OpenClaw versions before 2026.8.1 do not enforce proper access controls for operator command cron jobs. As a result, callers that can see the model can read and execute command jobs that have no owner defined. This flaw allows attackers to inspect stored environment variables and force the execution of disabled or unscheduled command jobs, thus gaining access to secrets and the ability to run operator‑authored commands. The primary consequence is the potential compromise of confidentiality, integrity, and availability of the system, with the flow culminating in the execution of arbitrary code supplied by an attacker.

The affected product is OpenClaw by OpenClaw. Only installations running a version earlier than 2026.8.1 are subject to this vulnerability. No additional product or vendor details were provided beyond the OpenClaw product line.

The CVSS score of 8.7 indicates a high severity, placing the flaw in the high and exploitability risk category. EPSS data is not available, so the current exploitation probability cannot be quantified; however, the flaw is not listed in the CISA KEV catalog. The likely attack vector is through an exposed model‑visible agent interface that can read operator command cron jobs. Once the attacker has read visibility, they can enforce the execution of previously disabled or unscheduled command jobs, directly enabling the retrieval of secrets and execution of arbitrary operator commands. This path suggests that a successful exploitation would provide attackers with the rights of an operator or higher, including full control over the affected system.