A critical vulnerability, CVE-2026-40217, affecting LiteLLM has been added to the CISA Known Exploited Vulnerabilities (KEV) catalog. This vulnerability, which has been exploited in the wild, allows for code execution within the proxy process due to a sandbox escape in the custom-code guardrail test endpoint. Users are advised to upgrade to version 1.83.10-stable or later, or to block POST /guardrails/test_custom_code requests at their gateway if an immediate upgrade is not possible. AI
IMPACT This critical vulnerability in LiteLLM, a tool for simplifying LLM API calls, could impact the security of AI applications relying on it.
RANK_REASON Security advisory for a specific software tool.
Read on Mastodon — mastodon.social →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →