Centrifugo vulnerabilities

Centrifugo before version 6.9.0 allows a remote client to spoof headers such as x-trusted-user in connect, refresh, subscribe, publish, RPC, and related proxy calls when the backend relies on that header for authentication or authorization. The vulnerability arises from copying client-controlled headers into the connection context and forwarding them as trusted backend headers.

In Centrifugo before version 6.8.4, the unidirectional WebSocket transport with compression enabled enforced message size limit before decompression, but after decompression there was no output cap. This allowed unauthenticated requests to /connection/uni_websocket to trigger high memory and CPU consumption.

In Centrifugo before version 6.8.1, dynamic JWKS endpoint verification could reuse a key for one allowed issuer to verify a JWT for another allowed issuer. This occurred because the JWKS cache and singleflight lookup were keyed only by the JWT header kid, not by the resolved JWKS endpoint, issuer, audience, or trust-domain namespace.