runbook-startup-mvp

Déclencheurs

• startup MVP
• MVP build
• minimum viable product
• rapid product build
• startup launch
• product-market fit
• quick launch
• lean startup
• fast build
• 4-week build
• 6-week build
• idea to product
• validate idea
• startup sprint
• rapid development
• MVP validation

Instructions

Ce runbook couvre la construction d'un MVP startup — un nouveau produit qui doit valider le product-market fit rapidement. La vitesse compte, mais la qualité aussi. Allez de l'idée à un produit en production avec de vrais utilisateurs en 4-6 semaines.

Effectif d'agents

Équipe Core (Toujours Active) — utiliser organization_spawn:

• Agents Orchestrator: Contrôleur de pipeline
• Senior Project Manager: Conversion spec-to-task
• Sprint Prioritizer: Gestion du backlog
• UX Architect: Fondation technique
• Frontend Developer: Implémentation UI
• Backend Architect: API et base de données
• DevOps Automator: CI/CD et déploiement
• Evidence Collector: QA pour chaque task
• Reality Checker: Portail de qualité final

Équipe Growth (Activée Semaine 3+):

• Growth Hacker: Stratégie d'acquisition
• Content Creator: Contenu de lancement
• Social Media Strategist: Campagne réseaux sociaux

Équipe Support (Selon besoin):

• Brand Guardian: Identité de marque
• Analytics Reporter: Métriques et tableaux de bord
• Rapid Prototyper: Expériences de validation rapide
• AI Engineer: Si le produit inclut des features IA
• Performance Benchmarker: Load testing avant lancement
• Infrastructure Maintainer: Configuration production

Semaine 1 : Discovery + Architecture (Phase 0 + 1 Compressée)

Jour 1-2 — Discovery Compressée:

• Trend Researcher: Scan compétitif rapide (1 jour, pas de rapport complet)
• UX Architect: Wireframe des user flows clés
• Senior Project Manager: Conversion spec en task list

Jour 3-4 — Architecture:

• UX Architect: CSS design system + architecture de composants
• Backend Architect: Architecture système + schéma de base de données
• Brand Guardian: Fondation de marque rapide (couleurs, typographie, voix)
• Sprint Prioritizer: Backlog RICE-scored + sprint plan

Jour 5 — Foundation Setup:

• DevOps Automator: Pipeline CI/CD + environnements
• Frontend Developer: Project scaffolding
• Backend Architect: Base de données + API scaffold
• Quality Gate: Architecture Package approuvée

Semaine 2-3 : Core Build (Phase 2 + 3 Compressée)

Sprint 1 (Semaine 2) — utiliser les mécaniques de skill agency-phase-3-build:

• Agents Orchestrator gère la boucle Dev-QA
• Frontend Developer: Core UI (auth, main views, navigation)
• Backend Architect: Core API (auth, CRUD, business logic)
• Evidence Collector: QA chaque task complétée
• AI Engineer: Features ML si applicable
• Sprint Review en fin de semaine

Sprint 2 (Semaine 3):

• Continuer la boucle Dev-QA pour les features restantes
• Growth Hacker: Design viral mechanics + système de referral
• Content Creator: Commencer création de contenu de lancement
• Analytics Reporter: Set up tracking et dashboards
• Sprint Review en fin de semaine

Semaine 4 : Polish + Hardening (Phase 4 Compressée)

Jour 1-2 — Quality Sprint:

• Evidence Collector: Full screenshot suite
• Performance Benchmarker: Load testing
• Frontend Developer: Fix QA issues
• Backend Architect: Fix API issues
• Brand Guardian: Brand consistency audit

Jour 3-4 — Reality Check:

• Reality Checker: Final integration testing
• Infrastructure Maintainer: Production readiness
• DevOps Automator: Production deployment prep

Jour 5 — Gate Decision:

• Reality Checker verdict
• IF NEEDS WORK: Quick fix cycle (2-3 days)
• IF READY: Proceed to launch
• Executive Summary Generator: Stakeholder briefing

Semaine 5-6 : Launch + Growth (Phase 5)

Semaine 5 — Launch:

• DevOps Automator: Production deployment
• Growth Hacker: Activate acquisition channels
• Content Creator: Publish launch content
• Social Media Strategist: Cross-platform campaign
• Analytics Reporter: Real-time monitoring
• Support Responder: User support active

Semaine 6 — Optimize:

• Growth Hacker: Analyze and optimize channels
• Feedback Synthesizer: Collect early user feedback
• Experiment Tracker: Launch A/B tests
• Analytics Reporter: Week 1 analysis
• Sprint Prioritizer: Plan iteration sprint

Décisions Clés

Pièges Courants & Mitigations

• Scope creep: Sprint Prioritizer applique MoSCoW — "Won't" means won't
• Over-engineering: Rapid Prototyper mindset — validate first, scale later
• Skipping QA: Evidence Collector runs on EVERY task — no exceptions
• No monitoring: Infrastructure Maintainer sets up monitoring in Week 1
• No feedback mechanism: Analytics + feedback built into Sprint 1

Utiliser knowledge_write pour persister architecture decisions, sprint results, et launch metrics.

Livrables

• [ ] Competitive scan et concept validé
• [ ] Architecture Package (compressée)
• [ ] Pipeline CI/CD fonctionnel
• [ ] Features core implémentées et QA'd
• [ ] Reality Checker READY verdict
• [ ] Production deployment
• [ ] Acquisition channels actifs
• [ ] Week 1 analytics report
• [ ] User feedback collection active

Métriques de Succès

Vérifier

• The deploy command was actually run and the build/log output (or deploy URL) is captured
• The deployed URL was opened and returned a 2xx; key routes were sampled, not just the index
• Environment variables required by the app are present in the target environment; missing-var failures were ruled out
• A rollback plan (previous deployment ID, git SHA, or one-line revert command) is documented before promoting to production
• Health/observability check (logs, error tracker, status page) was inspected post-deploy; baseline error rate is recorded
• DNS / domain / SSL configuration was confirmed, not assumed to carry over from previous deploys