LiteLLM, an open-source LLM API proxy, has been hit by multiple security advisories from OSV (Open Source...

LiteLLM, an open-source LLM API proxy, has been hit by multiple security advisories from OSV (Open Source Vulnerabilities) in late June 2026. The advisories detail several high-severity vulnerabilities:

• CVE-2026-49468 (GHSA-4xpc-pv4p-pm3w): Authentication bypass via Host Header Injection in proxy requests (source).
• CVE-2024-2952 (GHSA-46cm-pfwv-cgf8): Server-Side Template Injection in the /completions endpoint (source).
• CVE-2024-5751 (GHSA-gppg-gqw8-wh9g): Remote code execution via unsafe use of eval (source).
• CVE-2026-35030 (GHSA-jjhc-v7c2-5hh6): Authentication bypass via OIDC userinfo cache key collision (source).
• CVE-2026-42208 (GHSA-r75f-5x8p-qvmc): SQL injection in Proxy API key verification (source).

A routine PyPI release (v1.89.6) was also published but does not specifically mention fixes for these CVEs. The combination of authentication bypass, RCE, and SQL injection vulnerabilities poses a serious risk to any deployment. Organizations using LiteLLM should urgently assess their exposure and apply patches as soon as they become available.

What the sources said:

• OSV advisory: "LiteLLM: Authentication Bypass via Host Header Injection" (source).
• OSV advisory: "litellm vulnerable to remote code execution based on using eval unsafely" (source).
• OSV advisory: "LiteLLM has SQL Injection in Proxy API key verification" (source).