NewApp Monitoring is liveSign up free
AgentbeamBeta
AI agent vulnerability · OpenClaw

CVE-2026-53847: OpenClaw: Active Memory write scope could mutate global config

mediumCVSS 4.0 5.3GHSA-x629-46cc-7xgw

OpenClaw 2026.5.5 and earlier let a Gateway caller with only `operator.write` access to an Active Memory command change global configuration. The change should need `operator.admin`. Fixed in 2026.5.6.

Published Jun 18, 2026 · Analysis Oct 26, 2026

What an attacker can do

OpenClaw splits Gateway permissions into scopes such as operator.write and operator.admin. Global configuration is meant to be an admin action. In affected versions the Active Memory command applied global config changes on write scope alone (CWE-266, incorrect privilege assignment).

The attacker needs an authenticated Gateway session that holds operator.write and can reach the command. They send it and alter settings beyond the Active Memory feature. The CVSS 4.0 vector rates the effect as low integrity and low availability impact, with no confidentiality impact. Even so, a config change can weaken other controls, so check what global settings your Gateway exposes.

This matters on shared Gateways where some callers have write scope but not admin. The advisory keeps OpenClaw's trusted-operator model, and the exposure depends on the feature being enabled and reachable.

Who is affected

The npm package openclaw at 2026.5.5 and earlier, with Active Memory enabled and callers who hold operator.write but not operator.admin. Check the installed version:

npm ls openclaw

openclaw

Ecosystem
npm
Affected
<= 2026.5.5
Fixed in
2026.5.6
Check
npm ls openclaw

How to fix it

Upgrade to openclaw 2026.5.6, the first stable patched version:

npm install [email protected]

Until then, limit Active Memory write access to trusted operators. Keep channel and tool allowlists narrow, do not share one Gateway between mutually untrusted users, and disable the feature if you do not need it.

How to tell if you were hit

Compare your global Gateway configuration with a known-good copy and look for changes made by sessions that hold only operator.write. Review Gateway logs for Active Memory commands issued by non-admin callers around config changes. The advisory names no specific log line.

Original advisory

From the GitHub Advisory Database (CC BY 4.0).

Summary

Active Memory write scope could mutate global config. In affected versions, a Gateway caller with operator.write access to the affected command could change global configuration without requiring operator.admin.

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 apply configuration changes beyond the intended write scope. 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

limit Active Memory write access to trusted operators 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.

Credited to zsxsoft, KeenSecurityLab, qclawer.

Questions

Does the caller need admin rights to exploit CVE-2026-53847?

No, that is the flaw. operator.write access to the affected command was enough to change global config. Admin should have been required.

Which version is fixed?

openclaw 2026.5.6. Versions up to 2026.5.5 are affected.

What is the severity?

Medium. The CVSS 4.0 vector shows network access, low privileges, and low integrity and availability impact.

References

How this page is made: the data panel comes from the GitHub Advisory Database, CISA’s KEV catalog, the NVD and FIRST EPSS, and updates twice a day. The analysis is written by an AI model from the sources listed above and checked by automated rules before it is published. Check versions against the vendor's advisory before you act. Methodology

More OpenClaw vulnerabilities

All OpenClaw entries