eol-process
Run a product sunset end to end — decide, align, plan, prepare, announce, close
Install / Use
npx skills add deanpeters/Product-Manager-Skills --skill eol-processInstalls into whichever agent you are using.
SKILL.md
Installable skill definition
Quality Score
Category
Development & EngineeringSupported Platforms
Our assessment of eol-process
eol-process scores 90/100 on our quality scale, 1288th of 4,700 Development & Engineering skills we index (top 28%).
Its SKILL.md is 21 KB long, well organised into 48 sections with 1 code example: 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 eol-process is actively maintained.
- Our last check on 2026-10-30 found the source still online.
- 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 — 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 2026-09-26. Automated pattern scan on 2026-09-26. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.
eol-process compared with similar skills
All 4 of these similar skills score higher than eol-process; compare them before choosing.
| Skill | Score | Stars | Updated | Format |
|---|---|---|---|---|
| eol-process (this skill)by deanpeters | 90 | 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 eol-process?
- Run
npx skills add deanpeters/Product-Manager-Skills --skill eol-process. The install tabs above show the steps for each supported agent. - Which AI agents does eol-process 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 eol-process 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 eol-process still maintained?
- The repository was last updated about 3 months ago, so eol-process is actively maintained.
Skill content
View source on GitHubname: eol-process argument-hint: "[product being retired, and where you are in the process]" description: "Run a product sunset end to end — decide, align, plan, prepare, announce, close. Use when you need the whole EOL process, not just one artifact." intent: >- Orchestrate a complete product end-of-life from the go/no-go decision through post-retirement review, sequencing the decision, stakeholder alignment, operational plan, internal enablement, customer announcement, and closure. Right-sized so a feature deprecation runs in a week and a regulated hardware retirement runs across quarters. Use this to lose the product without losing the customer. type: workflow theme: eol-transition best_for:
- "Running a full product sunset from decision through post-EOL review"
- "Knowing what has to happen before you can announce — and in what order"
- "Rescuing a sunset already in motion by finding which phase got skipped" scenarios:
- "We've decided to retire a product and I own the whole thing — walk me through it end to end"
- "We announced an EOL and it's going badly; what did we skip and what do we do now?" estimated_time: "30-60 min to plan; weeks to quarters to execute"
EOL Process
Purpose
Run a product retirement end to end: decide whether to do it, align the people who can stop it, build the operational plan, ready the teams who will face customers, announce it, and close it out properly. Six phases with decision points between them.
The governing goal, and the sentence worth keeping in your head the whole way through: lose the product without losing the customer. Most of what follows exists to protect the second half of that sentence.
This is an orchestration skill. It doesn't replace the artifact skills — it tells you which one to reach for, when, and what has to be true before you move on.
Input
Works best with: The product being retired and where you currently are — considering it, decided, mid-plan, or already announced and struggling.
Also useful: Scale (customers, revenue, contracts), whether a replacement exists, and who already knows.
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 process opens by establishing what's being retired and where you are in it, then routes you to the right phase. If you're mid-sunset and something is going wrong, say so — the diagnostic in "Entering Mid-Stream" finds the skipped phase.
Example invocations:
Run the full EOL process for our legacy reporting module — decision made, nothing else started.We announced a sunset three weeks ago and Support is drowning. What did we skip?
Key Concepts
The Six Phases
| # | Phase | Question it answers | Primary artifact | |---|---|---|---| | 1 | Decide | Should we retire this, and how big is this? | Readiness assessment + intensity level | | 2 | Align | Who can stop this, and what do they know that we don't? | Stakeholder sequence | | 3 | Plan | What has to happen, when, and who owns it? | Phase-gated checklist | | 4 | Prepare | Are our people ready before our customers hear? | Internal enablement pack | | 5 | Announce | What do customers hear, and when? | Customer EOL message | | 6 | Close | Did we finish, and what did we learn? | Post-EOL review |
Phase 6 is the one everyone skips. It's also the phase that makes your next sunset cheaper. Budget for it up front, because nobody volunteers for it afterward.
This Is a Route, Not a Pipeline
Every skill in this suite stands alone. None requires another to have run first. This process is a recommended route through independent stops, not a conveyor belt — which means:
- You can enter at any phase. Decision already made and defensible? Start at Phase 2.
- You can skip phases when the level justifies it. A Level 1 sunset often collapses 2 through 4 into a single afternoon.
- You can go backwards, and sometimes must. Phase 2 routinely sends you back to Phase 1 — a Legal finding or a Sales commitment can invalidate the decision. That's the process working, not failing.
- Nothing hands off a format. Carrying context between phases means telling the next skill "Level 2" and pasting what you have. There's no schema to preserve.
The order earns its keep because each phase surfaces what the next one needs. Deviate deliberately, not accidentally.
Right-Size the Whole Process
Not all EOLs play out the same. The process compresses or expands with the sunset:
| | Level 1 — Light | Level 2 — Standard | Level 3 — Heavy | |---|---|---|---| | Typical scope | Feature, internal tool, API | Commercial product, active customers | Revenue-critical, hardware, regulated | | Elapsed time | Days to weeks | 6-12 months | 12-24 months | | Phase 2 stops | 3-4 | 7-8 | 10+ | | Phase 3 gates | 2-3 phases, no gate criteria | 4-5 phases with gates | All 6 phases, gates with approvers | | Phase 4 output | Support FAQ | + Sales points, objections, escalation | + Channel brief, training | | Phase 5 message | Brief notice | Standard with phase table | Full, phased, with compliance | | Phases 2-4 | Often collapse into one sitting | Distinct, sequential | Distinct, with their own workstreams |
Level 2 is the default. Set the level in Phase 1, and let it size everything downstream. Never default to Level 3 — process theater on a small sunset teaches everyone to ignore the process on a big one.
Entering Mid-Stream
Most people find this skill in the middle of a sunset, often a troubled one. Symptoms map to the phase that got skipped:
| Symptom | Phase skipped | What to do now | |---|---|---| | "Legal just found a contract problem" | 2 (Align) | Stop. Return to Phase 1 — the decision may not survive | | "Sales says they promised this forever" | 2 (Align) | Inventory field commitments before any further comms | | "Support is drowning in tickets" | 4 (Prepare) | Ship the FAQ and escalation ladder today; backfill the rest | | "Customers say they never heard" | 5 (Announce) | Re-announce with dates; the first notice didn't land | | "Nobody knows who owns what" | 3 (Plan) | Build the checklist; unowned items are the failure | | "We shut it off and things broke" | 3 (Plan) | Downstream readers were never inventoried | | "We did it and learned nothing" | 6 (Close) | Run the review now — memory decays fast |
Facilitation Source of Truth
Use workshop-facilitation as the interaction protocol when
running any phase conversationally. Give the heads-up, take context dumps, offer numbered choices,
and let people bail to a specific phase.
Anti-Patterns (what this is NOT)
- Not a project plan. It sequences decisions and artifacts, not tasks and resources.
- Not mandatory in full. Six phases at Level 1 is the ceremony failure this suite exists to prevent.
- Not a substitute for the conversations. Phase 2 is people talking, not a document.
- Not one-directional. Going back to Phase 1 is a success condition, not a rollback.
Application
Use template.md as the one-page tracker for a sunset in flight.
This workflow orchestrates 6 phases with a decision point after each. Elapsed time runs from days at Level 1 to two years at Level 3 — set the level in Phase 1 and let it size the rest. Each phase names the skill that produces its artifact, but every one of those skills also runs standalone, so enter wherever you actually are.
Phase 1: Decide
Question: Should we retire this, and how big is this?
Activities
- Name the trigger honestly — a metric, a strategy shift, a cost complaint, or an exec remark. The trigger predicts the failure mode.
- Assess retire signals against hold signals. Two or more strong retire signals is a real case; an obligation lock or a missing landing place can outweigh all of them.
- Set the intensity level. Recommend from blast radius, then choose deliberately. This sizes every phase that follows.
- Confirm the landing place: replacement, migration, or graceful exit — and whether it's ready.
Run eol-readiness-advisor for this phase.
Outputs
- A verdict: Go, Go-with-conditions, Hold, or Harvest
- An intensity level (1, 2, or 3)
- A named landing place with a readiness status
- A list of obligations to check before announcing
Decision Point 1: Is this a Go, and is the landing place real?
- Go → proceed to Phase 2
- Go-with-conditions → proceed to Phase 2, carrying the condition as a gate on Phase 5
- Hold → stop. Set a revisit trigger. Consider whether a cheaper move (End of Sale only, a price change, a bug fix) satisfies the actual need
- Harvest → stop the investment, not the product. Set a review date
If the landing place isn't ready, you may proceed through Phases 2-4 but not Phase 5. Do not announce a transition you can't yet support.
Phase 2: Align
Question: Who can stop this, and what do they know that we don't?
Activities
- Order the stops: Legal, Finance, Sales, Marketing, CS, difficult customers, Engineering, Support — filtered by level, plus Executives, Channel, and Regulatory at Level 3.
- For each stop, prepare what you need from them and what you owe to them.
- Hold the conversations in order. Each one informs the next.
- Capture what surfaced — especially field commitments, contract terms, and hidden dependencies.
Run eol-stakeholder-sequence for this phase.
Outputs
- A completed sequence with an output per stop
- An inventory of commitments, obligations, and dependencies discovered
- A revised impact list from your most difficult customers
Decision Point 2: Did anything invalidate the decision?
This is the most important gate in the process, and the one teams treat as a formality.
- Nothing blocking → proceed to Phase 3
- A contract, regulation, or commitment blocks the timeline → return to Phase 1. Reassess with the new evidence. The date moves, the scope changes, or the verdict flips
- The landing place turns out to be weaker than believed → return to Phase 1. This is the single most common finding, and it usually arrives from the difficult-customer stop
Returning to Phase 1 here is cheap. Discovering the same thing after Phase 5 is not.
Phase 3: Plan
Question: What has to happen, when, and who owns it?
Activities
- Select the lifecycle gates in scope — and name the ones you're deliberately not using.
- Build items per phase, per functional area, at the level you set. Every item: a verb, 4-8 words, a named owning function.
- Write gate criteria with approvers for each phase transition (Level 2+).
- Cover the four things sunsets strand: data, contracts, access, money.
- Put the Phase 4 "enablement complete" date on the plan, before the Phase 5 announcement date.
Run eol-checklist for this phase.
Outputs
- A phase-gated checklist with owners and dates
- Gate criteria with named approvers
- Post-EOL actions, including the Phase 6 review, owned
- Assumptions to validate
Decision Point 3: Is every item owned, and is every date real?
- Unowned items → that's the finding. Escalate rather than papering over it
- A date you can't defend → write
TBD, orNot scheduledwith the precondition. An invented EOL date is a promise you will break in public - Enablement date not before announcement date → fix it now; in practice they collapse
Phase 4: Prepare
Question: Are our people ready before our customers hear?
Activities
- Build the support FAQ, organized by what customers actually ask
Truncated for display — 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 — from basic concepts to advanced agents, with copy-paste templates that bring immediate value.
lobehub
85.6k🤯 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 — see the Safety scan above for what the skill file itself contains.
