---
id: "CVE-2025-15036"
page: "https://infrastack.ai/live/CVE-2025-15036"
kind: "vulnerability"
aliases: ["CVE-2025-15036", "GHSA-vhcx-3pq2-4fvc"]
title: "MLFlow path traversal vulnerability"
severity: "critical"
cvss: 10.0
epss: 0.0059
exploited: false
status: "open"
category: "ai-tooling"
products: [{"name": "mlflow", "fixed": ["3.9.0rc0"], "affected": "< 3.9.0rc0", "ecosystem": "pypi"}, {"name": "lfprojects mlflow", "fixed": ["3.9.0"], "affected": "< 3.9.0", "ecosystem": "vendor"}, {"via": "cpe", "name": "mlflow", "fixed": ["3.9.0"], "affected": "< 3.9.0", "ecosystem": "pypi"}]
ecosystems: ["pypi", "vendor"]
tags: ["ai-tooling", "cwe-22", "cwe-29", "ml-platform", "mlflow", "python", "shared-clusters"]
not_affected_if: ["single-user", "trusted-artifacts-only"]
published_at: "2026-03-30T02:16:14+00:00"
first_seen_at: "2026-11-16T15:37:42+00:00"
updated_at: "2026-11-16T16:08:50+00:00"
sources: [{"source": "ghsa", "url": "https://github.com/advisories/GHSA-vhcx-3pq2-4fvc", "first_seen_at": "2026-11-16T15:37:42+00:00"}, {"source": "nvd", "url": "https://nvd.nist.gov/vuln/detail/CVE-2025-15036", "first_seen_at": "2026-11-16T16:08:50+00:00"}]
exploit_sources: [{"type": "poc", "url": "https://huntr.com/bounties/36c314cf-fd6e-4fb0-b9b0-1b47bcdf0eb0"}, {"type": "epss", "score": 0.0059, "url": "https://api.first.org/data/v1/epss?cve=CVE-2025-15036"}]
---

# MLFlow path traversal vulnerability

Link for people: https://infrastack.ai/live/CVE-2025-15036 (this .md is for agents)

MLflow's artifact cache extracts tar.gz archives without checking member paths, so a crafted archive can write files outside the target directory. An attacker who controls the archive can overwrite arbitrary files or gain elevated privileges, escaping the sandbox in shared clusters.

_Summary, check and fix written by a model from the sources; labelled as such._

## Affected

- pypi:mlflow affected < 3.9.0rc0 fixed 3.9.0rc0
- vendor:lfprojects mlflow affected < 3.9.0 fixed 3.9.0
- pypi:mlflow affected < 3.9.0 fixed 3.9.0

## How to check

- Check the installed MLflow version with `pip show mlflow`; versions below `3.9.0rc0` are affected.
- Look for use of `extract_archive_to_dir` in `mlflow/pyfunc/dbconnect_artifact_cache.py` in your deployed code.

## Fix

- `mlflow` moves to `3.9.0rc0` or later.

## Timeline

- 2026-11-16T15:37:42+00:00 ghsa https://github.com/advisories/GHSA-vhcx-3pq2-4fvc
- 2026-11-16T16:08:50+00:00 nvd https://nvd.nist.gov/vuln/detail/CVE-2025-15036

## Documents

- [fix] Patch: github.com https://github.com/mlflow/mlflow/commit/3bf6d81ac4d38654c8ff012dbd0c3e9f17e7e346
- [poc] Exploit: huntr.com https://huntr.com/bounties/36c314cf-fd6e-4fb0-b9b0-1b47bcdf0eb0

## From the sources

A path traversal vulnerability exists in the `extract_archive_to_dir` function within the `mlflow/pyfunc/dbconnect_artifact_cache.py` file of the mlflow/mlflow repository. This vulnerability, present in versions before v3.7.0, arises due to the lack of validation of tar member paths during extraction. An attacker with control over the tar.gz file can exploit this issue to overwrite arbitrary files or gain elevated privileges, potentially escaping the sandbox directory in multi-tenant or shared cluster environments.

---

_Free to read, with fair use of about 60 pages a minute. For higher limits or API access: sales@infrastack.ai_
