tool-design-sprint-prototype-plan
À propos
Cette compétence génère le plan détaillé pour la journée de construction du prototype (jeudi) d'un Design Sprint. Elle produit les attributions de rôles, un brief prototype, un script d'entretien utilisateur et une liste de vérification pour la répétition générale, en se basant sur le storyboard approuvé de la veille. Utilisez-la pour structurer et déléguer les travaux de création du prototype, bien que la construction effective se fasse en dehors de l'IA.
Installation rapide
Claude Code
Recommandénpx skills add product-on-purpose/pm-skills -a claude-code/plugin add https://github.com/product-on-purpose/pm-skillsgit clone https://github.com/product-on-purpose/pm-skills.git ~/.claude/skills/tool-design-sprint-prototype-planCopiez et collez cette commande dans Claude Code pour installer cette compétence
Documentation
Design Sprint Prototype Plan (Thursday morning)
Thursday is the prototype build day. This skill produces the morning planning artifact that makes the build run; it does NOT build the prototype itself. The skill assigns roles, locks the fidelity bar, drafts the Five-Act interview script, defines the trial-run gate, and confirms Friday participants. The actual building (Figma frames, slideware decks, role-play scripts, paper assemblies) is craft work the team owns through the rest of Thursday.
Family contract: docs/reference/skill-families/design-sprint-skills-contract.md. This skill is a member of design-sprint-skills.
When to Use
- It is Thursday morning of the Design Sprint and Wednesday's storyboard is signed off.
- The Friday participant tracker needs a Thursday-morning confirmation check; cancellations now trigger buffer activation.
- The interview script needs to be drafted (or refined from a prior sprint's template) before the interviewer can mock-run it Thursday afternoon.
- The team needs role assignments before parallel build work begins; ad-hoc assignment Thursday afternoon wastes 1-2 hours.
When NOT to Use
- Wednesday is not closed. Return to
tool-design-sprint-decide-and-storyboard; without a locked storyboard, the prototype has no spec. - The team is trying to use this skill to build the prototype. The prototype is craft work (per Ratified Decision 1 in the DS integration plan). This skill plans Thursday; the build happens in Figma, Keynote, paper, or whatever the medium is.
- The team has decided to skip Friday's trial run. Trial run is the gate that catches "prototype is unreachable" or "interview script confuses customer" before live interviews start. Skipping it converts Friday into a debugging session.
- Friday participant cancellations have triggered cohort-shrinks-to-fewer-than-4. Pause the sprint and decide whether to postpone Friday or accept a reduced-evidence test.
What This Skill Produces
A single bundled artifact with five sections:
- Prototype role plan: assigns 5 canonical roles (Maker, Stitcher, Writer, Asset Collector, Interviewer) across the team. Roles can double up on small teams; cannot triple up except for Asset Collector.
- Prototype brief: one-page spec covering what to build (storyboard panels in scope), fidelity bar (what "real enough" means for the medium), time allocation per role, and what's explicitly NOT being built.
- Interview script (Five-Act): canonical Five-Act structure (Welcome, Context, Intro, Tasks, Debrief) with team-supplied task wording derived from the storyboard. The act sequence is fixed; only Tasks is heavily customized.
- Trial-run checklist: the gates the prototype must pass Thursday afternoon (canonically 16:00-17:00) before Friday begins. Each gate is a yes/no item the interviewer or Maker confirms.
- Participant confirmation tracker: Thursday morning re-confirmation of all 5 Friday slots; cancellations trigger buffer-slot activation; failures cascade to a cohort-decision moment.
See references/TEMPLATE.md for the canonical structure and references/EXAMPLE.md for the Brainshelf book-catalog Thursday morning artifact.
Five Canonical Roles
Sprint book Chapter 16 defines these 5 roles. The skill assigns them; the humans fill them.
| Role | What they do | Typical owner |
|---|---|---|
| Maker | Builds the prototype surface (Figma frames, Keynote slides, role-play props). Owns visual and interaction work. | Design lead |
| Stitcher | Connects the Maker's surfaces into a coherent end-to-end flow (Figma interactions, Keynote transitions, role-play sequence). | Design lead OR engineering lead |
| Writer | Writes prototype copy (button labels, body text, error states, microcopy). Critical for credibility; bad copy breaks trust. | PM OR researcher |
| Asset Collector | Sources images, icons, sample data (book covers, names, dates, fake-but-believable data). | Anyone with free hands; often whoever isn't on the critical path. |
| Interviewer | Runs Friday's interviews. Spends Thursday afternoon drafting and mock-running the Five-Act script. | Researcher OR PM with interview experience. |
On a 4-person team, Maker and Stitcher typically merge (one design owner); Writer and Asset Collector can merge; Interviewer is always a dedicated role.
Five-Act Interview Structure
Canonical structure from Sprint book Chapter 17. The act sequence is fixed; the Tasks act is heavily customized; the other acts are mostly canonical wording.
| Act | Time | Purpose | Canonicality |
|---|---|---|---|
| Welcome | 5 min | Friendly intro; consent confirmation; explain not testing the customer | Canonical wording with team-name swap |
| Context | 5-10 min | Learn the customer's current behavior and context related to the challenge | Canonical questions; team adapts to challenge |
| Intro | 5 min | Introduce the prototype; explain it's a prototype (some things won't work) and ask the customer to think aloud | Canonical wording |
| Tasks | 20-30 min | Customer attempts the storyboard tasks; interviewer observes and probes; team observes silently from breakout room | Team-supplied wording derived from storyboard |
| Debrief | 5-10 min | Open-ended reactions; pricing questions if relevant; thank-you | Canonical structure; team adapts pricing question |
Common Pitfalls
- Treating the prototype as a deliverable instead of a learning tool. The prototype dies Friday at 17:00 PT regardless of how Friday goes. Investing engineering time in polishing it past the fidelity bar wastes Thursday.
- Polishing instead of building enough realism. Fidelity bar question: "Will a customer believe this is real for the 30-minute interview?" If yes, the prototype is done. Polishing past that point steals time from trial run and recovery.
- Interview script with leading questions. "Did you find the camera flow easy?" leads; "Walk me through what you just did" doesn't. Writer (or Interviewer) reviews script for leading questions before trial run.
- Skipping trial run. Trial run catches dead Figma links, missing assets, confusing copy, and interviewer-script-too-long failures. Always run with a fake customer (a teammate playing a target-profile customer) Thursday afternoon.
- Asset Collector role left unassigned because "it's not important." Asset Collector is the role that prevents "lorem ipsum on the demo screen" and "stock photo of a person who clearly isn't the target customer." Always assign.
- Five-Act Tasks act over-scripted. Tasks should be open-ended ("Try to capture this book and find it later in your library"); over-scripted tasks become a tutorial, not a test.
- Cancellations Thursday morning not surfaced because "we'll figure it out." Thursday morning is the last clean window to activate buffer participants. By Thursday afternoon, recruiter has lower confidence in confirmation.
Cross-Skill Usage
Prerequisites: tool-design-sprint-decide-and-storyboard. Prototype-plan consumes Wednesday's storyboard as the build spec. Without a storyboard, this skill has no panels to map roles against.
This skill does NOT invoke tool-note-and-vote. Thursday morning planning has no voting moment; the role plan and script decisions are Facilitator-led with Decider sign-off.
Next invocation in the sprint: tool-design-sprint-test-and-score Friday morning.
Canonical Sources
- Knapp, J., Zeratsky, J., and Kowitz, B. Sprint. Simon and Schuster, 2016. Thursday chapter (Chapters 14-16) and Five-Act Interview (Chapter 17).
- GV Design Sprint Guide. "Sprint Week Thursday." https://www.gv.com/sprint/
- Character Capital. "Design Sprint Day 4." https://www.character.vc
- Google Design Sprint Kit. "Thursday agenda template + Five-Act Interview script template." https://designsprintkit.withgoogle.com/
Decider Checkpoint
This skill ends with a Decider Checkpoint in references/TEMPLATE.md. The Decider approves the role plan, the fidelity bar in the prototype brief, the interview script's Tasks act, and the trial-run gate criteria. Once signed, the team disperses into parallel build work; the Decider returns Friday PM for the post-interview Decider review.
Dépôt GitHub
Compétences associées
content-collections
MétaCette compétence propose une configuration éprouvée en production pour Content Collections, un outil axé sur TypeScript qui transforme des fichiers Markdown/MDX en collections de données typées de manière sûre avec une validation Zod. Utilisez-la lors de la création de blogs, de sites de documentation ou d'applications Vite + React riches en contenu pour garantir la sécurité de typage et la validation automatique du contenu. Elle couvre tout, de la configuration du plugin Vite et de la compilation MDX à l'optimisation des déploiements et la validation des schémas.
polymarket
MétaCette compétence permet aux développeurs de créer des applications avec la plateforme de marchés prédictifs Polymarket, incluant l'intégration d'API pour le trading et les données de marché. Elle fournit également une diffusion de données en temps réel via WebSocket pour surveiller les transactions en direct et l'activité du marché. Utilisez-la pour mettre en œuvre des stratégies de trading ou pour créer des outils traitant les mises à jour de marché en direct.
creating-opencode-plugins
MétaCette compétence aide les développeurs à créer des plugins OpenCode qui s'interconnectent avec plus de 25 types d'événements tels que les commandes, les fichiers et les opérations LSP. Elle fournit la structure du plugin, les spécifications de l'API événementielle et les modèles d'implémentation pour les modules JavaScript/TypeScript. Utilisez-la lorsque vous avez besoin d'intercepter, de surveiller ou d'étendre le cycle de vie de l'assistant IA OpenCode avec une logique personnalisée pilotée par les événements.
sglang
MétaSGLang est un framework de service LLM haute performance spécialisé dans la génération rapide et structurée pour les workflows JSON, regex et agentiques grâce à son cache de préfixe RadixAttention. Il offre une inférence nettement plus rapide, particulièrement pour les tâches avec des préfixes répétés, ce qui le rend idéal pour les sorties complexes et structurées ainsi que les conversations multi-tours. Choisissez SGLang plutôt que des alternatives comme vLLM lorsque vous avez besoin d'un décodage contraint ou que vous construisez des applications avec un partage étendu de préfixes.
