Arbitrary file overwrite via unchecked tar extraction
Published Mar 30, 2026 · Updated Mar 30, 2026
Path traversal in MLflow before 3.9.0 allows attackers to overwrite arbitrary files through a crafted tar.gz model artifact. The extract_archive_to_dir function in mlflow/pyfunc/dbconnect_artifact_cache.py passed archive members to tar extraction without validating that their resolved paths remained under the destination directory. Exploitation requires control of an archive that MLflow extracts and can escape a sandbox or inherit the MLflow process's file privileges in shared clusters.
Summary
What happened
Path traversal in MLflow before 3.9.0 allows attackers to overwrite arbitrary files through a crafted tar.gz model artifact. The extract_archive_to_dir function in mlflow/pyfunc/dbconnect_artifact_cache.py passed archive members to tar extraction without validating that their resolved paths remained under the destination directory. Exploitation requires control of an archive that MLflow extracts and can escape a sandbox or inherit the MLflow process's file privileges in shared clusters.
The record
- CVE
- CVE-2025-15036
- Published
- Mar 30, 2026
- Updated
- Mar 30, 2026
- Vendor
- MLflow Project
- Product
- MLflow
- Classifications
- CWE-29, CWE-22, T1565.001
- Attack vector
- network
- Privileges
- unauthenticated
Timeline
How it unfolded
- Mar 30, 2026CVE publishedPublication date reported by the CVE source.
- Mar 30, 2026Record updatedLatest update available in the CVE record.
Exploitability
Present is not the same as exploitable
Compare your product and version with the public record. A matching version still requires validation against your environment.
Is a vulnerable build present?
Compare these published version ranges with your installed build and any vendor patches.
- Affected versionversion=unspecified <3.9.0
What conditions does exploitation require?
What is affected?
Attacks
What attackers are doing with it
Daily unique IPs observed by Shadowserver honeypots for known exploited vulnerabilities (KEVs). Missing observations do not establish an absence of attacks.
Weakness, pattern, technique
Public exploit references
- Huntr MLflow path-traversal proof of conceptproof of concept · demonstrated
Labels summarize the accepted research assessment. They do not indicate a test against your environment.
Technologies
Your stack
See the directory against your own environment.
Your stack
Check the software in your environment
Book a demo to see how Hinoki identifies affected software and validates exploitability in your environment.
Book a demo