dragonked2/alphacode/incident-response

Incident Response Skill

Structured approach to handling production incidents.

Response Process

1. Detect — identify the incident (alerts, user reports, anomalies)
2. Triage — assess severity and impact scope
3. Mitigate — stop the bleeding (rollback, feature flag, hotfix)
4. Diagnose — root cause analysis
5. Remediate — fix the underlying issue
6. Review — blameless post-mortem

Rollback Strategy

• Feature flags for instant feature disable
• Database migration rollback scripts
• Blue-green deployment switch
• Traffic rerouting to healthy instances
• Cache invalidation if stale data is the issue

Post-Mortem Template

• Summary: What happened, when, impact
• Timeline: Key events in chronological order
• Root Cause: Technical root cause analysis
• Impact: Users affected, duration, revenue impact
• What Went Well: Things that helped during response
• What Went Wrong: Things that slowed response
• Action Items: Concrete steps to prevent recurrence (with owners and dates)