CVE-2026-42271 BerriAI LiteLLM Command Injection Vulnerability
BerriAI LiteLLM contains a command injection vulnerability that could allow any authenticated user, including holders of low-privilege internal-user keys, to run arbitrary commands on the host.
From the CISA KEV catalogue version 2026.11.06, with Cybergen analysis. Updated 6 Nov 2026.
Should You Worry?
- CISA confirmed exploitation in the wild and added it to the Known Exploited Vulnerabilities catalogue on 8 Jun 2026.
- Rated High severity (CVSS 8.8, scored by a CVE programme partner).
- It has been in the catalogue for 151 days. Systems still unpatched now have almost certainly been scanned for it.
What It Is
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.
Affected Versions
| Product | Versions |
|---|---|
| BerriAI litellm | >= 1.74.2, < 1.83.7 |
From the public CVE record, as supplied by the vendor. Always confirm against the vendor advisory.
What To Do Now
- 1
Find it
Search your asset inventory and external attack surface for BerriAI LiteLLM. Include cloud images, appliances and anything a supplier runs for you.
Attack Surface Management - 2
Fix it
Apply mitigations per vendor instructions, follow applicable BOD 22-01 guidance for cloud services, or discontinue use of the product if mitigations are unavailable.
Vulnerability Management - 3
Hunt for compromise
If an affected system was reachable before you patched, assume it may have been hit. Check logs, new accounts, web shells and outbound connections from the date of exposure.
Incident Response - 4
Prove it is fixed
Rescan or test the system after patching. Upgrades that fail silently and forgotten secondary instances are common.
Penetration Testing
The Compliance Clock
General guidance, not legal advice. Your obligations depend on your sector, scope and jurisdiction.
Not Sure If You Are Exposed?
Our attack surface and vulnerability teams can check your internet-facing estate for CVE-2026-42271 and every other exploited flaw on the list.