CVE-2026-42203

The record

LiteLLM is a proxy server (AI Gateway) to call LLM APIs in OpenAI (or native) format. From version 1.80.5 to before version 1.83.7, the POST /prompts/test endpoint accepted user-supplied prompt templates and rendered them without sandboxing. A crafted template could run arbitrary code inside the LiteLLM Proxy process. The endpoint only checks that the caller presents a valid proxy API key, so any authenticated user could reach it. Depending on how the proxy is deployed, this could expose secrets in the process environment (such as provider API keys or database credentials) and allow commands to be run on the host. This issue has been patched in version 1.83.7.

Affected packages

Related vulnerabilities

• GHSA-xqmj-j6mv-4862 LiteLLM: Server-Side Template Injection in /prompts/test endpoint Same vulnerability
• CVE-2026-96392 Improper Neutralization of Special Elements Used in a Template Engine vulnerability in… CWE-1336
• CVE-2026-96355 Improper Neutralization of Special Elements Used in a Template Engine vulnerability in… CWE-1336
• CVE-2026-104851 fsspec is a specification and Python implementation framework for filesystem interfaces. CWE-1336
• CVE-2026-9160 Improper neutralization of special elements used in a template engine vulnerability in… CWE-1336
• CVE-2026-19584 Velociraptor allows for the creation of notebook backups in its default enabled daily… CWE-1336
• CVE-2026-75650 Adobe Commerce is affected by an Improper Neutralization of Special Elements Used in a… CWE-1336
• CVE-2026-52762 YesWiki is a wiki system written in PHP. CWE-1336