Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Internal and external launch coordination playbook covering pre-launch, launch day, and post-launch with run-of-show, comms, RACI, rollback, and retro. Use to coordinate a GA launch, major release, or re-launch.
.claude/skills/borghei-launch-playbook/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 50% | 0% |
| case-14 | ✗→✓ | ▲ Improved | -16% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 248% | 0% |
| case-19 | ✗→✓ | ▲ Improved | 41% | 0% |
| case-20 | ✗→✓ | ▲ Improved | 81% | 0% |
A complete launch coordination playbook for software products and features. It covers the three windows that matter -- pre-launch (T-30 to T-1), launch day (T-0), and post-launch (T+1 to T+30) -- and produces five concrete artifacts: a run-of-show, an internal comms plan, an external comms checklist, a rollback plan, and a post-launch retro template.
Most failed launches are not failed builds; they are failed coordination. Engineering ships on time, but support has not been trained, sales does not have collateral, the changelog is wrong, the rollback path was never tested, and the executive sponsor hears about a customer complaint before hearing about the launch. This playbook prevents those failures by assigning every owner and every artifact before T-30.
beta-program/) and needs coordinated launch.When NOT to use: continuous-delivery bug fixes and minor improvements (use release-notes/ only); internal-only changes with no external comms (use a release runbook); product sunsets and deprecations (use eol-communication/).
Before building the launch plan, confirm these inputs. If any is unknown or vague, ASK — do not assume:
Stop rule: ask only the 2-3 that most change the output. If the user says "just draft it," proceed and list your assumptions at the top of the artifact.
Full timeline tables, RACI matrix, launch-type rules, and the 10-step workflow are in references/launch-timeline-and-execution.md.
references/launch-timeline-and-execution.md — read this when planning/running the launch: three-launches framework, full T-30→T+30 timeline, launch-type selection, RACI matrix, 10-step workflow, troubleshooting, and success criteria.references/launch-coordination-guide.md — read this for the reasoning behind the playbook: three-launches sequencing, RACI worked examples, big-bang vs progressive decision rules, and anti-patterns.references/red-flags.md — read this before signing off any launch artifact: common failure modes with bad/good examples and fixes.assets/launch-run-of-show.md — hourly run-of-show template for launch day.assets/internal-comms-plan.md — pre-launch, launch-day, and post-launch internal comms templates.assets/external-comms-checklist.md — T-30 to T-0 checklist for press, blog, social, email, sales enablement, support training, and legal review.assets/rollback-plan.md — rollback decision matrix, drill checklist, and execution runbook template.assets/post-launch-retro.md — 7-day and 30-day retrospective template.In Scope: GA launch coordination across Engineering, Product, PMM, Sales, Support, Legal; run-of-show, rollback plan, comms plans, post-launch retrospective; big bang, progressive rollout, and dark launch patterns; T-30 to T+30 timeline; RACI assignment for cross-functional launches.
Out of Scope: Beta program execution (beta-program/); release notes content generation (release-notes/); end-of-life / sunset communication (eol-communication/); incident response runbooks (delivery-manager/ and engineering on-call); marketing campaign performance analysis beyond the 30-day window (handoff to demand gen).
Important Caveats: A launch is not a moment; it is a 60-day window (T-30 to T+30) -- staff accordingly. The single most common cause of failed launches is a missing RACI assignment, not a missing artifact: if you can name the artifact but not the owner, the artifact will not exist. Big-bang launches concentrate risk on one date; default to progressive rollout unless marketing leverage justifies the risk.
| Integration | Direction | What Flows | |-------------|-----------|------------| | beta-program/ | Receives from | Greenlit beta hands off testimonials, known issues, and pricing decision | | create-prd/ | Receives from | PRD provides positioning, segments, and v1 scope | | release-notes/ | Feeds into | Launch produces the GA release notes | | daci-framework/ | Uses | Major launch decisions (go/no-go, rollback) use DACI | | eol-communication/ | Complementary | Launches that replace older features trigger sunset comms | | delivery-manager/ | Uses | Deployment, on-call rota, and rollback execution | | senior-pm/ | Reports to | Exec sponsor receives T-1, T+1, T+7, T+30 updates | | summarize-meeting/ | Feeds into | War-room standups and retro produce structured summaries | | status-update-generator/ | Feeds into | Post-launch metrics roll up into weekly status |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 34,234 | 39,226 | +15% | 1 | 1 | 0% | 5,290 | 7,924 | +50% | 0 | 0 | — |
case-02 | pass→pass | 36,979 | 37,764 | +2% | 1 | 1 | 0% | 5,678 | 7,563 | +33% | 0 | 0 | — |
case-03 | pass→pass | 34,029 | 38,253 | +12% | 1 | 1 | 0% | 5,212 | 7,912 | +52% | 0 | 0 | — |
case-04 | pass→pass | 11,941 | 9,932 | -17% | 1 | 1 | 0% | 1,758 | 3,095 | +76% | 0 | 0 | — |
case-05 | pass→fail | 15,640 | 15,373 | -2% | 1 | 1 | 0% | 2,245 | 3,920 | +75% | 0 | 0 | — |
case-06 | pass→pass | 13,296 | 10,556 | -21% | 1 | 1 | 0% | 1,857 | 3,313 | +78% | 0 | 0 | — |
case-07 | pass→pass | 15,586 | 9,172 | -41% | 1 | 1 | 0% | 2,216 | 3,063 | +38% | 0 | 0 | — |
case-08 | pass→pass | 10,707 | 12,488 | +17% | 1 | 1 | 0% | 1,588 | 3,603 | +127% | 0 | 0 | — |
case-09 | pass→pass | 14,522 | 14,042 | -3% | 1 | 1 | 0% | 2,153 | 3,787 | +76% | 0 | 0 | — |
case-10 | pass→pass | 13,592 | 14,096 | +4% | 1 | 1 | 0% | 2,049 | 3,945 | +93% | 0 | 0 | — |
case-11 | fail→fail | 22,545 | 36,606 | +62% | 1 | 1 | 0% | 3,620 | 7,527 | +108% | 0 | 0 | — |
case-12 | fail→fail | 6,992 | 6,178 | -12% | 1 | 1 | 0% | 1,036 | 2,637 | +155% | 0 | 0 | — |
case-13 | fail→fail | 17,911 | 29,458 | +64% | 1 | 1 | 0% | 2,934 | 6,495 | +121% | 0 | 0 | — |
case-14 | fail→pass | 22,135 | 6,887 | -69% | 1 | 1 | 0% | 3,241 | 2,732 | -16% | 0 | 0 | — |
case-15 | fail→pass | 12,236 | 30,574 | +150% | 1 | 1 | 0% | 1,885 | 6,556 | +248% | 0 | 0 | — |
case-16 | pass→pass | 15,208 | 11,037 | -27% | 1 | 1 | 0% | 2,275 | 3,434 | +51% | 0 | 0 | — |
case-17 | pass→pass | 9,898 | 12,493 | +26% | 1 | 1 | 0% | 1,431 | 3,512 | +145% | 0 | 0 | — |
case-18 | pass→pass | 7,335 | 4,239 | -42% | 1 | 1 | 0% | 979 | 2,241 | +129% | 0 | 0 | — |
case-19 | fail→pass | 20,114 | 14,195 | -29% | 1 | 1 | 0% | 2,805 | 3,963 | +41% | 0 | 0 | — |
case-20 | fail→pass | 12,334 | 10,664 | -14% | 1 | 1 | 0% | 1,755 | 3,185 | +81% | 0 | 0 | — |
case-21 | pass→pass | 5,689 | 5,042 | -11% | 1 | 1 | 0% | 817 | 2,343 | +187% | 0 | 0 | — |
case-22 | pass→pass | 11,302 | 5,564 | -51% | 1 | 1 | 0% | 1,706 | 2,485 | +46% | 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. The headline lift of +18 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
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.