OpenClaw Authorization Bypass Puts Operator Privileges at Risk

The National Vulnerability Database has disclosed CVE-2026-41299, a high-severity authorization bypass in OpenClaw before version 2026.3.28. This flaw specifically impacts the chat.send gateway method, allowing authenticated operator clients to spoof Advanced Control Protocol (ACP) identities.

The vulnerability stems from insufficient validation; ACP-only provenance fields are gated by self-declared client metadata during the WebSocket handshake, rather than verified authorization states. This means an attacker can manipulate client metadata during connection to inject reserved provenance fields, effectively bypassing intended ACP bridge restrictions and escalating privileges.

For defenders, this is a clear signal to audit your OpenClaw deployments immediately. An attacker doesn’t need to bypass authentication, only to be an authenticated operator. This significantly lowers the bar for exploitation, as internal threats or compromised operator accounts can leverage this to inject unauthorized commands or data, leading to integrity violations and potential operational disruption. Patching to version 2026.3.28 or later is critical to mitigate this risk.