Code execution via unsafe AgentExecutor pickle loading
Published Jun 4, 2024 · Updated Aug 2, 2024
Unsafe deserialization in MLflow 2.5.0 and later allows remote attackers to execute arbitrary code through a malicious AgentExecutor model. The mlflow/langchain/utils.py _load_from_pickle function passes the model tools pickle directly to cloudpickle.load without restricting object reconstruction. Code runs only after a user retrieves and loads the attacker-uploaded model, giving it the privileges and data access of that user process.
Summary
What happened
Unsafe deserialization in MLflow 2.5.0 and later allows remote attackers to execute arbitrary code through a malicious AgentExecutor model. The mlflow/langchain/utils.py _load_from_pickle function passes the model tools pickle directly to cloudpickle.load without restricting object reconstruction. Code runs only after a user retrieves and loads the attacker-uploaded model, giving it the privileges and data access of that user process.
The record
- CVE
- CVE-2024-37058
- Published
- Jun 4, 2024
- Updated
- Aug 2, 2024
- Vendor
- MLflow Project
- Product
- MLflow
- Classifications
- CWE-502, T1204.002
- Attack vector
- network
- Privileges
- unauthenticated
Timeline
How it unfolded
- Jun 4, 2024CVE publishedPublication date reported by the CVE source.
- Aug 2, 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=2.5.0
- Affected versionversion=2.5.0 <=*
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
- HiddenLayer AgentExecutor cloudpickle 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