What problem does it solve? When production is down, teams often waste critical minutes hunting root causes, arguing over severity, or stacking speculative fixes. This Skill enforces a disciplined incident command procedure: mitigate first with reversible actions, assign a single Incident Commander, communicate on a fixed cadence, and verify recovery against real signals before declaring resolved. ## Core Features & Use Cases - Severity Declaration: Maps impact to SEV1-SEV4 using a severity matrix, with a round-up rule when impact is ambiguous. - Mitigation-First Algorithm: Prioritizes reversible levers (rollback, feature flag, failover) over live root-cause analysis, one change at a time. - Communication Cadence: Separates internal responder channels from external status pages with committed update intervals per severity. - Blameless Postmortem: Drafts timelines, 5-whys on contributing factors, and owned action items without blaming individuals. - Use Case: Paged at 2 AM because checkout is throwing 500s after a deploy. The Skill walks you through confirming impact, declaring SEV1, naming an IC, rolling back the deploy, posting stakeholder updates, and verifying recovery before writing the postmortem. ## Quick Start Run the incident procedure for the current production outage where checkout is returning 500 errors after the latest deploy.
