Junglewise Threat Intelligence

CVE-2026-42208: BerriAI LiteLLM SQL injection in Proxy API key verification

CVE-2026-42208 · Severity: critical · CVSS 9.8 · Exploited in the wild · Published 2026-05-08

Technologies: Berriai LiteLLM. Vendors: Berriai, PyPI.

Executive brief

LiteLLM is an AI gateway used to manage and proxy requests to various Large Language Model (LLM) providers. A security flaw allows unauthenticated attackers to bypass security checks and access the underlying database. This could result in the theft of sensitive API keys, modification of system data, or full unauthorized access to the proxy service and the AI credentials it manages.

Technical details

A SQL injection vulnerability exists in LiteLLM's proxy API key verification mechanism. The application fails to properly parameterize database queries when checking API keys, instead concatenating caller-supplied values directly into the query string. An unauthenticated attacker can trigger this vulnerable path by sending a specially crafted Authorization header to any LLM API route (such as /chat/completions), which reaches the query via the proxy's error-handling logic. Successful exploitation allows for arbitrary data exfiltration or modification within the proxy's database, potentially leading to the compromise of all managed LLM credentials. The issue is fixed in version 1.83.7 by implementing parameterized queries.

Affected products

  • BerriAI LiteLLM >= 1.81.16, < 1.83.7

Timeline

  • 2026-04-19: patched: Version 1.83.7-stable released
  • 2026-04-20: advisory: GitHub Security Advisory GHSA-r75f-5x8p-qvmc published
  • 2026-05-08: disclosed: CVE-2026-42208 published to NVD
  • 2026-05-08: kev added: Added to CISA Known Exploited Vulnerabilities catalog

References

Related threats