Executive brief
OpenClaw is a library used to build bot command systems with access control policies. The QQBot feature contains a flaw where authenticated users can invoke slash commands before access control policies (allowFrom checks) are applied, potentially allowing blocked senders to execute commands. Impact depends on the operator's configuration and what the deployed commands do.
Technical details
This is an incorrect authorization (CWE-863) vulnerability in the QQBot feature of OpenClaw. The root cause is that slash command dispatch occurs before the configured allowFrom access control policy is evaluated, allowing the authorization check to be bypassed. The attack vector is network-based and requires an authenticated sender with ability to invoke slash commands; no additional privileges or user interaction is required. An attacker can dispatch a slash command to a target that should have been blocked by allowFrom policy. The vulnerability is fixed in version 2026.4.27 and later. Operators should upgrade immediately or restrict QQBot slash command exposure until patched.
Affected products
- OpenClaw OpenClaw < 2026.4.27
Timeline
- 2026-05-28: disclosed: GHSA-77pv-3w4q-vrj5 (primary advisory) published
- 2026-06-12: advisory: NVD published CVE-2026-53834
- 2026-06-13: advisory: GHSA-35c7-4r45-9gv3 (duplicate) published and later withdrawn
- 2026-04-27: patched: Version 2026.4.27 released with fix