launch-plan
Build a launch plan sized by tier: owners, dates, go/no-go criteria, and a readiness scorecard
Install / Use
npx skills add deanpeters/Product-Manager-Skills --skill launch-planInstalls into whichever agent you are using.
SKILL.md
Installable skill definition
Quality Score
Category
Development & EngineeringSupported Platforms
Tags
Our assessment of launch-plan
launch-plan scores 86/100 on our quality scale, 2325th of 4,698 Development & Engineering skills we index (top 50%).
Its SKILL.md is 22 KB long, well organised into 40 sections and no code examples: a thorough specification that gives an agent plenty to work with.
With 7,063 GitHub stars, it is one of the more widely adopted skills in the catalogue.
Maintenance, license and trust
- The repository was last updated about 3 months ago, so launch-plan is actively maintained.
- No license is declared. By default that means all rights are reserved: you can read it, but reusing or redistributing it is not clearly permitted. Ask the author before building on it commercially.
- Its trust signals score 88/100, with 1 caution from licensing, adoption, age or documentation. These come from repository metadata, not a code audit โ read the skill file before letting an agent act on it.
Safety scan
No issues foundOur scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful.
AI review by kimi-k2.7-code on 2026-11-16. Automated pattern scan on 2026-11-16. It catches known dangerous patterns, not every risk โ read a skill before letting an agent act on it.
launch-plan compared with similar skills
All 4 of these similar skills score higher than launch-plan; compare them before choosing.
| Skill | Score | Stars | Updated | Format |
|---|---|---|---|---|
| launch-plan (this skill)by deanpeters | 86 | 7.1k | 3mo ago | SKILL.md |
| taste-skillby Leonxlnx | 100 | 116.8k | 8d ago | CLAUDE.md |
| ai-job-searchby ai-job-search-dev | 100 | 72.1k | today | CLAUDE.md |
| claude-howtoby luongnv89 | 100 | 47.3k | 2d ago | CLAUDE.md |
| lobehubby lobehub | 100 | 85.6k | today | MCP Server |
Frequently asked questions
- How do I install launch-plan?
- Run
npx skills add deanpeters/Product-Manager-Skills --skill launch-plan. The install tabs above show the steps for each supported agent. - Which AI agents does launch-plan work with?
- It is written for Zed, as a SKILL.md file. Other agents that read the same format can often use it too.
- Is launch-plan safe to use?
- Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful. It declares no license and scores 88/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
- Is launch-plan still maintained?
- The repository was last updated about 3 months ago, so launch-plan is actively maintained.
Skill content
View source on GitHubname: launch-plan argument-hint: "[what is launching, and the date that drives it]" description: "Build a launch plan sized by tier: owners, dates, go/no-go criteria, and a readiness scorecard. Use when a launch needs every function ready at once and nobody has defined 'ready'." intent: >- Make a launch a coordinated business event instead of a ship date, so every function that touches the customer is ready, the effort fits the size of the launch, and the go/no-go call is made on criteria the team agreed before anyone saw the status. Right-sized so a minor variant gets a short punch list and a platform launch gets the full cross-functional plan. type: workflow theme: strategy-positioning best_for:
- "Planning a launch across Sales, Support or Service, Marketing, Legal, and Operations"
- "Deciding how much launch effort an offering deserves before the work starts"
- "Writing go/no-go criteria before the gate, not after the status is known"
- "Reviewing a launch plan for the function everyone forgot" scenarios:
- "We launch the route optimizer in six weeks and Support has not seen it yet. Build the plan"
- "Is this a Tier 1 or Tier 2 launch, and what does each function owe us?"
- "Review our launch plan. I think we forgot the distributors and the spare parts"
- "Help me write go/no-go criteria for a hardware launch with a trade-fair date" estimated_time: "30-60 min to plan; weeks to execute"
Launch Plan
Purpose
Build the plan that makes a launch land: a checklist of what each function must have ready, sized by launch tier, with a named owner, a date, and a status on every row; go/no-go criteria agreed before the gate; and a one-page Launch Readiness Scorecard the launch forum uses to decide.
The sentence to keep in your head: a launch is a business event, not a ship date. Engineering finishing the product is one row. A customer experiences a launch as finding out about it, getting a price, buying it, receiving it, using it, and getting help when it breaks. Each of those belongs to a different function, and the customer sees whichever one is not ready.
The PM orchestrates. The PM does not own every row. This is influence work: make the gaps visible early enough that the people who own them can close them.
Input
Works best with: What is launching and the date that drives it (a conference, a contract, a season, a budget year).
Also useful: The positioning statement, the price decision, the channels you sell through (direct, partners, distributors, marketplaces), the last launch's lessons learned, and who will make the go/no-go call.
Anything supplied with the invocation itself (text after the skill name, a pasted context dump, or
an appended ARGUMENTS: line) counts as answers already given. Use it and skip whatever it covers;
don't re-ask.
Arriving empty-handed? That works too. The skill opens by asking what is launching and what
fixes the date, proposes a tier from your answers, and builds from there. Anything you can't supply
becomes a labeled [NEEDS SOURCE] blank instead of an invented one.
Example invocations:
Launch plan for our route optimizer. GA is June 12, Sales and Support are not briefed.Review our launch checklist for the NFA-500 controller. Trade fair is April 21. I think we forgot distributor price files.
Key Concepts
Right-Size the Launch (the Tier Dial)
Not every launch deserves the same effort. A tier is an agreed size class that decides how many rows apply and how deep each goes.
| | Tier 1: Full | Tier 2: Standard | Tier 3: Light | |---|---|---|---| | Typical launch | New platform or generation, new market, new business model, large revenue at stake, new regulatory exposure | Significant new variant, major feature or option, adjacent segment on existing channels | Minor variant, small feature, cost-down invisible to customers | | Functions engaged | All | Most, lighter depth | Price list, release note, internal notice | | Gates | Formal go/no-go (often two: order intake, first shipment or GA) | One gate | None unless compliance is touched | | Output feel | Cross-functional playbook | Working checklist with owners | Punch list |
Score the tier on six criteria, each low / medium / high with a one-line reason: revenue and margin at stake, newness to the customer, newness to the company (new channel, model, or production process), regulatory and safety exposure, installed-base impact (replaces, cannibalizes, or must coexist with what customers own), and channel impact (do partners need new skills, stock, or contracts).
Tier 2 is the default. Never default to Tier 1. Recommend a tier, say why in one line, and let the user move it in either direction. If they go lighter, name what drops out so the choice is informed. If the scores split (low revenue, high regulatory exposure) say so: a Tier 2 launch can carry a Tier 1 compliance workload. Watch for tier inflation (everything is Tier 1, so the checklist becomes noise) and tier deflation (a Tier 1 change labeled Tier 3 to dodge the work).
Readiness by Function
Organize the checklist by function, because that is how ownership works in a matrix organization. These are the functions that most often fail:
- Sales and channel. Briefed and trained before customers hear; partners told early enough to plan stock, training, and their own messaging; quoting and configuration tools updated; demo environments or demo units ready.
- Support, Service, and spare parts. Trained before the first customer; runbooks, diagnostic tools, and escalation paths released; capacity planned for the launch wave; for hardware, spare parts defined, priced, and stocked where the field needs them.
- Operations, supply chain, and production. Capacity and lead times confirmed; ramp sized to the realistic forecast, not the hopeful one; first-article or first-release quality passed.
- Pricing and commercial. Price, discounts, and packaging approved; price lists, billing, and quoting entries created; partner price files sent in time for their cycles.
- Legal, compliance, and certification. Terms, privacy, and security reviews done; for hardware, conformity assessment and country approvals planned. A missing certificate stops shipments no matter how ready everything else is.
- Documentation and training. Guides, safety instructions, and customer training in every language and market you sell into.
- Marketing. Messaging from the positioning statement, web page, data sheets, demos, press, event material.
- Internal enablement. Order handling, customer success, finance, and legal know what changes for them. FAQs and escalation contacts exist.
Four Dates, Not One
Software launches often collapse these into one day. Industrial launches cannot. Name all four, and make the gap between each a deliberate choice:
- Announce. Customers and the market hear.
- Order opens. Quotes and orders accepted, delivery promised.
- First delivery. The product reaches a customer (first shipment, or first tenant live).
- Full support. Service, training, and spare parts available everywhere you sell.
Announcing long before you can deliver stalls sales of the current product while customers wait. Opening orders before capacity is confirmed breaks delivery promises. Delivering before support is ready turns the first customers into the worst reference stories.
Go/No-Go Criteria Are Agreed Before the Gate
A gate is a scheduled decision. Its criteria must be written down, with owners and thresholds, before anyone sees the status. Otherwise the team looks at a mixed picture and talks itself ready.
- Must-pass: if one fails, the launch does not proceed (support trained in launch regions; signed declaration of conformity; security review closed).
- Watch: a failure needs a mitigation and an owner but does not stop the launch (translated guide for a secondary market).
The PM prepares the gate. The forum with decision rights makes the call. This skill never makes the go/no-go decision.
The Launch Readiness Scorecard
The checklist has dozens of rows. The forum needs one page: a row per function, a column per readiness criterion, with status, owner, and evidence.
| Function | Readiness criteria | Status | Owner | Evidence | |---|---|---|---|---| | Support | Agents trained, runbook live, escalation path tested | Ready / At risk / Not ready | Named person | Training roster, runbook link |
Three rules keep it honest:
- Status needs evidence. Ready means the evidence in the criterion exists. An owner's optimism is "At risk" until it does.
- Define the statuses before the first review. What makes a row Ready, At risk, Not ready.
- Must-pass rows are marked. One must-pass "Not ready" means the scorecard says no-go, however green the rest looks.
RACI for the work, DACI for the decision. Give each function row a RACI (Responsible, Accountable, Consulted, Informed). Give each gate a DACI: the Driver (usually the PM) prepares it, one Approver decides, Contributors supply status and evidence, the Informed hear the outcome. A go/no-go with two approvers is not a decision.
Owners Are People, Not Departments
"Marketing" is not an owner. Every row needs a named person (or role, if names are confidential) who has agreed to the date. A row the owner has not agreed to is a wish. Mark it At risk.
The Premortem: Imagine the Launch Already Failed
A checklist finds the rows a team forgot. It does not find the failure nobody wants to say out loud. Gary Klein's premortem does: imagine the launch has already failed and explain why. Treating failure as a fact that already happened makes people name more and sharper causes than "what could go wrong?", and it gives quiet doubters permission to speak.
Run it once, after the checklist and criteria are drafted, before the gate:
- Set the scene. "It is six months after launch, and it failed. What happened?"
- Write failure stories alone first. Each function owner writes what failure looks like from where they sit. Writing alone keeps the most senior voice from setting the story.
- List the causes. Actions, inactions, assumptions, dependencies, partners.
- Rank by likelihood and impact. Keep the few drivers that matter.
- Convert every driver into the plan. Each becomes a readiness criterion (must-pass or watch) or a risk with an owner, a trigger, and a response. A story that changes nothing was not finished.
The premortem is a team exercise. AI can prompt it and propose typical failure stories as labeled suggestions. It cannot know why your launch would fail.
Pre-Approval Check: Will This Plan Hold?
Before the forum approves, test it as an approver would:
- Owners: separate owners for delivery, adoption, operation, and measurement. No risk owned by "the team."
- Dependencies: the critical path is named, including outside parties and capacity assumptions.
- Risks: each has a cause, trigger, consequence, response, owner, and remaining exposure.
- Adoption: someone owns whether sales, partners, and customers actually change behavior, not just whether they were trained.
- Evidence: milestones produce evidence (a signed record, a training roster), not just dates.
- Rollback: what pauses orders, what happens to units already delivered, and who decides.
Match the depth to the tier and to how reversible the step is.
Facilitation Source of Truth
Use workshop-facilitation as the interaction protocol when
running this conversationally: heads-up, context dump, numbered choices, one question per turn.
Default entry mode: Guided ยท set in pm-skills.config.yaml (library-wide default or a per-skill override); the user's own choice in the s
Truncated for display โ read the full file on GitHub.
Related Skills
taste-skill
116.8kAgent skills that give coding agents taste: art direction, reference images and a browser check instead of generic AI pages
ai-job-search
72.1kThe job search that runs on your machine. AI job application framework built on Claude Code: evaluate postings, tailor CVs, write cover letters, prep interviews. Fork it and own it.
claude-howto
47.3kA visual, example-driven guide to Claude Code โ from basic concepts to advanced agents, with copy-paste templates that bring immediate value.
lobehub
85.6k๐คฏ LobeHub - Super-Agents on Your Team. Add super-agents to the chats your team already uses.They take on work by themselves and get better every week.
Languages
Trust signals
From repository metadata: license, adoption, age and documentation. Not a code audit โ see the Safety scan above for what the skill file itself contains.
