OpenClaw: risks for the users and how to mitigate them

The report analyzes security risks in the OpenClaw AI agent ecosystem, documenting hundreds of vulnerabilities and widespread distribution of malicious skills on the public hub (hundreds of malicious skills and over a thousand malicious accounts), describes active detections by Kaspersky, and recommends layered defenses including scanning, sandboxing, network monitoring, and strict AI/tool usage policies.