Writes an operational runbook for a service covering symptoms, diagnostic checks, mitigations, and escalation paths. Use when shipping a new service or when an existing service lacks incident procedures.
Click to play with sound.
---
name: Runbook Writer
description: Writes an operational runbook for a service covering symptoms, diagnostic checks, mitigations, and escalation paths. Use when shipping a new service or when an existing service lacks incident procedures.
---
# Runbook Writer
A runbook is a recipe for a tired engineer at 2 AM. It must answer: what is broken, how do I confirm it, what do I try first, and when do I give up and escalate.
## Required Sections
Every runbook gets exactly these sections in this order.
### 1. Service Overview (3-5 lines)
State what the service does, its criticality (which SEVs it can cause), its SLO target, and its direct dependencies. Skip history and rationale.
### 2. Symptoms and Alerts
List each alert that fires for this service. For each:
- Alert name and what condition triggers it.
- What it means in plain language (not the PromQL).
- Expected SEV range when it fires.
### 3. Diagnostic Checks… load the full skill through Skill Me