Code execution via unsafe model deserialization
Published Apr 15, 2025 · Updated Apr 21, 2025
Deserialization in XorbitsAI Xinference 1.4.0 and 1.4.1 allows attackers to execute code via malicious CosyVoice model files. The CosyVoiceModel.load method passes llm_model, flow_model, and hift_model to torch.load without weights_only=True, permitting pickle objects to invoke code. The attacker must be able to supply one of those model paths to the load routine; execution inherits the Xinference process's permissions.
Summary
What happened
Deserialization in XorbitsAI Xinference 1.4.0 and 1.4.1 allows attackers to execute code via malicious CosyVoice model files. The CosyVoiceModel.load method passes llm_model, flow_model, and hift_model to torch.load without weights_only=True, permitting pickle objects to invoke code. The attacker must be able to supply one of those model paths to the load routine; execution inherits the Xinference process's permissions.
The record
- CVE
- CVE-2025-3622
- Published
- Apr 15, 2025
- Updated
- Apr 21, 2025
- Vendor
- XorbitsAI
- Product
- Xinference
- Classifications
- CWE-502, CWE-20
- Attack vector
- adjacent
- Privileges
- authenticated
Timeline
How it unfolded
- Apr 15, 2025CVE publishedPublication date reported by the CVE source.
- Apr 21, 2025Record 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=1.4.0
- Affected versionversion=1.4.1
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.
Public exploit references
No public exploit references are available in this record.
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