huntr / Protect AI | 2026-09-11 | basefortify.eu | CVSS 5.4 MEDIUM, EPSS | Disposition |
| CVE-2024-7053 | huntr / Protect AI | 2026-08-13 | basefortify.eu | CVSS 9.0 CRITICAL, EPSS | Disposition |
| CVE-2024-7806 | huntr / Protect AI | 2026-08-31 | basefortify.eu | CVSS 8.8 HIGH, EPSS | Disposition |
| CVE-2024-7959 | huntr / Protect AI | 2026-07-16 | basefortify.eu | CVSS 7.7 HIGH, EPSS, AI summary | Disposition |
| CVE-2024-7990 | huntr / Protect AI | 2026-08-13 | basefortify.eu | CVSS 8.4 HIGH, EPSS, AI summary | Disposition |
| CVE-2024-7999 | huntr / Protect AI, as a duplicate of CVE-2024-53981 | 2025-04-15 | basefortify.eu | EPSS, AI summary | Disposition |
| CVE-2024-8060 | huntr / Protect AI | 2026-08-13 | basefortify.eu | EPSS, AI summary | Disposition |
| CVE-2024-9840 | huntr / Protect AI, as a duplicate of CVE-2024-53981 | 2025-04-15 | basefortify.eu | EPSS, AI summary | Disposition |
| CVE-2024-12534 | huntr / Protect AI | 2026-08-13 | basefortify.eu | CVSS 7.5 HIGH, EPSS, AI summary | Disposition |
| CVE-2024-12537 | huntr / Protect AI | 2026-08-13 | basefortify.eu | CVSS 7.5 HIGH, EPSS, AI summary | Disposition |
| CVE-2024-12868 | huntr / Protect AI, as a duplicate of CVE-2024-47874 | 2025-04-15 | basefortify.eu | EPSS | Disposition |
| CVE-2025-15603 | VulDB | 2026-06-18 | basefortify.eu | EPSS | Disposition |
| CVE-2025-29446 | MITRE | 2026-06-29 | basefortify.eu | CVSS 3.3 LOW, EPSS, AI summary | Disposition |
| CVE-2025-63391 | MITRE | 2026-06-29 | basefortify.eu | CVSS 7.5 HIGH, EPSS, AI summary | Disposition |
| CVE-2026-0765 | Zero Day Initiative | 2026-09-02 | basefortify.eu | CVSS 8.8 HIGH, EPSS, AI summary | Disposition |
| CVE-2026-0766 | Zero Day Initiative | 2026-09-02 | basefortify.eu | CVSS 8.8 HIGH, EPSS, AI summary | Disposition |
| CVE-2026-0767 | Zero Day Initiative | 2026-09-02 | basefortify.eu | CVSS 6.5 MEDIUM, EPSS, AI summary | Disposition |
CVE-2024-7999, CVE-2024-9840 and CVE-2024-12868 were rejected as duplicates of CVEs filed against third-party libraries.
CVE-2024-7959 is a typical example. Beside the Rejected badge and the withdrawal notice it shows a CVSS base score of 7.7 HIGH, an EPSS probability of 24.4% at the 97.6th percentile, a CPE for Open WebUI 0.3.8 and an attack-flow graph from CWE-918 to the CAPEC pattern for server-side request forgery. Its AI Executive Summary states that the flaw "can be exploited to access internal services and potentially execute commands by accessing instance secrets".
The page states the record is rejected, and then describes the vulnerability
On 2026-10-30, 19 entries still did this, and CVE-2024-7959 above is one of them. The clearest case we recorded is CVE-2025-15603 as it stood on 2026-08-08, described below. BaseFortify has since corrected that entry: on 2026-10-30 it carried a "Rejected - Do Not Use" badge, no CVSS data and no AI summary, and still showed its EPSS probability.
The BaseFortify entry for CVE-2025-15603 reproduces the CNA's withdrawal in full, verbatim, in its Description field, including the vendor explanation recorded alongside it:
** REJECT ** DO NOT USE THIS CANDIDATE NUMBER. ConsultIDs: none. Reason: This candidate was withdrawn by its CNA. Further investigation showed that it was not a security issue. Notes: The vendor explains: "The 't0p-s3cr3t' default was dead code on every supported startup path: start.sh, start_windows.bat and
open-webui serveall set or auto-generate WEBUI_SECRET_KEY before the backend imports env.py. It was only ever reachable by invoking uvicorn directly, which is unsupported and unsafe (the app would then sign tokens/cookies with a public, hardcoded key)."
That text includes our own explanation of why the reported behaviour was never exploitable. It is on the page.
On 2026-08-08, directly beside it, in an AI Quick Actions panel occupying the right-hand column of the same screen, a generated Executive Summary reached the opposite conclusion:
The vulnerability arises because the argument WEBUI_SECRET_KEY is manipulated in a way that leads to insufficiently random values being generated for security-critical keys. [...] The vulnerability can be exploited remotely without authentication, although the attack requires a high level of complexity and is considered difficult to execute.
There is no vulnerability and it was never remotely exploitable. The withdrawal notice explaining both sat level with the generated summary, on the same screen, and both were visible at once.
A reader therefore saw the correct answer and the wrong one in the same glance, and only one of the two was written in the register of an assessment. The rejection was a quoted administrative notice. The contradiction beside it was a confident explanation of how the vulnerability works.
The word "rejected" appears once, in the description, and nowhere else
On 2026-08-08, every other element of the entry presented an ordinary, current finding:
- The status badge on the record read Received.
- The title read "Insufficient Entropy in open-webui JWT Key Handler Allows Remote Attack".
- CVSS scores were offered across versions 4.0, 3.1 and 2.0, with a base score of 2.9, a base severity of LOW and a radar chart plotting the vector across eight axes.
- EPSS Scores were displayed as current risk data, with a probability and a percentile.
- An Affected Vendors & Products table listed a CPE for
open-webui, from 0.6.0 through 0.6.16. - An Attack-Flow Graph mapped the record onto CWE-330 and CWE-310, then onward to CAPEC attack patterns including "Brute Force", "Signature Spoofing by Key Recreation" and "Session Credential Falsification through Prediction", and from there to MITRE ATT&CK techniques.
A withdrawn record had been given a severity, an affected version range, an exploitation-probability score and a threat-technique mapping. None of it describes anything that exists.
It also meant the rejection was invisible to anything that does not parse the description prose. A consumer reading the status field saw "Received". A consumer reading severity saw LOW. A consumer reading the affected range saw 0.6.0 through 0.6.16.
The withdrawal was a single field inside a complete presentation of a live vulnerability: a generated summary explaining how it is exploited, a scored vector rendered as a chart, an exploitation-probability estimate, an affected version range, a weakness classification and a mapped attack chain. Take the description away and nothing on the page would have looked unusual for a current, unpatched finding.
Generated assessments need a rejection check before anything else
Presenting machine-generated exploitability analysis on a rejected record is worse than presenting nothing. It converts a correction the CNA already published into a fluent, authoritative paragraph telling the reader the opposite.
The failure is not specific to this vendor, and the fix is one check placed before generation rather than after it. We have written that up separately, since it applies to every system that puts a model in front of vulnerability data: AI Summaries of Rejected CVEs.
Contact log
| Date | Channel | Outcome |
|---|---|---|
| 2026-08-08 | [email protected], the address in their security.txt | Awaiting response |
See also
- Rejected CVEs in Vulnerability Databases, the overview and how to verify any record yourself.
- CVE-2025-15603 vendor disposition
- Vendor Dispositions, our assessment of each identifier listed above.