How Exposed OpenClaw Deployments Turn Agentic AI Into an Attack Surface

SecurityScorecard STRIKE research identifies tens of thousands of publicly exposed OpenClaw agentic-AI instances—about 35.4% flagged as vulnerable to Remote Code Execution—creating widespread infrastructure exposure that attackers can abuse for code execution, lateral movement, botnets, or cryptomining; the report emphasizes that adoption is outpacing hardening and recommends segmentation, least privilege, MFA, and isolated deployments, with live tracking available at declawed.io.