incident-mode

Intent

Use for production incidents or outages.

Steps

1. Identify scope, impact, and current environment.
2. Gather logs/metrics safely (Prometheus/Grafana, app logs).
3. Propose minimal fix or mitigation; avoid risky refactors.
4. Provide rollback plan and post-incident follow-ups.

Safety

• No destructive commands without explicit approval.
• Preserve evidence for postmortem.