CVE-2026-114612

LiteLLM is a proxy server (AI Gateway) to call LLM APIs in OpenAI (or native) format. From 1.87.0 until 1.100.4, 1.101.3, 1.102.2, 1.103.1, and 1.104.0-rc.2, the proxy reuses one cryptographic context for at-rest secrets and session tokens, allowing an authenticated internal user to have crafted metadata containing a forged administrator credential encrypted and returned by the API-key flow. Replaying that ciphertext as a bearer token causes the proxy to trust the forged proxy administrator identity, permitting administrative access and arbitrary commands through the MCP stdio endpoint. Versions 1.87.0 through 1.90.x require EXPERIMENTAL_UI_LOGIN to be enabled, while versions after 1.91.0 are exploitable in the default configuration. This issue was fixed in LiteLLM 1.100.4, 1.101.3, 1.102.2, 1.103.1, and 1.104.0-rc.2.

• CWE-269
• CWE-345
• CWE-441

• https://github.com/BerriAI/litellm/commit/1d5fdc87fed66b49e1df8cf7d708ad93249eb7d0
• https://github.com/BerriAI/litellm/commit/9161d7f9d6073603ba61ee4aa5e2c0ef5b37a81b
• https://github.com/BerriAI/litellm/commit/9fa1a641119dd0d4fe43e93622eae5f482ceb63f
• https://github.com/BerriAI/litellm/commit/a80eacacee118fbed54e3d98f239d17a63fac818
• https://github.com/BerriAI/litellm/commit/f7cd90f09e9716652b38daec9f0640cdf49d582b
• https://github.com/BerriAI/litellm/releases/tag/v1.100.4
• https://github.com/BerriAI/litellm/releases/tag/v1.101.3
• https://github.com/BerriAI/litellm/releases/tag/v1.102.2
• https://github.com/BerriAI/litellm/releases/tag/v1.103.1
• https://github.com/BerriAI/litellm/releases/tag/v1.104.0-rc.2
• https://github.com/BerriAI/litellm/security/advisories/GHSA-7hp6-4w63-5g45

Published

Oct 22, 2026

Last Modified

Oct 22, 2026

Status

Received

• Same Subject, Wrong User: A Cross-Issuer Account Takeover in n8nStrix found an identity-binding bug in n8n's token-exchange flow enabling account takeover.
• Autonomous PentestingAI agents that find and validate exploitable vulnerabilities like this one across your applications.
• PR ReviewsPentest every pull request so vulnerable code is caught before it ships to production.
• AI Penetration TestingHow AI-driven penetration testing continuously covers your attack surface.

• CVE-2026-114607Better Auth is an authentication and authorization library f…9.1
• CVE-2026-114608Pylint is a static code analyser for Python. From 0.15.2 unt…5.4
• CVE-2026-114609fpdf2 is a PDF creation library for Python. Prior to 2.8.8, …5.8
• CVE-2026-11461A vulnerability has been found in NousResearch hermes-agent …6.3
• CVE-2026-114610fpdf2 is a PDF creation library for Python. From 2.5.0 until…6.5
• CVE-2026-114611OCO-Server is a self-hosted endpoint management application …9.1
• CVE-2026-114613MessagePack is the serializer implementation for Python msgp…7.5
• CVE-2026-11462A vulnerability was found in Chengdu Everbrite Network Techn…7.3
• CVE-2026-11463A vulnerability was determined in USCiLab Cereal up to 1.3.2…7.3
• CVE-2026-114633WWBN AVideo before 29.4.0 contains an incorrect authorizatio…6.5
• CVE-2026-114634WWBN AVideo before 29.4.0 contains a blind server-side reque…4.3
• CVE-2026-114635Traefik before v2.11.58 and v3.7.14 contains an authenticati…8.7