CVE-2026-59822 — LiteLLM is a proxy server (AI Gateway) to call LLM APIs in OpenAI (or native) fo

LiteLLM is a proxy server (AI Gateway) to call LLM APIs in OpenAI (or native) format. Prior to 1.84.0, LiteLLM's MCP Streamable HTTP endpoint allowed an unauthenticated attacker to use a fabricated Authorization header to trigger an OAuth2 passthrough fallback path that repl

01What is this vulnerability?

LiteLLM is a proxy server (AI Gateway) to call LLM APIs in OpenAI (or native) format. Prior to 1.84.0, LiteLLM's MCP Streamable HTTP endpoint allowed an unauthenticated attacker to use a fabricated Authorization header to trigger an OAuth2 passthrough fallback path that replaced failed LiteLLM key validation with an empty UserAPIKeyAuth() object, allowing requests to reach MCP tooling without a…

02Affected products

No structured affected-product list available — see references below for vendor advisories.

03Active exploitation status

Yes — actively exploited. Added to the CISA KEV catalog on —.

04Recommended remediation

• Patch to a fixed version listed in the vendor advisory (see references below).
• Mitigate with WAF rules, network egress filters, or feature flags where the patch is not yet available.
• Hunt historical logs for exploitation indicators — see Detection signatures below.

05Technical details

For the full vendor write-up, exploit chains, and reference implementations, see the references list in section 09.

06Detection signatures

Open the Sigma generator with a pre-filled prompt for this CVE to draft a starting detection in your stack of choice:

Open in Sigma generator →

07Related CVEs

No directly-cited follow-up CVEs in the KB record for this advisory. The references list in section 09 carries the vendor cross-references.

08Timeline

• Published: 2026-07-08 (1 month ago)
• Last modified: 2026-09-03

09References

No additional references available in the KB record.