host-firewall-defense

Host And Firewall Defense

Use this when the question is specifically about GuardianAgent host-monitor alerts, sensitive path drift, new external destinations, host firewall posture, or gateway firewall findings.

Workflow

1. Start with the narrowest security view that answers the question.
   • security_alert_search with source: "host" or source: "gateway" when the user is asking about active alerts.
   • host_monitor_status or gateway_firewall_status when they want posture plus counts.
2. Refresh only when stale state is the problem.
   • host_monitor_check
   • gateway_firewall_check
3. Separate:
   • confirmed drift or suspicious activity
   • baseline-relative noise
   • alert-hygiene candidates
4. If the finding is repetitive but expected, use the alert lifecycle tools with a reason and a time bound.

Interpretation Rules

• sensitive_path_change means the fingerprint changed, not that compromise is confirmed.
• new_external_destination means GuardianAgent observed a previously unseen outbound address, not that the address is malicious.
• Treat host and gateway firewall disablement or rule drift as higher-confidence control degradation than a single medium host anomaly.

Boundaries

• Use native-av-management for Windows Defender, Malwarebytes coexistence, scans, signatures, and Controlled Folder Access.
• Use network-recon for deeper network diagnostics beyond the alert queue itself.
• Use security-triage when the work has become a broader incident review rather than focused host or firewall interpretation.
• Use security-alert-hygiene when the primary task is acknowledge/resolve/suppress workflow rather than technical interpretation.

Gotchas

• Do not treat GuardianAgent's own state churn as proof of tampering without corroborating evidence.
• Do not call normal SaaS traffic malicious just because the destination is new to the local baseline.
• Do not suppress broad classes of alerts without narrowing by source, evidence pattern, and time window.