CVE-2026-53834: OpenClaw < 2026.4.27 - Authorization Bypass in QQBot Pre-dispatch Slash Commands

Synopsis

An authorization bypass vulnerability exists in OpenClaw's QQBot pre-dispatch slash command handling, affecting all versions before 2026.4.27. The flaw is reachable over the network without any authentication, allowing any sender to invoke slash commands before the configured allowFrom access control policies are evaluated. Successful exploitation lets a blocked sender trigger command handling that the operator intended to restrict, bypassing the access control layer entirely. A patched-image rebuild at version 2026.4.27 is available on HarborGuard for environments running an affected version.

HarborGuard Coverage

Exploit Conditions

• Network reachabilityRequired

  The vulnerable slash command endpoint is exposed over the network, so an attacker must be able to reach the service remotely.

• AuthenticationNot required

  No credentials or account are needed; any network sender can trigger the pre-dispatch command path before access control is evaluated.

• Victim interactionNot required

  The attacker sends a crafted slash command directly; no action from an operator or user is required to complete the exploit.

• Attack complexityDetail

  Base exploit logic is straightforward and condition-free, though the CVSS vector notes an attack requirement (AT:P) indicating that specific operator-defined allowFrom policy configuration must be present for the bypass to be meaningful.

Blast Radius

• An attacker submits slash commands that the operator configured access controls to block, causing those commands to execute as if the sender were permitted.
• Restricted command handlers run with their full intended effect, potentially modifying bot behavior, triggering automated workflows, or altering data that those commands control.
• The integrity of the access control boundary is broken; operator-defined allowFrom policies no longer reliably gate command execution.