agentleFS
lukesw55lukesw55/ai-pm-toolkit/skills/pm-phase-define/SKILL.md

pm-phase-define

Skill1 starsChanged 42 days agoMITView on GitHub

Lean Double Diamond **Define phase** skill for PMs. Invoke when discovery evidence exists but the team has not yet chosen what to build, how to measure success, how to sequence bets, or how to justify the investment. Trigger on requests for product vision/strategy, North Star metrics, KPI trees, opportunity sizing (TAM/SAM/SOM, bottoms-up), business cases, PRFAQs, pricing & packaging, prioritisation (RICE/WSJF/Kano/cost-of-delay), roadmap narratives, or decision memos (DACI/ADR). Also trigger on "o que priorizar?", "construa árvore de KPI", "write the annual plan", "choose what not to do", "build a scorecard", "justify this investment". Outputs strategic choices that are explicit, non-goals that are named, and measurement that shapes behaviour.

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Write docs that make a recommendation explicit, clarify trade-offs, and enable asynchronous decision-making.

    line 115
  2. Distinctive:

    every roadmap initiative traces to a strategic pillar

    line 159

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

PM Phase — Define

Double Diamond Phase 2 of 4. Turns discovery evidence into explicit choices and a measurement architecture that makes the choices legible and auditable.

Progressive loading

Load this SKILL.md first. For large or specialized tasks, use references/progressive-loading.md to choose the narrowest supporting reference before reading more.

When to use this skill

Invoke when:

  • discovery has produced evidence but the team has not committed to a specific problem/metric/sequence
  • a roadmap exists without clear rationale ("why this, why now, why not X?")
  • metrics are proliferating but success is ambiguous
  • an initiative needs a credible size or business case before funding
  • leadership asks "what's our strategy/North Star/priority order?"
  • pricing/packaging decisions are being discussed
  • multiple stakeholders are pulling in different directions and a crisp recommendation memo is needed
  • quarterly or annual planning is starting

Skip to pm-phase-develop if the wedge is already defined and the work is about execution. Go back to pm-phase-discover if the underlying problem is still fuzzy.

Prime directive

Define is choices + measurement. A strategy that avoids trade-offs is not a strategy. KPIs that cannot be moved by anyone on the team are not KPIs. Priorities without invalidation conditions are opinions. Expert PMs leave this phase with: one sharp problem, one measurable outcome, ranked bets, and a set of things explicitly NOT being done.

Calibrated disagreement

The default failure in Define is dressing a weak prioritisation or strategy rationale in rigorous-looking structure. Challenge a ranking whose stated evidence doesn't actually support the order before formatting the table — see ../DOCTRINE.md.

Core sub-skills

1. Strategy formulation

Define direction, strategic pillars, and the set of bets that connect customer evidence, market conditions, and business goals. Seniority shows up in horizon, ambiguity tolerance, and willingness to commit to non-goals on the record.

Outputs: strategy memo (1–3 pages), vision statement, "where we play / how we win" narrative, explicit non-goals.

Anti-patterns: calling a roadmap a strategy, confusing goals with initiatives, avoiding trade-offs, overfitting to the latest escalation.

→ Deep-dive: references/strategy-memo.md

2. KPI architecture and metric-tree design

Build a measurement system that links customer value → product outcomes → business results via a North Star and supporting input metrics.

Outputs: North Star with formula, 2–3 layer metric tree, metric dictionary with owners, guardrail set, leading-vs-lagging taxonomy.

Anti-patterns: metrics that are easy to query but not causal, revenue-only scorecards, mixing health and target metrics, changing definitions mid-stream.

→ Deep-dive: references/kpi-tree.md

3. Opportunity sizing and demand modelling

Estimate scale × reachability × economics. Blend top-down TAM/SAM/SOM with bottoms-up cohort-level reasoning and stress-test the 3 load-bearing assumptions.

Outputs: sizing model, scenario table (base/upside/downside), assumptions sheet, sensitivity analysis.

Anti-patterns: TAM theatre, false precision, one heroic adoption curve, ignoring acquisition constraints.

→ Deep-dive: references/opportunity-sizing.md

The stage-4 artefact this skill commonly produces is the One Pager — see references/one-pager.md for the template (stage 4 of 8 in ../WORKFLOW.md); it cites the opportunity-tree node it comes from and lists the open assumptions (../pm-phase-discover/references/opportunity-solution-tree.md).

4. Business case and PRFAQ reasoning

Translate an initiative into costs, benefits, operating implications, and return logic — suitable for investment decisions. Use PRFAQ when the decision needs executive-level alignment or when working backwards from a press-release-quality outcome sharpens the bet.

Outputs: investment memo or PRFAQ, cost/benefit model with baseline, staged gate plan, downside scenario.

Anti-patterns: upside-as-certainty, ignoring ongoing ops cost, no baseline, single-point estimates, no migration cost.

→ Deep-dive: references/business-case-prfaq.md

5. Pricing and packaging

Design value metrics, plans, tiers, and monetisation structures aligned to customer value and GTM model. Shared with Product Marketing and Finance; this skill frames the PM share of the work.

Outputs: pricing memo, tier matrix, packaging proposal, migration FAQ, willingness-to-pay estimate, monetisation experiment plan.

Anti-patterns: copying competitors mechanically, cost-plus anchoring, changing price without packaging clarity, ignoring migration friction.

→ Deep-dive: references/pricing-packaging.md

6. Prioritisation frameworks

Apply an explicit method to compare options and rank work according to impact, effort, confidence, timing, and strategic alignment. The expert move is choosing and adapting the framework to context, not treating the score as truth.

Outputs: prioritisation matrix, scoring rubric, ranked initiative list, rationale memo.

Anti-patterns: treating the score as truth, hiding judgement behind arithmetic, scoring with weak evidence, never recalibrating.

→ Deep-dive: references/prioritisation-frameworks.md

At stage 5, use prioritisation only when several validated One Pagers compete for the same capacity. Do not score the original discovery backlog again. Record the selected bet, then pair with ../pm-phase-develop/references/backlog-scope-slicing.md to produce scope-slices.md before the PRD.

7. Roadmap narrative

Build a roadmap that tells a story: strategy → sequenced bets → dependencies → expected outcomes. Roadmaps create understanding, not just visibility.

Outputs: roadmap narrative + deck + FAQ, now-next-later view, stage gates, review pack.

Anti-patterns: date theatre, feature laundry lists, roadmap without rationale, failing to distinguish commitments from bets from options.

→ Deep-dive: references/roadmap-narrative.md

8. Decision documentation (DACI / PRFAQ / ADR)

Write docs that make a recommendation explicit, clarify trade-offs, and enable asynchronous decision-making. Senior PMs increasingly work through written alignment.

Outputs: decision memo, DACI page, ADR record, PRFAQ (cross-posted with §4).

Anti-patterns: status without recommendation, hidden trade-offs, burying the ask, no record of who decided what.

→ Deep-dive: references/decision-memo-daci.md

Workflow

  1. Load context when relevant — for project work, read .ai/memory/active-context.md first when present and resolve <slug> before opening relevant project files: profile, decisions, experiments, strategy/KPI artefacts, and discovery outputs. If no pointer exists, do not infer a project; bootstrap only when durable project context is needed and the name is known. When present, read .ai/memory/org/goals.md before building a KPI tree or ranking bets: the tree hangs off the cycle's objectives.
  2. Classify the ask — use strategy, KPI, sizing, business case, pricing, prioritisation, roadmap, or decision memo to choose the right reference. State it only when doing so clarifies scope or the next step.
  3. Name the choices — every artefact must include "what we chose NOT to do" and why.
  4. Ground in evidence — cite discovery outputs, data, or assumptions (flagged as such). Missing evidence → name it explicitly.
  5. Stress-test assumptions — for sizing/business case, identify top 3 load-bearing assumptions and show sensitivity.
  6. Persist durable outputs — when this task creates or materially changes project context, save strategy, decisions, KPI trees, or priorities under .ai/memory/projects/<slug>/ in the appropriate files. If a bet changes a cycle goal, update .ai/memory/org/goals.md and record the evidence in the relevant project changelog. A read-only answer needs no memory entry.

Output contract

## [Sub-skill] — [short title]

### Context read
### Choice / definition
### Rationale (why this over alternatives)
### What we are NOT doing
### Assumptions (load-bearing, with invalidation conditions)
### Evidence / baseline
### Risks and downside
### Next step (what would make this firm or force a re-look)
### Memory updates

Integration

  • Upstream: pm-phase-discover feeds segments, opportunities, hypotheses.
  • Downstream: pm-phase-develop (KPIs become PRD success criteria, priorities become backlog order), pm-phase-deliver (tree becomes dashboard + experiment primary metric).
  • Transversais: pm-transversal-stakeholder for DACI/exec reporting, pm-transversal-docs for Confluence/Jira publication.
  • Archetype lenses adapt the framing to the product type during the Define phase.

Communication modes follow ../../docs/patterns/COMMUNICATION_MODES.md. Per-skill: Lean (default) is the 1-page choice + rationale + non-goals; Standard is the full memo for leadership; Caveman is a 5-line bet / metric / size / risk / next.

Success criteria

  • every roadmap initiative traces to a strategic pillar
  • KPIs move because of shipped work (not just collected)
  • sizing/business cases survive finance scrutiny
  • the team can cite 2 bets killed this quarter and why
  • priorities change only with new evidence, not with loud voices

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.