Skip to main content

Improper Neutralization of Special Elements used in an OS Command ('OS Command Injection')

CVE-2024-9053

High
9.8/10

Summary

The vllm-project vllm version contains a vulnerability in the "AsyncEngineRPCServer()" RPC server entrypoints. The core functionality "run_server_loop() "calls the function "_make_handler_coro()", which directly uses "cloudpickle.loads()" on received messages without any sanitization. This can result in remote code execution by deserializing malicious pickle data. This issue affects versions from 0.5.4 and after.

  • LOW
  • NETWORK
  • HIGH
  • UNCHANGED
  • NONE
  • NONE
  • HIGH
  • HIGH

CWE-78 - OS Command Injection

The OS command injection weakness (also known as shell injection) is a vulnerability which enables an attacker to run arbitrary OS commands on a server. This is done by modifying the intended downstream OS command and injecting arbitrary commands, enabling the execution of unauthorized OS commands. This has the potential to fully compromise the application along with all of its data, and, if the compromised process does not follow the principle of least privileges, it may compromise other parts of the hosting infrastructure as well. This weakness is listed as number ten in the 'CWE Top 25 Most Dangerous Software Weaknesses'.

References

Advisory Timeline

  • Published