Claude Docs for Product Managers: 41 Claude Skills to Write the Docs the Job Runs On

What these skills are

The documents of the product job, from the first brief to the post-launch review, as installable Claude skills.

This is Claude for product documentation that starts from the reader and the decision, not the blank page: Claude drafts the document from your own sources, in your own voice, and you are the one who signs it.

The skills follow the documents in the order they come up. You set up once — a voice guide built from docs you wrote, your templates as reusable outlines, and a six-line brief that fixes the reader and the length before a word is written — then discover, shape the vision, define the work, decide, align, report up, ship and learn, run the team, and last of all cut what nobody will read.

The red line: Claude drafts from your sources, in your voice, at the length the reader will actually read. It never invents a number, a quote or a decision, and a named person signs every decision a doc records.

Each skill is one finished document you walk away with: a six-line brief, a customer interview guide, a lean canvas, a problem statement, a PRD with a summary and a detail tab, a trade-off memo with the costed no ready to send, a PR/FAQ, a roadmap narrative, a QBR, a sunset notice, an executive summary under 120 words. Each works from what you paste — the transcripts, this week's tickets, the metrics export, the figures finance gave you — and each one names its reader, holds a length budget, and ends with the decision a named person makes, by a date.

Each skill also says where the document lives: Claude Docs, a Google Doc from chat, or Claude for Word with your company template and tracked changes. When a surface is not on your plan, the skill gives you the same document as plain chat output.

Mechanically, each skill is one folder with a SKILL.md file. The 41 are grouped into ten stages of the work, and each one gives full value on its own: run one, run a stage, or work through the lot in the order the documents come up.

Download all 41 skills. One zip: ready-to-install skill zips, readable SKILL.md files, and the notes naming the source behind every method. Free, no signup. The same 41 skills are open on GitHub: polar-bear-org/claude-skills.

The 41 skills you get

The set follows the documents, in ten stages: the set-up that runs once, discovery, shaping the vision, defining the work, deciding, aligning, reporting up, shipping and learning, running the team, and the pass over what is already written.

1 · Set up once

1. Writing Style Guide

Use when: Everything Claude writes sounds like AI, and your team can tell

Output: One-page voice guide from your own docs, banned words list, length defaults per doc type

2. Doc Template Kit

Use when: AI ignores your template and invents its own headings

Output: Your PRD, memo and update templates as reusable outlines, section rules, a filled sample

3. Doc Brief

Use when: Before any doc, so it is the length people will read

Output: Purpose, reader, decision wanted, length budget, sources, what to leave out, in six lines

2 · Discover

4. Customer Interview Guide

Use when: You get 30 minutes with a customer, once

Output: Research questions, screener, story-based questions, note sheet

5. User Research Summary

Use when: Notes and transcripts pile up and nobody reads them

Output: Findings with counts and sources, what we still do not know, opportunities, top line

6. Competitive Analysis

Use when: A competitor launches and it lands on the roadmap by Monday

Output: Comparison table with dated sources, where we win and lose, what we will not copy

3 · Shape the vision

7. Opportunity Assessment

Use when: An idea arrives and needs one page before anyone commits a team

Output: One page on the standard opportunity questions, riskiest unknowns, go or not now

8. Lean Canvas

Use when: A new bet needs its business logic on one page

Output: One-page canvas with the riskiest assumptions marked untested

9. Business Model Canvas

Use when: The feature changes how the business makes money, sells or delivers

Output: The nine blocks, what changes for the business, untested assumptions

10. Value Proposition Canvas

Use when: Marketing, sales and product each describe the product differently

Output: Jobs, pains and gains from research, value map, fit notes, one positioning line

4 · Define

11. Problem Statement

Use when: Leadership saw a demo and wants it shipped before anyone asks why

Output: Who, situation, evidence, cost of the problem, success, a "how might we" line

12. Product One-Pager

Use when: A bet needs a yes before anyone writes a full spec

Output: Problem, appetite, approach, rabbit holes, no-gos, open questions on one page

13. PRD (Product Requirements Document)

Use when: A long spec was read differently by everyone

Output: Goals and non-goals, users, prioritised requirements, edge cases, metrics, length budget

14. AI Feature Spec and Eval Plan

Use when: Nobody can say how anyone will test what good means for an AI feature

Output: Model inputs and outputs, must and must-never behaviour, success criteria, test cases

15. User Stories and Acceptance Criteria

Use when: Engineers say the requirements are not detailed enough

Output: Stories, Given/When/Then criteria, INVEST check, refinement questions

5 · Decide

16. Trade-off Memo

Use when: You must write why you are not building it, then defend it

Output: The request, what it displaces, cost of delay of each, recommendation, the costed no

17. Decision Memo

Use when: Sign-off happens, then people change their minds

Output: Options with evidence, recommendation, reversibility, who decides by when, signature line

18. Working Backwards PR/FAQ

Use when: A big feature has to prove customers care before engineering starts

Output: One-page press release, external FAQ, internal FAQ with the hard questions

19. Business Case

Use when: Finance will pick apart any ROI projection

Output: Case for change, options including do nothing, ranges with named assumptions, the ask

6 · Align

20. Product Strategy

Use when: A leader says everything is high priority

Output: Diagnosis, guiding policy, coherent actions, what we will not do

21. Roadmap Narrative

Use when: A date on the roadmap becomes a promise

Output: Written Now, Next, Later with the problem behind each item, confidence, what changed

22. Product OKRs

Use when: Quarterly planning, or OKRs that read like task lists

Output: 1 to 3 objectives, outcome key results with baselines, what we stop, check-in rhythm

23. Weekly Product Update

Use when: Friday afternoons on a report people skim in seconds

Output: Decision needed at the top, what changed, risks with evidence, what is next, under 250 words

7 · Report up

24. Weekly Metrics Review

Use when: The dashboard exists and nobody writes down what it means

Output: Input and output metrics with base and period, what moved and why, actions, owner questions

25. Roadmap Review Pre-Read

Use when: The roadmap review is spent explaining instead of deciding

Output: What changed, now-next-later with confidence, what is off and why, the decision asked

26. Quarterly Business Review Doc

Use when: The quarter ends and leadership wants the review written first

Output: Goals against results, shipped against promised, lessons, next bets, the asks

27. Board Memo, Product Section

Use when: The CEO needs the product section of the board pack by Friday

Output: Progress against plan, the few metrics that matter, risks, the question for the board

8 · Ship and learn

28. Product Launch Plan

Use when: Support hears about a feature from customers

Output: Launch tier, readiness checklist by team, dates and owners, go or no-go criteria, comms list

29. Sales Enablement Brief

Use when: Sales promised it before it shipped, or cannot explain it after

Output: Who it is for, the problem in customer words, what it does not do, objections, internal FAQ

30. Release Notes

Use when: Every release, or a changelog nobody reads

Output: Customer notes grouped Added, Changed, Fixed, Removed, benefit first, plus an internal version

31. Feature Sunset Notice

Use when: You have to retire a feature customers still use

Output: Internal decision note, customer notice, dates, replacement, migration steps, support FAQ

32. Experiment Readout

Use when: An A/B test ends and is read as the answer it hoped for

Output: Hypothesis, design, sample, result with interval, guardrails, decision, lessons

33. Post-Launch Review

Use when: Customers still use the old workflow two weeks after launch

Output: Goals against actuals with sources, adoption, surprises, keep, change, stop, owners

9 · Run the team

34. Meeting Notes and Decisions

Use when: Every review someone disputes what was decided last time

Output: Decisions, actions with owners and dates, open questions, sent the same day

35. Decision Log

Use when: Decisions get reopened because nobody can find why

Output: Dated table of decisions with context, options, decider, status, link

36. Team Charter

Use when: A new team forms, or nobody knows who decides

Output: Mission, customers, scope, roles, decision rights, working agreements, review date

37. Handover Doc

Use when: You change teams, go on leave or take over mid-way

Output: State of each workstream, open decisions, risks, key roles, first two weeks

38. New PM Onboarding Doc

Use when: A new PM would otherwise learn the product from chat scrollback

Output: Product, users, strategy, metrics, who decides what, open threads, first 30 days

10 · Review and edit

39. RFC and Design Doc Review

Use when: A 30-page RFC lands, or a polished spec misses an edge case

Output: Comments by severity, edge cases, missing requirements, questions, a go, revise or discuss call

40. Executive Summary

Use when: Execs want the bottom line first and your doc buries it

Output: The answer in one line, three supporting points, the ask, under 120 words

41. Plain Language Edit

Use when: Eight paragraphs could have been one sentence

Output: Shorter version, word count before and after, AI tells removed, edits listed

How to use it

There is no compulsory sequence. You bring one document you owe somebody — the PRD due Thursday, the Friday update, the memo that has to say no to the feature the CEO wants this quarter — and open the skill that matches it. The three set-up skills are worth running first all the same: doc-writing-style-guide and doc-template-kit are saved once to a Project, and doc-brief fixes the reader, the decision and the length before anything is written, which is what keeps every later document to the length people will read.

Most of them chain. doc-interview-guide fills doc-research-summary, which feeds doc-problem-statement and doc-opportunity-assessment; doc-lean-canvas, doc-business-model-canvas and doc-value-proposition-canvas put the bet's logic on one page. doc-one-pager wins the yes that doc-prd then specs, and the PRD splits into doc-user-stories and, for a model-backed feature, doc-ai-feature-spec. doc-trade-off-memo and doc-decision-memo settle what gets built and doc-decision-log keeps the answer findable, while doc-product-strategy, doc-roadmap-narrative and doc-okrs put it where the company can see it, and doc-weekly-update, doc-metrics-review, doc-qbr and doc-board-memo carry it upward. Shipping runs from doc-launch-plan through doc-sales-enablement-brief and doc-release-notes to doc-experiment-readout and doc-post-launch-review. The last three — doc-rfc-review, doc-executive-summary and doc-plain-language-edit — work on documents that already exist, including the ones the other skills drafted.

The sources behind each method are in resources/evidence-and-sources.md inside the pack.

Setup guide

1. Download the pack. One zip: an install folder with 41 ready-to-upload skill zips, a skills folder with the same 41 skills as readable SKILL.md files, and resources/evidence-and-sources.md, which names the source behind every method.
2. Install your skills. In Claude Code, add the marketplace and install docs-for-product-managers-pack, and all 41 load at once. In Claude, turn on code execution and file creation in Settings, Capabilities, then go to Customize, Skills and upload one zip per skill from the install folder. On Team and Enterprise plans an owner enables Skills and cloud code execution first. Prefer working from files? Add the SKILL.md files to your Project knowledge instead; it works, just less cleanly.
3. Set up once, then bring one doc. Start with doc-writing-style-guide on three to five documents you wrote yourself, and doc-template-kit on your team's PRD and memo templates, both saved to a Project. After that every skill works from what you paste — the transcript, this week's tickets, the metrics export, the figures finance gave you — and writes in your voice at the length the reader will read.

Where to start

The quality bar

Every skill in the pack holds the same standard, the one we hold for the documents we write ourselves:

• One document, one method: each skill runs the real mechanics of a named method or artifact — its sections, its grid, its formula, the order of the pass — not a generic gather, analyse, recommend
• Every document names who reads it and how long it may be, and the answer or the decision sits at the top
• Every skill says when not to use it, and which skill fits instead
• The same shape every time: a fixed output template, a short done-when check before handover, and the next skill to run
• No invented numbers: a figure, a quote or a date you have not given stays a [placeholder] until you set it
• Work is reviewed, people are not: the skills assess options, requirements, risks, metrics and documents, never colleagues or individual customers
• The red line: Claude drafts from your sources, in your voice, at the length the reader will actually read; it never invents a number, a quote or a decision, and a named person signs every decision a doc records

Who made this

Polar Bear is a people ops consultancy for human-size teams (20 to 200 people). Built by ex-McKinsey founders with a dream to make AI work for People, not instead of them. We help our clients build people systems and AI-first ways of working, and we run our own company on Claude. This pack is the free, self-serve version of how we work.

The pack carries one product desk's writing. When you want your whole team working this way, AI carrying the overhead so people do the part only people can do, across hiring, management and everyday operations, that's what we build with clients.

Meet Pauline. A PRD nobody read the same way, or a doc that buries the ask? Bring the doc and the reader you need a decision from. Book a 30-minute call · Pauline on LinkedIn

Install in Claude Code

Two lines, and every skill in the pack loads at once.

/plugin marketplace add polar-bear-org/claude-skills
/plugin install docs-for-product-managers-pack@polar-bear-skills

Using Claude on the web instead? Download the zip and upload each skill from its install folder under Customize → Skills.