CVE-2026-12774 is a server-side request forgery vulnerability affecting BerriAI litellm up to version 1.82.2. The issue exists in the _execute_with_mcp_client function of the litellm/proxy/_experimental/mcp_server/rest_endpoints.py file, part of the MCP Server Connection Testing component. This flaw allows an attacker with limited privileges to remotely trigger the server to send crafted requests to arbitrary destinations, potentially bypassing network restrictions or accessing internal resources. The vulnerability is publicly disclosed, but no vendor remediation or patch has been announced. The CVSS 4.0 vector indicates network attack vector, low attack complexity, no privileges required, no user interaction, and low to low impact on confidentiality, integrity, and availability.

Successful exploitation of this SSRF vulnerability could allow an attacker to make unauthorized requests from the affected server to internal or external systems. This could lead to information disclosure or interaction with internal services not directly accessible to the attacker. However, the impact is rated medium due to the limited scope and low impact on confidentiality, integrity, and availability as indicated by the CVSS vector. No known active exploitation in the wild has been reported.

As of the current information, no official patch or remediation is available from the vendor. Users should monitor the vendor's advisory channels for updates and apply any forthcoming fixes promptly. Until a fix is released, restrict access to the vulnerable function if possible and consider network-level controls to limit the server's ability to make arbitrary outbound requests. Patch status is not yet confirmed — check the vendor advisory for current remediation guidance.