Launch Checklist

Launch Checklist

Trigger

Activate on "launch checklist", "pre-launch check", "ready to launch?", "launch prep".

Behavior

Step 1: Get Context

Ask:

1. What are you launching?
2. Launch date
3. How many users affected?
4. Risk level (high/medium/low)

Step 2: Generate Checklist

Pre-Launch

• [ ] Feature complete and QA'd
• [ ] Rollback plan documented and tested
• [ ] Success metrics defined with baselines
• [ ] Monitoring dashboards set up
• [ ] Support team briefed with FAQ
• [ ] Release notes drafted
• [ ] Feature flag configured (if gradual rollout)

Launch Day

• [ ] Deploy during low-traffic window
• [ ] Verify metrics are tracking
• [ ] Monitor error rates for 2 hours
• [ ] Send internal launch comms
• [ ] Publish release notes

Post-Launch (48 hours)

• [ ] Check primary metric vs baseline
• [ ] Check guardrail metrics
• [ ] Review support tickets for new issues
• [ ] Share initial results with team

Post-Launch (1 week)

• [ ] Full metrics review
• [ ] Ship or iterate decision
• [ ] Retrospective if needed

Risk Scaling

Scale the checklist to risk level:

| Risk Level | When | What to Add | |-----------|------|-------------| | Low | Config change, copy update, small UI tweak | Minimal — deploy, monitor for 30 min, done | | Medium | New feature behind flag, migration with rollback | Full checklist above. 2-hour monitoring window | | High | Pricing change, auth flow, data migration, public-facing API | Everything above PLUS: staged rollout (1% → 10% → 50% → 100%), war room channel, exec notification, 24-hour monitoring |

Example: Good vs Bad Launch Checklist Items

Bad: "Make sure everything works" Good: "Verify signup flow completes end-to-end in staging (email → verification → first dashboard load). Test with Gmail, Outlook, and corporate SSO."

Bad: "Tell the team" Good: "Post in #eng-launches: what shipped, who's on-call, rollback command, link to monitoring dashboard."

Bad: "Watch metrics" Good: "Monitor error rate in Datadog for 2 hours post-deploy. Baseline: 0.3%. Alert threshold: 0.5%. Rollback threshold: 1.0%."

Rules

• Higher risk = more checklist items. Scale accordingly.
• Every item needs an owner and a deadline. Not "soon" — a time.
• Never launch on Friday afternoons
• If rollback takes more than 5 minutes, add more testing time
• Every item must be verifiable. "Done" means someone can confirm it, not just feel it.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: aakashg
• Source: aakashg/pm-claude-code-setup
• License: MIT
• Homepage: https://www.news.aakashg.com/p/pm-os

Install and usage instructions live in the source repository linked above.