incident-commander
Production incident response protocol, blameless post-mortem authoring, root-cause triage, and mitigation runbooks.
SKILL.mdskills/incident-commander/SKILL.md
GitHubIncident Commander & Production Triage Protocol
When investigating live service disruptions or responding to production outages, follow this structured incident lifecycle.
1. Triage & Stabilization Phase
- Assess Severity (P1-P4):
- P1: Complete service outage or data loss affecting all users.
- P2: Major feature failure without acceptable workaround.
- P3: Minor degradation with available workaround.
- Mitigate First, Debug Later:
- Roll back to the previous stable release commit or toggle feature flags before attempting complex code surgery in production.
2. Blameless Post-Mortem Structure
Document all P1/P2 incidents in docs/incidents/YYYY-MM-DD-title.md:
- Executive Summary: Impact duration, affected users, and revenue/reputation blast radius.
- Timeline: Chronological sequence from trigger event, detection, triage, mitigation, to resolution.
- Root Cause (5 Whys): Deep structural cause, not superficial operator error.
- Action Items: Preventative engineering tasks with assigned owners and due dates.