Kong Gateway authenticated-response cache flaw CVE-2026-92443

Evidence record

Primary or origin disclosure used to qualify the underlying mechanism, stated prerequisites and limits. CVE-2026-92443 GHSA-qhwv-f2p7-7fxg: proxy-cache, ai-semantic-cache and GraphQL cache can store responses to Authorization/Proxy-Authorization requests and serve to another caller. Separate related sanitizer recover_redacted bug cached restored caller PII. Wrong nginx variable used in one semantic-cache authorization check.

Read the original source ↗

Claim-level citations (3)

• supportsKong AI Gateway could reuse authenticated or restored AI responses across callers: CVE-2026-92443 covers cached responses to requests carrying Authorization or Proxy-Authorization headers being served across callers; the public description does not establish that a bearer token itself is included in the leaked response.

  GHSA-qhwv-f2p7-7fxg, Description and referenced Kong changelog

• supportsKong AI Gateway could reuse authenticated or restored AI responses across callers: The public disclosure does not identify a live leaked customer prompt, prove actual token disclosure, or establish that every Kong cache plugin was susceptible in every configuration.

  GHSA-qhwv-f2p7-7fxg, authorized response caching and potential cross-caller replay

• supportsKong AI Gateway could reuse authenticated or restored AI responses across callers: Kong disclosed that some enterprise and AI Gateway caching plugins could store a response generated for a request with Authorization or Proxy-Authorization headers and return it to another client without preserving the originating principal boundary.

  GHSA-qhwv-f2p7-7fxg, authorized response caching and potential cross-caller replay