Free AI Safety Workshop — in collaboration with explainx.aiSave your seat
AgentbeamBeta
AI agent vulnerability · OpenClaw

GHSA-2rqg-gjgv-84jm: OpenClaw: Gateway `agent` calls could override the workspace boundary

highCVSS 3.1 8.8

Summary The public gateway agent RPC allowed an authenticated operator with operator.write to supply attacker-controlled spawnedBy and workspaceDir values. That let the caller re-root the agent run outside its configured workspace boundary. Impact A non-owner operator could escape the intended workspace boundary and run normal file and exec tools from an arbitrary process-accessible directory.

Published Mar 13, 2026

Affected packages

openclaw

Ecosystem
npm
Affected
<= 2026.3.8
Fixed in
2026.3.11
Check
npm ls openclaw

Advisory details

From the GitHub Advisory Database (CC BY 4.0).

Summary

The public gateway agent RPC allowed an authenticated operator with operator.write to supply attacker-controlled spawnedBy and workspaceDir values. That let the caller re-root the agent run outside its configured workspace boundary.

Impact

A non-owner operator could escape the intended workspace boundary and run normal file and exec tools from an arbitrary process-accessible directory.

Affected versions

openclaw <= 2026.3.8

Patch

Fixed in openclaw 2026.3.11 and included in later releases such as 2026.3.12. The gateway now enforces the configured workspace boundary for agent runs regardless of caller-supplied overrides.

Credited to tdjackey.

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. No analysis has been written for this entry yet. Methodology

More OpenClaw vulnerabilities

All OpenClaw entries