The vulnerability stems from unauthenticated HTTP requests bypassing ACL checks when mTLS is disabled. Client agents handle HTTP requests and forward them to servers via RPC. The root cause would be in the client's HTTP request processing logic where authorization checks were omitted for non-mTLS scenarios. While exact function names aren't provided in advisory sources, the architectural pattern implies the vulnerability exists in the client agent's HTTP handler responsible for processing API requests before RPC forwarding. The high confidence comes from the clear description of the attack vector (client HTTP endpoint + RPC flow) and associated CWEs (Missing/Improper Authorization).