Launch
Coordinates a software or AI-powered product launch. Two phases: pre-launch QA (checklist against staging environment) and go-live execution (tracking, rollback, distribution). Requires FINAL positioning and an ICP card as upstream inputs. Triggers launch-readiness-review between Phase 1 and Phase 2. For hardware launches, use hardware-gtm/launch instead.From its SKILL.md
npx -y skills add 0xF4ng/aether-growth-fieldwork --skill launchAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
3 things to look at
- 28 days oldThe repository was created 28 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
- no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
- 4 stars4 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
15.9 KB, ~3.6k tokens by cl100k_base, as published. Nobody here has run it
Launch (Software)
Before starting
Confirm (ask or infer) before running:
- Positioning status — has
/positioningbeen run? Is the output DRAFT or FINAL? - ICP card — does one exist from
/icp-research? (required for audience targeting) - Launch scope — full release / beta / feature / segment launch?
- Target date — fixed or flexible? (determines if Phase 1 QA has time to run)
- Stakeholder sign-off — who needs to approve before go-live? (support, legal, exec?)
- Rollback plan — what is the fastest way to revert if something breaks post-launch?
If you detect hardware signals (BOM data, supply chain context, physical product, distributor channel, certifications), stop and redirect to hardware-gtm/launch/SKILL.md instead.
Contract
This skill guarantees:
- No launch proceeds without a Phase 1 QA record with explicit PASS/FAIL
- No launch achieves GO status without
launch-readiness-reviewpassing - Software launches require FINAL positioning before distribution begins
- Rollback option is declared for every launch — "N/A" requires written justification
Role: Launch Coordinator. You prevent the two most common launch failure modes: launching before the product is ready (wrong timing), and launching with the right product but the wrong go-to-market motion.
Inputs
Required before proceeding:
- Product name and description
- Positioning artifact (ideally with FINAL status from positioning-review)
- ICP card (required for audience targeting)
- Launch scope (global / segment / geography / beta / full launch)
- Target launch date
Pre-flight check
STEP S1 — Positioning check
IF positioning_artifact.status ≠ FINAL → BLOCK.
Return: "Software launch requires FINAL positioning.
Run /positioning and pass positioning-review before launch.
Launching without clear positioning produces copy that fails across every channel."
STEP S2 — ICP check
IF icp_card = undefined → WARN.
Return: "No ICP card found. Without an ICP card, audience targeting defaults
to 'everyone' — which means missing the best-fit users.
Strongly recommend: run /icp-research before launch.
Proceeding without ICP card: channels will be listed but without audience
precision. Revisit after launch with cohort retention data."
STEP S3 — Scope confirmation
Confirm: is this a full launch, segment launch, beta, or feature launch?
Each requires different checklist depth and distribution scope.
Phase 1 — Pre-launch QA
Run against staging/preview environment. NEVER skip this phase.
Pre-launch QA record:
Environment: [URL / build number / branch — confirm not production]
Date of QA: [Date]
Checks:
[ ] Content quality gate: run /content-quality-review on ALL primary launch copy
(homepage headline + subheadline, primary CTA, primary email, any paid ad copy)
Required: content_quality_review.status = APPROVE
IF not yet run → STOP. Run /content-quality-review first, then return here.
IF status ≠ APPROVE → STOP. Address REVISE/REJECT items before proceeding.
[ ] Load & render: primary launch page loads without blocking console errors
[ ] Tracking: UTM parameters, conversion events, analytics tags fire correctly;
privacy banners present if applicable to jurisdiction
[ ] Happy path: one full user journey completes end-to-end
(sign up → activate → core action) without error
[ ] Negative path: one invalid input returns a clear error
(not a silent failure or blank page)
[ ] Copy: final approved copy in place; no placeholder text [LOREM], [TBD], or draft language
[ ] Mobile: critical elements render correctly on mobile viewport
(test 375px wide minimum)
[ ] Accessibility: keyboard navigation works on primary CTA; alt text on images
[ ] Load time: primary page loads in < 3 seconds on simulated 4G
Result: PASS | FAIL
Blockers: [List or "none"]
QA reviewer: [Name or "self-review with a second pair of eyes"]
IF result = FAIL → STOP. Do not proceed to Phase 2.
IF blockers = critical → STOP. Address before any go-live step.
Launch-readiness-review gate
After Phase 1 QA passes, trigger the review before proceeding to Phase 2:
→ Trigger /launch-readiness-review
→ If verdict = GO → proceed to Phase 2
→ If verdict = CONDITIONAL-GO → address conditions with deadlines, then proceed
→ If verdict = NO-GO → resolve blockers before attempting launch
Phase 2 — Go-live execution
Launch log:
Objective: [What this launch was for — one sentence]
Environment: [Production URL / feature flag name / campaign ID]
Target audience: [ICP segment name from ICP card]
Go-live steps:
1. [Step description] — PASS/FAIL
2. [Step description] — PASS/FAIL
3. [Step description] — PASS/FAIL
...
Verification: [Which check confirmed the launch is live and working correctly]
Example: "Trial signup completed successfully from an incognito browser.
Conversion event fired in analytics. No error in logs."
Rollback option:
[How to undo this launch if something goes wrong post-launch]
Example: "Disable feature flag [X] to revert to old flow. Takes effect immediately."
"N/A" is acceptable only with justification: [explain why rollback is not needed]
User-visible impact: [What changed for users from their perspective]
Post-launch monitoring:
[ ] Conversion events firing for first real users (monitor first 30 minutes)
[ ] Error rate not elevated vs baseline (check in first 2 hours)
[ ] At least one successful end-to-end conversion by a real user
Distribution plan
After confirming go-live, distribution is the launch:
Distribution plan:
Primary channel: [Most important surface for this launch — where Core ICP is]
Secondary channels: [2-3 additional surfaces]
Assets ready: [Confirm: copy, images, links, UTMs are all in place]
Publication time: [Optimal time for primary channel]
HN note: if this is a Show HN launch, write the first comment before posting.
LinkedIn note: first line works as standalone in preview truncation.
Email note: subject line is the most important asset; test variant if possible.
Launch asset checklist and dependency order
Required structured output when launch date is set. Produce this section before Distribution plan is finalized — dependency order determines which assets can be produced in parallel and which cannot.
Dependency enforcement rule
Assets depend on each other. The agent enforces ordering and does not assume parallel production is safe. Each asset row declares what must exist before production begins. Do not mark an asset "in progress" if its upstream dependency has not been approved.
Pre-checklist gate:
BLOCK: IF launch date is set AND checklist has >2 assets with no assigned owner →
Return: "Launch checklist has unassigned assets. Assign owners before continuing
launch planning. Unowned assets at T-2d are the single most common cause of
day-of scramble and copy regression."
WARN: IF Demo video and Landing page are assigned to different owners AND no
dependency lock-date has been agreed →
Return: "Demo video and landing page are being produced in parallel — verify
landing page does not depend on final video content that is not yet locked.
Agree on video content lock date before landing page copy is written."
Core launch assets
| Asset | Format | Must exist before this | Typical lead time before launch |
|---|---|---|---|
| Demo video or product GIF | 60–90s video / animated GIF | Nothing — create first | T-21d |
| Landing page | Web | Demo video or screenshot (locked content) | T-14d |
| Docs / changelog entry | Markdown | Feature complete (engineering sign-off) | T-7d |
| Announcement copy — short (2–3 variants) | 280 chars + subject line | Landing page headline approved | T-7d |
| Social posts (3–5 variants per platform) | Platform copy | Announcement copy approved | T-3d |
| Launch email — existing customers | HTML email | Announcement copy approved | T-3d |
| Launch email — prospects | HTML email | Landing page live in staging | T-3d |
| HN / Product Hunt post draft | Text post | All above approved and stable | T-2d |
| Internal enablement brief | 1-page PDF or doc | All customer-facing copy approved | T-1d |
Checklist output format
## Launch asset checklist
Launch date: [Date]
Checklist produced: [Date]
| Asset | Owner | Status | Dependency met? | Due date |
|---|---|---|---|---|
| Demo video or product GIF | [Owner] | [ ] Not started / [ ] In progress / [✓] Approved | — (first asset) | T-21d |
| Landing page | [Owner] | [ ] Not started / [ ] In progress / [✓] Approved | [ ] Demo video locked? | T-14d |
| Docs / changelog entry | [Owner] | [ ] Not started / [ ] In progress / [✓] Approved | [ ] Feature complete? | T-7d |
| Announcement copy (2–3 variants) | [Owner] | [ ] Not started / [ ] In progress / [✓] Approved | [ ] Landing page headline approved? | T-7d |
| Social posts (3–5 variants/platform) | [Owner] | [ ] Not started / [ ] In progress / [✓] Approved | [ ] Announcement copy approved? | T-3d |
| Launch email — existing customers | [Owner] | [ ] Not started / [ ] In progress / [✓] Approved | [ ] Announcement copy approved? | T-3d |
| Launch email — prospects | [Owner] | [ ] Not started / [ ] In progress / [✓] Approved | [ ] Landing page live in staging? | T-3d |
| HN / Product Hunt post draft | [Owner] | [ ] Not started / [ ] In progress / [✓] Approved | [ ] All above approved? | T-2d |
| Internal enablement brief | [Owner] | [ ] Not started / [ ] In progress / [✓] Approved | [ ] All customer-facing copy approved? | T-1d |
Asset owner coverage: [N] of [Total] assets have assigned owners
Unblocked assets (no unmet dependency): [List]
Blocked assets (dependency not yet met): [List]
Output format
## Launch Brief
**Product:** [Name]
**Launch scope:** [Full / Beta / Segment / Feature]
**ICP segment:** [Name]
**Target date:** [Date]
### Upstream status
Positioning artifact: [FINAL / DRAFT / Missing]
ICP card: [Present / Missing]
### Phase 1 — Pre-launch QA
QA environment: [URL / system identifier]
Result: [PASS / FAIL]
Blockers: [List or "none"]
### Launch-readiness-review
Status: [Pending / Triggered / GO / CONDITIONAL-GO / NO-GO]
### Phase 2 — Go-live execution
[Complete after review passes]
Objective: ...
Steps: ...
Verification: ...
Rollback option: ...
User-visible impact: ...
### Distribution plan
Primary channel: ...
Publication time: ...
### Launch asset checklist
[Complete when launch date is set — see "Launch asset checklist and dependency order" section]
Asset owner coverage: [N] of [Total] assets assigned
Unblocked: [List or "all clear"]
Blocked: [List or "none"]
Brain reads / writes
If a companion brain repo is connected:
Before starting:
- Read
knowledge/icp-map.md— confirm which ICP segment this launch targets - Read
channels/channel-history.md— check which channels have performed for this product type before - Read
experiments/experiment-log.md— check if any messaging experiment results inform copy selection
Brain write (after GO verdict):
- Write launch decision log to
decisions/including: launch date, scope, distribution plan, and first 48-hour monitoring results - Update
channels/channel-history.mdwith which channels were used and initial result signal
Brain not connected: proceed normally; note in output that launch learnings will not be persisted.
Anti-patterns
| Anti-pattern | Why it fails | Fix |
|---|---|---|
| Launching without FINAL positioning | Copy becomes generic; every channel says something slightly different | Block until positioning-review passes |
| Skipping Phase 1 QA "because we know it works" | UAT catches what you assume away; tracking bugs, copy regressions, mobile breakages are invisible until QA runs | Never skip Phase 1 |
| No rollback plan defined | Post-launch issue becomes a crisis; no one knows who makes the call to revert | Define rollback before go-live |
| Distributing without confirmed tracking | Conversion events that don't fire create permanent data gaps | Confirm UTMs and events fire in staging before Phase 2 |
| "Soft launch" without audience definition | Soft launches are real launches; without ICP targeting, distribution is inefficient | Always define the audience even for beta / limited launches |
| Running /launch without /content-quality-review on copy | Launch copy regressions ship silently | Gate Phase 2 on APPROVE from content-quality-review |
| Producing all launch assets simultaneously without dependency order | When upstream assets change (video content shifts, headline revs), every downstream asset that copied from them requires rework; parallel production without declared dependencies creates cascading revision loops | Declare the dependency order first; lock upstream assets before starting dependent production |
Benchmarks (2025–2026)
| Benchmark | Value | Notes |
|---|---|---|
| Show HN launch typical upvote range | 100–500 upvotes | Established OSS/dev-tools; requires strong technical first comment |
| Email launch open rate (warm list, developer audience) | 35–50% | Plain text outperforms HTML by 15–20% for developer audiences |
| Trial conversion from launch email (dev tools) | 8–18% | Drops to 2–5% without ICP segmentation |
| Product Hunt Day 1 ranking | Top 5 in category | Requires pre-launch coordination; quality gallery; maker responses |
| Time from launch event to first conversion signal | Under 2 hours for instrumented products | If no signal within 2 hours: check tracking first |
| Software launch copy approval rate (content-quality-review) | Target: APPROVE on first submission | Average: 60% first-pass approval on public launch copy |
| Phase 1 QA time for feature launch | 1–2 hours | Full product launch: 4–8 hours for thorough QA |
Related skills
| Skill | When to use |
|---|---|
pmm/positioning/SKILL.md | Before launch: FINAL positioning required |
pmm/icp-research/SKILL.md | Before launch: ICP card required for audience targeting |
pmm/launch-readiness-review/SKILL.md | Between Phase 1 and Phase 2: GO verdict required |
pmm/content-review/SKILL.md | On launch copy: APPROVE required before Phase 2 |
hardware-gtm/launch/SKILL.md | If hardware signals are detected: use this skill instead |
foundations/leader/composite-launch/SKILL.md | Full orchestration: runs ICP → positioning → launch as a single chain |
Validation criteria
- Pre-flight check completed (positioning FINAL, ICP card present or documented as missing)
- Phase 1 QA completed with explicit PASS/FAIL and blockers listed
- Launch-readiness-review triggered before go-live
- Phase 2 log includes verification and rollback option
- Distribution plan specified
- When launch date is set: launch asset checklist produced with dependency order, assigned owners (or unassigned flagged as a blocker), and due dates relative to launch date
References & Sources
Tier 1:
- Product Marketing Alliance 5-phase framework (2026): "Get Set" phase — QA checklist structure
- growth-launch-checklist (growth-skills v1.0, score 7.5/10): two-phase launch structure, software track checklist
What ships with it
Read from the repository
Just SKILL.md. No reference files, no scripts.