Source: LiteLLM GitHub security advisories list

BerriAI's published GitHub security advisories for LiteLLM; as of November 2026 the first page shows ten advisories from May to November 2026, four of them critical.

Contents

Summary

The advisory list shows a steady stream of proxy vulnerabilities through 2026, concentrated in authentication, credential handling and request parameters that make the server fetch URLs or read files. Only the first page was read; more advisories exist on later pages.

Key claims

  • Critical: authentication bypass via Host header injection (GHSA-4xpc-pv4p-pm3w, 2026-05-28); privilege escalation to proxy admin via reuse of the salt key (GHSA-7hp6-4w63-5g45, 2026-09-30); unauthenticated file read that leaks the master key on Enterprise proxies (GHSA-w4c7-34wj-hqm8, 2026-10-27).
  • High: an internal_user_viewer can read local files via vertex_ai_credentials (GHSA-g5ff-637f-6q2m, 2026-10-01); authenticated secret exfiltration via Azure Entra credential parameters in the request body (GHSA-2rwg-2wh4-7c4f, 2026-11-12).
  • Moderate: SSRF and provider-credential exfiltration through unvalidated request-body routing parameters (GHSA-3cv6-jpf6-8222) and via user_config (GHSA-hx8v-g79f-8w5f), both 2026-08-26; path traversal in Skills archive extraction (GHSA-5jmr-gcrj-2c9q, 2026-06-30); reflected XSS in /sso/debug/callback (2026-09-30).
  • Low: local file read via request-supplied OIDC file references (GHSA-4g5m-c9r5-49xf, 2026-06-30).

Feeds

LiteLLM

Suggest changes →