Product Strategy

What this skill does

product-strategy produces product/strategy.md — the strategy layer of the product, between the vision (why) and the roadmap (when/how). It defines who the product is for (personas), how it is positioned, the measurable goals, the monetization model, and the time-to-market urgency. It is authored from human intent and reconciled against the approved vision. The roadmap and the capabilities are sequenced and shaped to serve it.

Where the state tree lives

Generic skill — hard-codes no path. Resolve the state-tree root from the project's CLAUDE.md (project-declared paths, DESIGN_PRODUCT_PIPELINE.md §2); default docs/, and say which you used. {state-root} is that root, so the strategy lives at {state-root}/product/strategy.md and the vision at {state-root}/product/vision.md.

STOP — how this skill works

Before writing strategy.md you MUST, in order:

1. CHECK THE VISION. Read product/vision.md — it must exist and be

status: approved. No approved vision → STOP; the strategy serves the vision and cannot precede it.

1. ELICIT INTENT. The strategy is authored from the human's intent about

market, customers, positioning, monetization, and urgency. If you do not have that intent, ASK. Do NOT derive the strategy from the codebase, the capabilities, or the readied-idea stub.

1. RECONCILE against the vision (see "Reconciliation").
2. WRITE the strategy (see "What strategy.md must contain").
3. SEEK HUMAN APPROVAL — the strategy is a high-stakes artifact; explicit

human approval, like the vision.

Hard rules:

• Authored from human intent, never derived from what exists.
• Concise. A strategy nobody can internalize and reference in daily decisions

is dead. No 50-page documents.

• It is the WHAT, not the WHY (that is the vision) and not the WHEN (the

roadmap). It does not restate the vision or pre-empt the roadmap.

• You do not approve the strategy yourself. Approval is an explicit human act.

Reconciliation

Hold the drafted strategy against the approved vision. The strategy must serve the vision, not contradict it. Surface, as findings flagged for a human:

• a strategy element (a persona, a monetization model, a goal) that the vision

does not support — either the vision needs revisiting or the element is wrong;

• a vision goal the strategy does nothing to advance — a gap in the strategy.

Do not resolve these by bending the strategy or the vision. Flag them. If the strategy genuinely requires the vision to change, that goes back to product-vision — the strategy does not edit the vision.

What strategy.md must contain

Write to {state-root}/product/strategy.md.

Frontmatter:

---
revision: 1
status: draft
approved_by:
approved_on:
why-link: product/vision.md
reconciled-against: vision.md revision N
---

Body — every section required; concise throughout:

• Target customers & personas — the specific segments the product serves,

as personas: their context, needs, behaviours, what they do today. This is the home of personas — they belong here, not in the vision.

• Positioning — how the product is different from the alternatives, in a

sentence or two; who it is explicitly NOT for.

• Measurable goals — the concrete, trackable objectives for the product

(the vision's goals are aspirational; these are measurable). Each should serve a vision goal.

• Monetization — how, or whether, the product makes money. If the product

is to be paid, this is where that decision lives — not in the vision and not in the roadmap.

• Time-to-market urgency — how fast the product needs to reach users, and

WHY (competitive window, runway, the need to test an unproven bet). This is the input the roadmap uses to sequence for earliest shippable value.

• Reconciliation report — the findings from "Reconciliation".
• Changelog — per DESIGN_PRODUCT_PIPELINE.md §2b (Changelog); first entry

revision 1 — <date> — initial strategy.

The strategy is the WHAT. It names no capabilities, no roadmap items, no implementation. It is concise enough to be internalized.

Re-running the strategy

strategy.md is durable state and can go stale as the vision changes. product-strategy is re-runnable: re-read the current vision, reconcile, and re-propose. On every run, bump revision, set reconciled-against to the vision revision used, and append a ## Changelog entry naming the change. A re-run produces a reconciliation report of what changed and is human-approved like any strategy.

Approval

strategy.md is approved only by an explicit human act — the same rule as the vision. You cannot approve it; a sub-agent cannot; vague assent does not count.

1. Present the strategy to the user FOR REVIEW — the personas, the

positioning, the goals, the monetization decision, the time-to-market urgency, and any reconciliation findings. Explicitly invite rejection or revision.

1. On unambiguous human approval, write status: approved, approved_by,

approved_on, and the current revision.

1. Any later edit to an approved strategy INVALIDATES the approval — reset to

draft, clear the fields, seek fresh approval.

On approval, before writing status: approved, append a ## Changelog entry for this revision: the new revision number, the date, and one or two sentences on what changed and why. Bumping revision without the matching changelog entry is a defect — they are one act.

When this skill applies

• Defining or revising the product's strategy — this skill.
• Setting the product's vision — NOT this skill; that is product-vision.
• Sequencing what to build next — NOT this skill; that is product-roadmap.
• A single feature idea — NOT this skill; that is idea-grill / change-scope.

Precedence

vision.md outranks strategy.md — the strategy serves the vision and never edits it. The strategy in turn outranks the roadmap and the capability map: the roadmap sequences to serve the strategy, capabilities are shaped to serve it. If the strategy exposes that the vision is wrong, that is a finding for product-vision and a human — flag it, do not resolve it by bending the strategy.

Document versioning

Per DESIGN_PRODUCT_PIPELINE.md §2b. product-strategy sets its own revision and reconciled-against: vision.md revision N; a vision whose revision exceeds that value means this strategy is STALE and must be re-run.