Arbitrary command execution via crafted prediction paths
Published Aug 1, 2023 · Updated Oct 15, 2024
OS command injection in MLflow before 2.6.0 allows local users to execute arbitrary commands through crafted prediction paths. The PyFuncBackend predict workflow interpolates input and output path values into a Python command string that is passed through a shell without safe argument separation. Exploitation requires access to invoke the local models predict CLI and runs commands with the invoking process's privileges.
Summary
What happened
OS command injection in MLflow before 2.6.0 allows local users to execute arbitrary commands through crafted prediction paths. The PyFuncBackend predict workflow interpolates input and output path values into a Python command string that is passed through a shell without safe argument separation. Exploitation requires access to invoke the local models predict CLI and runs commands with the invoking process's privileges.
The record
- CVE
- CVE-2023-4033
- Published
- Aug 1, 2023
- Updated
- Oct 15, 2024
- Vendor
- MLflow Project
- Product
- MLflow
- Classifications
- CWE-78, T1059.004
- Attack vector
- local
- Privileges
- authenticated
Timeline
How it unfolded
- Aug 1, 2023CVE publishedPublication date reported by the CVE source.
- Oct 15, 2024Record 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=0 <2.6.0
- Affected versionversion=unspecified <2.6.0
What conditions does exploitation require?
What is affected?
Published CVSS scores
CVSS describes severity. EPSS estimates exploitation probability.
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
- MLflow prediction-path command-injection regression payloadsproof 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