OpenClaw Security Issues: Add a “Security Guardrail” to Your AI Application

The NSFOCUS blog warns that the OpenClaw AI agent (formerly Clawdbot/Moltbot) poses significant enterprise risks: two high-risk RCEs were disclosed and actively exploited, default/autonomous behaviors and high-permission deployments enable unauthorized access and lateral movement, and a permissive third‑party plugin system creates supply‑chain poisoning opportunities. Enterprises are advised to detect shadow deployments, restrict privileges, monitor agent network activity, and apply layered protections (AI agent discovery, isolation, and firewall inspection).