Before you go live, run this command: openclaw security audit. OpenClaw ships with a built-in security audit that scans your entire Gateway configuration for vulnerabilities.
What does it check for? Agents with the Lethal Trifecta, overly permissive tool access, missing mention gating on group chats, and agents without fallback models. Each finding gets a severity rating: critical, warning, or info. Fix all critical findings before going live. Warnings are worth addressing but won't immediately put you at risk. Run the audit after every configuration change to catch regressions. Make it a habit.