| name | product-design |
| description | Product-level design methodology covering D01–D14 — vision, discovery, UX/CX design, and concept validation. Produces the "what and why" before any feature work begins. Provides the methodology for each of the 14 product-design artifacts; any single artifact workflow can be applied independently. Triggers include "product vision", "discovery", "user personas", "journey mapping", "user testing", "friction points", and any "produce D0X" request. |
| version | 1.0.0 |
Product Design Skill
This skill holds the methodology for product-level design — from first idea to validated concept. It covers the 14 product-design artifacts (D01–D14): how to produce each one well, what makes each good or bad, and the reusable frameworks behind them.
Each artifact has a dedicated workflow describing the procedure for producing it. Each workflow is self-contained methodology and can be applied on its own. The frameworks the workflows lean on live in context/.
Announce at start: "I'm using the Product Design skill."
The 14 Artifacts
The artifacts span four design concerns. Each produces exactly one D-numbered document.
Vision
| Artifact | Workflow | What it captures |
|---|
| D01 — Vision Statement | workflows/d01-vision.md | The problem, target user, solution direction, scope, and bets |
| D02 — Success Criteria | workflows/d02-success-criteria.md | Falsifiable, measurable definition of success |
Discovery
| Artifact | Workflow | What it captures |
|---|
| D03 — Market Research | workflows/d03-market-research.md | Competitive landscape, market size, positioning gap |
| D04 — Feasibility Report | workflows/d04-feasibility.md | Technical, operational, and business feasibility |
| D05 — Stakeholder Insights | workflows/d05-stakeholder-insights.md | Synthesized user/stakeholder research themes |
UX Design — User Research Synthesis
| Artifact | Workflow | What it captures |
|---|
| D06 — User Insights | workflows/d06-user-insights.md | Behavioral clusters and cross-cutting patterns |
| D07 — User Needs | workflows/d07-user-needs.md | Solution-agnostic, prioritized need statements |
| D08 — User Personas | workflows/d08-user-personas.md | Behaviorally-grounded persona archetypes |
CX Design — Journey Strategy
| Artifact | Workflow | What it captures |
|---|
| D09 — Journey Maps | workflows/d09-journey-maps.md | End-to-end persona journeys with emotional arcs |
| D10 — Touchpoints | workflows/d10-touchpoints.md | Catalog of interaction points, critical ones flagged |
| D11 — Friction Points | workflows/d11-friction-points.md | Prioritized inventory of where the experience breaks down |
Validation
| Artifact | Workflow | What it captures |
|---|
| D12 — Test Results | workflows/d12-test-results.md | Raw, structured output of user-testing sessions |
| D13 — User Feedback | workflows/d13-user-feedback.md | Themed, prioritized synthesis of test results |
| D14 — Validated Designs | workflows/d14-validated-designs.md | A verdict per design, with persona-anchored evidence |
Context Reference Files
These hold the frameworks and worked examples the workflows reference but don't duplicate. Read the relevant one before applying the corresponding workflow.
| File | Frameworks it holds | Relevant to |
|---|
context/vision-anatomy.md | 9-component vision anatomy, anti-patterns, worked example | D01, D02 |
context/measurement-frameworks.md | Metric schema, leading/lagging, SMART/OKR/HEART/AARRR, guardrails | D02 |
context/research-techniques.md | JTBD, Five Whys, competitive analysis, market sizing, survey design | D03, D05 |
context/feasibility-frameworks.md | Tech/ops/business feasibility matrices, ROI, POC criteria, decision records | D04 |
context/design-patterns.md | UX/UI pattern library, accessibility, design systems | D06–D11 |
context/validation-techniques.md | Test methods, severity rubric, persona-evidence rules, failure modes | D12–D14 |
How These Artifacts Build On Each Other
Each artifact draws on evidence captured in earlier ones — these are evidence dependencies, not a required running order:
- D02 success criteria each trace to a D01 vision element.
- D03–D05 scope their research against the D01 target user and problem.
- D06 synthesizes D03 + D05 into behavioral clusters.
- D07 needs derive from D06 clusters.
- D08 personas are built from D06 clusters and D07 needs.
- D09 journeys are mapped per D08 persona, anchored to D07 needs.
- D10 touchpoints are extracted from D09 stages.
- D11 friction points reference D10 touchpoint IDs and D09 emotional valleys.
- D12 testing exercises the D11 friction points against D08 personas.
- D13 synthesizes D12 and reconciles each D11 friction point.
- D14 renders verdicts from D13 with D08 persona-anchored evidence.
When applying a workflow, read the artifacts it builds on so your output stays anchored to that evidence. Each workflow's "Builds on" note lists what it expects to read.
Inputs From Bootstrap (Track + Architecture)
Bootstrap runs before product-design and lays down docs/architecture/. Product-design reads it — it is not blind to the stack:
- Read the marker
docs/architecture/bootstrap.md for the track (brownfield or greenfield) and the stack summary.
- Read
docs/architecture/tech-stack.md and integration-map.md when present.
Feed this forward:
- D01 may surface the recorded stack as a stated constraint — it does not design it (the stack stays in
docs/architecture/; D01's redirect principle still holds).
- D04 assesses feasibility against the recorded stack and frames it by track: fixed for brownfield, a revisable proposal for greenfield. On a "feasible with changes" / "not feasible" verdict it returns a stack-feedback signal (verdict + gap + driver + recommended route) in its report. Acting on it is the host's job, not the skill's: the
/shark product-design verb re-runs bootstrap in reconcile mode (greenfield) or files a tech-debt entry / constraint note (brownfield).
If no marker or architecture docs exist, product-design runs exactly as before.
Core Principles (Apply Across All Workflows)
- Elicit, don't invent. Vision (D01–D02) and validation (D12–D14) are human decisions. Use
AskUserQuestion. Never fabricate a user quote, metric baseline, participant, or approval.
- Persona-anchored evidence. From D08 onward, every observation, decision, and verdict ties to a named persona ID.
- One question at a time. Vision and validation work rewards depth. Prefer multiple choice when the answer space is small; open-ended when it isn't.
- Mark unknowns explicitly. Write
TBD — <owner> by <date> — never guess.
- Trace every claim to a source. A user statement, a session observation, a cited study, or a metric reading. No orphan assertions.
Output Standards (All D-Artifacts)
- Each artifact uses its exact D-numbered filename (e.g.,
D01-vision-statement.md), stored together in a single product-design directory (conventionally docs/product/).
- Every claim traces to a source: a user statement, a session observation, a cited study, or a metric reading.
- Unknowns marked
TBD — <owner> by <date>.
- Cross-reference the artifacts an output builds on (D02 cites D01 sections; D13 cites D12 session IDs; etc.).
- Footer:
Version: 1.0 — YYYY-MM-DD — author: <name>.
Output Standards (Feature-Level Artifacts)
When product-design evidence feeds feature-level design:
- Reference D08 persona IDs and D09 journey stage names verbatim.
- Include the rationale for design decisions, not just the decisions.
Tools Used
AskUserQuestion — All elicitation, approvals, and human decisions.Read — Read the artifacts a workflow builds on, and the relevant context/ file, before producing output.Write — Create the artifact.WebSearch — Market research (D03), technical research (D04).
