Generate an operational runbook from template. Use when the user says "write a runbook", "create an ops guide", "incident response procedure", "how to handle [X] in production", or "document the on-call procedure for".
Installation
Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.
Generate an operational runbook from template. Use when the user says "write a runbook", "create an ops guide", "incident response procedure", "how to handle [X] in production", or "document the on-call procedure for".
disable-model-invocation
false
user-invocable
true
argument-hint
[operation or incident type]
Generate Runbook
Path Resolution
Read workflow.json in the project root
If it exists and docsRepo is ".": this IS the docs repo — use local paths
If it exists and docsRepo is a repo name: resolve via pwsh .claude/skills/tool-worktree/scripts/resolve-repo.ps1 <docsRepo> to get the docs root path. Templates at <resolved>/templates/, output to <resolved>/runbooks/
If no workflow.json: templates at templates/, output to docs/runbooks/
Instructions
Resolve paths (see Path Resolution above)
Read the template at <templates>/runbook.md
Understand the operation/incident:
What triggers this runbook? (Alert, symptom, scheduled task)
What services are involved?
What access/permissions are needed?
Generate the runbook with emphasis on:
Concrete commands: Every diagnosis and resolution step should have copy-pasteable commands
Decision tree: Clear "if X then Y" logic for diagnosis
Escalation path: When and how to escalate
Verification: How to confirm the issue is resolved
Save to <output>/[slug].md
Quality Checklist
Every diagnosis step has a concrete command or dashboard link
Common causes are ordered by likelihood (most likely first)
Resolution steps are copy-pasteable (not abstract instructions)
Rollback procedure is included
Escalation criteria are specific (time-based or condition-based)
Verification steps confirm the fix worked
Prerequisites list required access and tools
Tips
Write for someone who's been paged at 3 AM — be explicit, don't assume context
Include actual commands, not descriptions of commands
Order diagnosis steps by likelihood to minimize mean time to resolution
Link to monitoring dashboards and log queries
Update runbooks after every incident (add what was learned)
