A critical code injection vulnerability in xorbitsai inference allows unauthenticated remote attackers to execute arbitrary commands on the host server.

This is an improper neutralization of directives vulnerability (CWE-95) where the application passes attacker-influenced input directly to the eval function. An unauthenticated remote attacker can supply a malicious prompt to the chat completion API, resulting in arbitrary code execution within the server process context.

The ability for an unauthenticated attacker to execute arbitrary commands poses a severe risk to organizational infrastructure. Successful exploitation could lead to full system compromise, unauthorized data exfiltration, or the deployment of persistent threats within the internal network. With a CVSS score of 10.0, this represents the highest level of urgency for remediation.

Immediate Action: Upgrade xorbitsai inference to version 2.7.0 or later immediately to eliminate the vulnerable code path.

Proactive Monitoring: Review application access logs for suspicious prompts containing Python-like syntax or unexpected tool-call patterns.

Compensating Controls: Deploy a Web Application Firewall (WAF) with rules designed to detect and block malicious payloads directed at the /v1/chat/completions endpoint.

Public Exploit Available: Unknown

This vulnerability is critical and requires immediate patching. Organizations should prioritize updating their Xinference deployments to version 2.7.0 to prevent potential remote code execution and full system takeover.