--- name: cutover-day-runbook description: Produce the cutover-day operational runbook for a migration — hour-by-hour plan, go/no-go gates, comms cadence, rollback triggers, and a minute-by-minute first-60-minutes script. version: 1.0.0 author: VantagePoint Networks author_url: https://www.vpnetworks.co.uk audience: IT Managers, Migration Leads, Project Managers, MSP Technical Consultants, Platform Engineers output_format: Formatted Markdown cutover runbook with pre-cutover checks, hour-by-hour plan, named roles, go/no-go gates, comms cadence, rollback plan, and post-cutover verification. license: MIT last-reviewed: 2026-04 --- # Cutover-Day Runbook A Claude Code skill for the day you're actually flipping the switch — not the migration plan you wrote months ago, but the hour-by-hour runbook the on-the-night team holds in their hand. ## How to use this skill 1. Download this `SKILL.md` file. 2. Place it in `~/.claude/commands/` (macOS/Linux) or `%USERPROFILE%\.claude\commands\` (Windows). 3. In Claude Code, run `/cutover-day-runbook`. Describe the cutover in plain English. Answer the clarifying questions. Receive a runbook ready for the bridge. ## When to use this - You're cutting over a system this weekend and the migration-planner output is too high-level for the team on the bridge at 2am. - A vendor is running the technical work and you need to brief your side's observers and rollback decision-makers. - You want a single shareable document for the runbook, comms plan, and rollback plan — no Jira-ticket hunting during the change window. - Your CAB record references a detailed runbook and you need to attach one. - You want pre-cutover-day dress-rehearsal steps and day-of go/no-go criteria on one page. ## What you'll get A single Markdown document containing: - **Pre-cutover checklist** (T-7 days through T-0) - **Named roles** (lead, comms, observer, business sponsor, rollback approver, on-call escalation) - **Bridge details** (call link, backup channel, start time, expected duration) - **Go/no-go decision points** with explicit pass criteria - **Hour-by-hour plan** (T-0 start through T+end) - **First-60-minutes minute-by-minute script** - **Comms cadence** (internal, stakeholder, external customer, social/status page) - **Rollback triggers + rollback plan** - **Point of no return** (explicit timestamp) - **Post-cutover verification checklist** (0-30 min / 30 min-2 h / 2-24 h) - **Known risks and who decides** ## Clarifying questions I will ask you 1. **What's cutting over?** (system, scope, source → target in one sentence) 2. **When?** (date, start time, timezone, expected duration) 3. **What business service depends on this?** (who feels it if something goes wrong) 4. **Is there user-facing downtime?** (yes — how long / no — but degraded for) 5. **Vendor or internal execution?** 6. **Who's on the bridge?** (roles, not names if you don't know yet) 7. **What's the go/no-go criteria at T-0?** 8. **What's your point of no return?** (after which you can't roll back) 9. **Has this been rehearsed?** (full dress rehearsal / partial / never) 10. **Primary rollback approach?** (restore snapshot, failover to standby, DNS flip back, replace hardware, revert config) 11. **Who's the named rollback approver?** (one person — not a committee) 12. **Business communications required?** (none, internal-only, external customer, regulator) ## Output template ```markdown # Cutover Runbook — — **Change reference:** CHG- · **Cutover window:** – **Bridge:** · **Backup:** **Lead:** · **Rollback approver:** ## 1. Cutover Summary ## 2. Roles on the bridge | Role | Responsibility | Who (name) | |---|---|---| | Change Lead | Runs the bridge, calls go/no-go | | | Technical Lead | Executes the steps | | | Observer | Independent verification | | | Comms Lead | Drafts and sends updates | | | Business Sponsor | Authorises scope/scheduling changes | | | Rollback Approver | Single decision-maker on rollback | | | On-Call Escalation | Reachable if rollback approver is not | | ## 3. Pre-Cutover Checklist ### T-7 days - [ ] Runbook reviewed and approved by - [ ] Dress rehearsal completed — issues logged, remediated - [ ] Snapshot / backup confirmed of all affected systems - [ ] Rollback tested end-to-end ### T-48 hours - [ ] Change notice sent to stakeholders (template in §10) - [ ] Monitoring alerts silenced for the window (if appropriate) - [ ] Out-of-hours contacts confirmed for each role - [ ] Bridge link and backup channel tested ### T-24 hours - [ ] Go/no-go meeting held and decision recorded - [ ] Final backup confirmed running and verified - [ ] Known dependencies checked (DNS TTLs lowered if needed, pending jobs paused) ### T-1 hour - [ ] Bridge open - [ ] Roll-call of all named roles - [ ] Final weather check (system health, change-freeze status, external alerts) ## 4. Go / No-Go Gate (T-0) All must be ✅ to proceed: - [ ] All roles present on bridge - [ ] Backup / snapshot timestamp < hours old - [ ] Source system healthy (no active incident) - [ ] Target system healthy (no active incident) - [ ] No external dependency in degraded state - [ ] Comms pre-approved message staged - [ ] Rollback approver confirmed reachable until **Decision:** GO / NO-GO (recorded by: ______) ## 5. Hour-by-Hour Plan | Time | Action | Owner | Expected duration | Verification | |---|---|---|---|---| | T-0:00 | Comms: "Cutover starting" | Comms | — | Channel updated | | T-0:05 | | Tech | | | | T-0:15 | | Tech | | | | T-0:30 | Go/no-go gate 2 (checkpoint) | Lead | — | — | | T-0:45 | | Tech | | | | T-1:00 | Comms: progress update | Comms | — | — | | T-1:30 | | Tech | | | | T-2:00 | Point of no return: rollback no longer possible via | Lead | — | Decision recorded | | T-2:30 | | Tech | | | | T-3:00 | Verification phase (see §8) | Observer | | Sign-off | | T-3:30 | Comms: "Cutover complete — in verification" | Comms | — | — | | T-4:00 | Hand-back to operations / close bridge | Lead | — | Bridge closed | ## 6. First-60-Minutes Script (the one you'll read from) **T+0:00** — Lead says: "Cutover is starting at