GHSA-p73f-w79w-jqr5: OpenClaw: Native command authorization could skip owner-command enforcement

OpenClaw lets operators mark some commands as owner-only. For native commands, the authorization step in affected versions did not apply that owner-command enforcement. A sender who reached native command handling could therefore authorize a command that policy meant only for the owner.

The attacker needs to be a sender on a channel that reaches native command handling, with low privileges. They send an owner-style native command and it is accepted. The CVSS vector scores high integrity impact and low confidentiality and availability impact, so the effect is unauthorized actions, not full takeover. Which actions depends on which commands your deployment keeps owner-only. The advisory does not name them or the configuration key.

The advisory keeps the trusted-operator model unchanged and says impact depends on the configuration and on whether lower-trust input can reach the path. Gateways shared among several users, or connected to group channels, are the most exposed.

The npm package openclaw at 2026.5.5 and earlier, with native commands enabled and reachable by senders who are not the owner.

npm ls openclaw

Use npm ls -g openclaw for a global install. Versions up to 2026.5.5 are affected.

Upgrade to [email protected], the first stable patched version:

npm install [email protected]

Until then, the advisory says to keep native command surfaces limited to trusted senders. It also advises narrow channel and tool allowlists, not sharing one Gateway between mutually untrusted users, and disabling the feature when it is not needed.

Review Gateway and channel history for native commands that you configured as owner-only and that were issued by senders other than the owner. Check for settings or state that changed after such messages. The advisory gives no log string, so compare the commands that ran against your owner-only policy.

Summary

Native command authorization could skip owner-command enforcement. In affected versions, a sender able to trigger native command handling could authorize a native command without enforcing the configured owner-only command policy.

This advisory is scoped to the named feature and configuration. It does not change OpenClaw's trusted-operator model: authenticated Gateway operators, installed plugins, and intentional local execution surfaces remain trusted unless a separate policy, approval, allowlist, sandbox, or auth boundary is crossed.

Impact

When the affected feature is enabled and reachable, this could run an owner-style command from a sender that should not have that command access. Practical impact depends on the operator's configuration and whether lower-trust input can reach that path.

Patched Versions

The first stable patched version is 2026.5.6.

Mitigations

keep native command surfaces limited to trusted senders until patched. As general hardening, keep channel and tool allowlists narrow, avoid sharing one Gateway between mutually untrusted users, and disable the affected feature when it is not needed.