CVE-2024-1483 is a high severity (CVSS 7.5) path traversal in MLflow 2.9.2. Crafted HTTP POST requests with specially formed artifact_location and source values can reach files outside the intended directory on the server. The record names no fixed version, so upgrade past 2.9.2 and check the release notes.
What's affected
The record says MLflow version 2.9.2 is affected. The record names no fixed release. Its only reference is a huntr bounty report, which I did not retrieve.
Because MLflow publishes frequent releases, the practical step is to move to the newest release and read its notes for path handling fixes. Several path traversal records were filed against 2.9.2 around the same date, including CVE-2024-1558, CVE-2024-1560, CVE-2024-1593 and CVE-2024-1594, so a single upgrade should address the whole group. Confirm that against the release notes.
How to check if you're exposed
Check the installed version in each environment:
pip show mlflow
Look in virtual environments, containers and notebooks, not only on the tracking server. Then find out who can reach the tracking server over HTTP. An MLflow server often runs with no authentication on an internal network, which makes any host on that network a possible requester.
To check a whole codebase at once, run the free in-browser dependency scan at gold.safeguard.sh/scan. You can also look up the CVE in the Gold directory, and use reachability analysis to see whether your code calls the affected path.
What to do now
- Find every MLflow tracking server and its version.
- Upgrade to a release newer than 2.9.2 and read its release notes.
- Put the server behind authentication and a network allow list.
- Run the server under an account that can read only the files it needs.
- Review the server's file permissions for secrets such as cloud credentials and SSH keys.
Why it matters
The record classes this as CWE-22, path traversal. User-supplied locations were validated in a way that a fragment marker in the URI could slip past, so a path that looked local and safe was later resolved to somewhere else.
The danger is the file the server process can read. ML hosts commonly hold cloud keys, tokens and data. The score of 7.5 reflects a network request with no privileges and a confidentiality impact. The record's KEV flag is false.
Try it yourself
Click through it at your own pace, or open it full screen.
Quick check
Is a fixed version named?
No. The record only says 2.9.2 is affected, so check the MLflow release notes.
Are there related CVEs?
Yes. CVE-2024-1558, CVE-2024-1560, CVE-2024-1593 and CVE-2024-1594 are also path traversal records against MLflow.
Sources
- huntr report (listed in the NVD record, not retrieved)
- NVD record data for CVE-2024-1483
Last reviewed: 19 October 2026.
