incident-response
Contain an incident, verify recovery and preserve a factual postmortem. Triggers on "incident", "事故响应".
SKILL.mdskills/09-operate/incident-response/SKILL.md
GitHubIncident Response
When to use
- Contain an incident, verify recovery and preserve a factual postmortem.
Not for: debugging for non-urgent diagnosis; shipping for planned releases.
Steps
Assess impact and actual response authority; preserve timeline and evidence and avoid unapproved messaging.
Contain with authorized platform recovery before full root-cause diagnosis.
Check user impact and telemetry after each step; distinguish known facts from suspicions in authorized communication.
Diagnose after stabilization, verify durable fixes and prepare an approved postmortem with impact, cause and owned actions.
Output: Suggested paths, subject to human audit: docs/postmortem/YYYY-MM-DD-<slug>.md at the approved path.
Verify
- Recovery has user-flow and operational evidence; timeline and follow-up remain factual and authorized.
References
- engineering-principles.md — shared decision boundaries, scope and verification.
- human-audit.md — document consent, readability and human review.
- references/runbook-template.md — runbook format for alert-linked playbooks
- references/postmortem-template.md — blameless postmortem template + action-item tracking
- references/dr-planning.md — load during postmortem action-item planning for proactive DR; backup strategy, restore testing, DR drills, failover orchestration