Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Generate a comprehensive pre-launch, launch day, and post-launch checklist for any product release. Use when preparing for a product launch, feature release, or major update. Produces a role-assigned, tiered checklist covering engineering readiness, marketing and comms, support, and post-launch monitoring.
.claude/skills/mohitagw15856-product-launch-checklist/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-06 | ✗→✓ | ▲ Improved | 126% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 49% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 117% | 0% |
| case-22 | ✓→✗ | ▼ Worse | 128% | 0% |
| case-14 | ✓→✓ | = Same ✓ | 66% | 0% |
Never launch without checking everything. Generate a complete, role-assigned checklist covering pre-launch readiness, launch day execution, and post-launch monitoring.
Once the checklist is approved, it can be executed: hand the items to action-runner, which previews them (dry-run, risk-rated), runs only what you approve via the connected action MCP (GitHub/Linear/Slack), and records what was done back to the brain. Typical: open an issue per checklist item in the named repo/project (🟡), and post the launch summary to Slack (🔴 — approved individually). This skill proposes; action-runner gates and runs — never silently.
Ask the user for these if not provided:
If a professional-brain (brain/) exists, use it before asking:
entities/ feature being launched and related decisions/ (scope, dates, owners).decisions/. This skill can also hand the checklist to action-runner to file the tickets — which records what was actually done back to the brain, closing the loop.Provide:
The skill generates a tiered checklist. Tier 3 launches use only the Essentials section. Tier 2 adds Marketing & Comms. Tier 1 uses all sections.
Launch Tier: 1 / 2 / 3] Launch Owner: PM Name] Engineering Lead: Name] Go/No-Go Decision By: Date and time — typically 24 hours before launch]
| Criteria | Status | Owner | |---|---|---| | All critical bugs resolved | 🟢 / 🔴 | Eng Lead | | QA sign-off complete | 🟢 / 🔴 | QA | | Rollback tested | 🟢 / 🔴 | Eng Lead | | Help centre articles live | 🟢 / 🔴 | Support | | Monitoring active | 🟢 / 🔴 | Eng Lead | | PM sign-off | 🟢 / 🔴 | PM |
Go / No-Go Decision: GO / NO-GO] Decision Owner: PM + Eng Lead jointly]
This skill ships with support files — use them when they are available:
references/launch-tiering.md — Launch Tiering: Matching Ceremony to Stakes. Apply it while producing the output; it carries the calibration and judgment calls the method summary above compresses.templates/launch-plan.md — a fill-in version of the deliverable with the quality gates inline. Offer it when the user wants to work the document themselves rather than have it generated.Score any output of this skill before handing it over; 32+ is ship-quality.
| Dimension | 0 | 5 | 10 | |---|---|---|---| | Tier calibration | No tier stated, or checklist depth obviously mismatched (full Tier 1 ceremony for a copy tweak, or a bare list for a new revenue line) | Tier stated but not justified; some sections included or dropped inconsistently with the tier | Tier stated with the reasoning (pricing, legal, blast radius — not engineering effort), sections match the tier, and any tier disagreement is resolved on record | | Ownership & timing | Items and gates owned by "the team" or nobody; no dates | Most items have owners but key gates (Go/No-Go, expansions, retro) lack a named individual or a specific time | Every checklist item, gate, and decision has one named individual and a date/time, including the Go/No-Go decision time set ~24h before launch | | Rollback & staged rollout | No rollback plan, or flag flips to 100% on day one | Rollback documented but untested; staging exists but expansion steps have no pass criteria | Rollback tested with a known revert time ("X minutes, tested on date]"), flag staged 5–10% → 50% → 100%, and each expansion gated on specific named checks | | Go/No-Go integrity | No gate, or a gate that is theatre — everything green by default, blocked work checked off | Gate exists with criteria, but blocked items are softened, statuses aspirational, or a NO-GO has no revised plan | Honest statuses (reds shown with blocker and owner), the decision follows the table even when that means slipping, and a NO-GO produces a dated re-gate — plus the retro booked at launch time |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 22,673 | 29,725 | +31% | 1 | 1 | 0% | 2,875 | 5,622 | +96% | 0 | 0 | — |
case-02 | fail→fail | 17,909 | 22,830 | +27% | 1 | 1 | 0% | 2,727 | 4,514 | +66% | 0 | 0 | — |
case-03 | fail→fail | 25,735 | 24,116 | -6% | 1 | 1 | 0% | 2,821 | 5,194 | +84% | 0 | 0 | — |
case-04 | fail→fail | 20,723 | 16,172 | -22% | 1 | 1 | 0% | 867 | 4,253 | +391% | 0 | 0 | — |
case-05 | fail→fail | 17,339 | 35,567 | +105% | 1 | 1 | 0% | 1,649 | 5,670 | +244% | 0 | 0 | — |
case-06 | fail→pass | 26,734 | 14,418 | -46% | 1 | 1 | 0% | 1,633 | 3,685 | +126% | 0 | 0 | — |
case-07 | fail→fail | 18,720 | 23,708 | +27% | 1 | 1 | 0% | 2,523 | 4,565 | +81% | 0 | 0 | — |
case-08 | fail→pass | 23,138 | 18,775 | -19% | 1 | 1 | 0% | 3,032 | 4,511 | +49% | 0 | 0 | — |
case-09 | fail→fail | 24,214 | 26,147 | +8% | 1 | 1 | 0% | 2,760 | 5,544 | +101% | 0 | 0 | — |
case-10 | fail→fail | 27,395 | 26,157 | -5% | 1 | 1 | 0% | 2,966 | 5,227 | +76% | 0 | 0 | — |
case-11 | fail→fail | 23,876 | 18,299 | -23% | 1 | 1 | 0% | 2,583 | 4,532 | +75% | 0 | 0 | — |
case-12 | fail→fail | 13,726 | 22,616 | +65% | 1 | 1 | 0% | 1,908 | 4,980 | +161% | 0 | 0 | — |
case-13 | fail→fail | 30,201 | 22,311 | -26% | 1 | 1 | 0% | 3,453 | 5,064 | +47% | 0 | 0 | — |
case-14 | pass→pass | 28,012 | 26,292 | -6% | 1 | 1 | 0% | 3,142 | 5,211 | +66% | 0 | 0 | — |
case-15 | fail→pass | 21,445 | 26,041 | +21% | 1 | 1 | 0% | 2,275 | 4,948 | +117% | 0 | 0 | — |
case-16 | fail→fail | 21,677 | 22,872 | +6% | 1 | 1 | 0% | 2,678 | 5,077 | +90% | 0 | 0 | — |
case-17 | pass→pass | 29,021 | 29,156 | +0% | 1 | 1 | 0% | 3,162 | 5,084 | +61% | 0 | 0 | — |
case-18 | fail→fail | 20,196 | 21,987 | +9% | 1 | 1 | 0% | 2,977 | 4,535 | +52% | 0 | 0 | — |
case-19 | pass→pass | 23,869 | 32,757 | +37% | 1 | 1 | 0% | 2,753 | 5,469 | +99% | 0 | 0 | — |
case-20 | pass→pass | 23,842 | 29,283 | +23% | 1 | 1 | 0% | 2,929 | 5,231 | +79% | 0 | 0 | — |
case-21 | pass→pass | 23,962 | 36,219 | +51% | 1 | 1 | 0% | 3,135 | 6,382 | +104% | 0 | 0 | — |
case-22 | pass→fail | 22,703 | 24,833 | +9% | 1 | 1 | 0% | 2,230 | 5,089 | +128% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 21 counted toward the lift figure. The other 1 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +9 percentage points is the difference between those two pass rates over the 21 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
The publisher has shipped newer versions since this run, so these numbers describe v3, not the version currently listed.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.