CVE intelligence and bounded remediation

CVE-2026-42271 — BerriAI LiteLLM Command Injection Vulnerability

High CVSS 8.8 CISA KEV

LiteLLM is a proxy server (AI Gateway) to call LLM APIs in OpenAI (or native) format. From version 1.74.2 to before version 1.83.7, two endpoints used to preview an MCP server before saving it — POST /mcp-rest/test/connection and POST /mcp-rest/test/tools/list — accepted a full server configuration in the request body, including the command, args, and env fields used by the stdio transport. When called with a stdio configuration, the endpoints attempted to connect, which spawned the supplied command as a subprocess on the proxy host with the privileges of the proxy process. The endpoints were gated only by a valid proxy API key, with no role check. Any authenticated user — including holders of low-privilege internal-user keys — could therefore run arbitrary commands on the host. This issue has been patched in version 1.83.7.

Severity
High
CVSS
8.8 (3.1)
Published
2026-05-08
CISA KEV
Known exploited
Ecosystem
software/application
Weaknesses
CWE-77, CWE-78

Affected products

  • litellm / litellm
  • redhat / openshift_ai
  • redhat / openshift_ai / 3.4

Showing 3 representative product identities from 4 source matches. Confirm exact affected versions with the linked vendor and NVD evidence.

Matched remediation archetype

Command, code, expression, and template injection

This catalog composition supplies bounded fallback guidance. Explicitly reviewed curated workflows load with the complete record below.

Check exposure

  • Trace untrusted values to process execution, interpreters, evaluators, template engines, dynamic imports, and administrative scripting features.
  • Determine whether the affected path is reachable across each trust boundary and which service account or host privilege it inherits.
  • Review configuration for optional execution features, unsafe compatibility modes, and shell invocation.

Remediate safely

  • Replace string-built commands or evaluated code with fixed operations and structured argument APIs that do not invoke a shell.
  • Use strict allowlists for operation identifiers and reject unexpected input before it reaches any interpreter.
  • Update the affected component and add inert regression tests covering metacharacters, encoding variants, and alternate request paths.

Authoritative sources

Complete CVE record and remediation plan

The detailed catalog view below loads this exact record, its source evidence, and the full seven-phase agentic change plan.