OpenClaw Security Checklist

Post-Installation / Pre-Deployment Checklist

Work through each item and verify before treating your instance as production-ready. Each item links to detailed steps where applicable.

• ☐ Gateway bound to localhost (or private IP only)
  The OpenClaw gateway must not be reachable from the public internet. Bind to 127.0.0.1 or a private VLAN; use Tailscale or SSH tunneling for remote access. Network Isolation.
• ☐ Firewall rules applied; gateway port not exposed to the internet
  Use a host firewall (e.g. UFW on Linux) so the gateway port is not accessible from the internet. Network Isolation.
• ☐ Secrets in environment variables or a secrets manager; no secrets in repo
  API keys and tokens must not live in config files committed to version control. Use env vars or a secrets manager; restrict file permissions (e.g. chmod 600) on any local secret files. Credential Management.
• ☐ API spending limits set where supported
  Configure spending or usage limits on LLM and other API keys to limit impact of abuse or leaks. Credential Management.
• ☐ Sandbox or tool restrictions enabled as appropriate
  Enable sandbox mode or restrictive tool policies so the agent cannot run arbitrary shell commands or access sensitive paths unless required. Best Practices (Sandbox & tool restrictions).
• ☐ Logging and (if possible) audit logging configured
  Enable application and, where available, audit logs; set up log rotation and monitor for anomalies. Monitoring & Logging.
• ☐ Installed skills audited; only trusted or reviewed skills in use
  Third-party skills from ClawHub can contain vulnerabilities or malicious code; research (e.g. Snyk) has found a notable portion with credential leak risks. Review source/publisher, pin versions, and run openclaw security audit if available. Skills Security, Known Vulnerabilities.
• ☐ Regular update process (e.g. weekly or after security advisories)
  Plan for weekly OpenClaw and critical skill updates, and apply patches promptly when CVEs or advisories are published. Known Vulnerabilities, Best Practices – Regular Maintenance.

Docker-Specific Checks

If you run OpenClaw in Docker, add these items:

• ☐ docker.sock is not mounted unless strictly required (container escape risk).
• ☐ Only minimal volumes are mounted (config, data-not the whole filesystem).
• ☐ Resource limits (CPU, memory) are set where possible.

Full guidance: Docker Hardening. Installation: Docker Deployment.

Quick Verification Commands

Use these to sanity-check your setup (exact commands may vary by OpenClaw version):

# Check gateway and service status
openclaw status

# Validate configuration
openclaw config validate

# Run security audit (if available)
openclaw security audit

# View recent logs
openclaw logs --tail 50

For diagnostics, see Troubleshooting Guide and Security Best Practices.

Scenario-Based Quick Reference

Tail the checklist to your risk level (from Security for Different Scenarios):

• Personal use (low-stakes): All 8 core items; optional extra firewall and minimal skill set.
• Small business (medium-risk): All items plus VPN/Tailscale for remote access, formal skills review, and consider managed hosting if no dedicated IT.
• Enterprise / regulated (high-stakes): Full isolation, secrets manager, strict sandbox and tool allowlists, audit logging, incident response plan, and regular reviews; align with GDPR/HIPAA as needed.

Related Resources

Next Steps

After completing the checklist:

• Schedule a quarterly run-through and after any major config or skill changes.
• Bookmark Known Vulnerabilities and official security advisories for CVE and ClawHub incident updates.
• Explore Use Cases and ClawHub skills with security in mind-audit before installing.