saas-init
Skill2 starsChanged 6 months agoMITView on GitHub
Bootstrap complet d'un projet SaaS from scratch. Système de Q/R interactif pour configurer chaque aspect.
- Reads credentials
In this repository32 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- apexSkill
- api-genSkill
- assets-pipelineSkill
- brainstormSkill
- debugSkill
- deep-updateSkill
- deploy-checkSkill
- depsSkill
- docs-genSkill
- e2e-testsSkill
- env-setupSkill
- featureSkill
- fixSkill
- forkSkill
- meta-promptSkill
- migrateSkill
- onboardSkill
- oneshotSkill
- perfSkill
- prdSkill
- qualitySkill
- refactorSkill
- reviewSkill
- schema-genSkill
- scope-taskSkill
- session-startSkill
- stripeSkill
- ui-auditSkill
- uiSkill
- variationsSkill
The file
Files
Instructions
Phase 0 — Interview Projet (Q/R obligatoire)
Utiliser AskUserQuestion pour chaque question. Ne RIEN supposer.
Question 1 — Type de projet :
- SaaS B2B (Recommandé)
- SaaS B2C
- Marketplace
- API / Backend only
Question 2 — Framework :
- Next.js 15 (Recommandé)
- Remix
- Nuxt 4
- SvelteKit
- FastAPI + React/Vue
Question 3 — Base de données :
- PostgreSQL + Drizzle (Recommandé)
- PostgreSQL + Prisma
- Supabase (inclut auth + storage)
- MongoDB + Mongoose
Question 4 — Auth :
- Auth.js v5 (Recommandé)
- Clerk (rapide mais payant)
- Supabase Auth
- Better Auth
- Custom (JWT + sessions DB)
Question 5 — Paiements :
- Stripe (Recommandé)
- Lemonsqueezy (simple, Merchant of Record)
- Pas encore
Question 6 — UI :
- shadcn/ui + Tailwind (Recommandé)
- Radix + Tailwind
- Mantine
- Material UI
Question 7 — Déploiement :
- Vercel (Recommandé)
- Railway
- Docker + VPS (Coolify/Dokku)
- AWS (ECS/Lambda)
Question 8 — Package Manager :
- pnpm (Recommandé)
- bun
- npm
- yarn
Question 9 — Extras : (multiSelect)
- Email transactionnel (Resend)
- File upload (UploadThing / S3)
- Real-time (WebSockets / SSE)
- Background jobs (Inngest / BullMQ)
- Analytics (PostHog / Plausible)
- Monitoring (Sentry)
- AI features (Vercel AI SDK)
Question 10 — MCP Servers : (multiSelect)
- Context7 — Docs à jour (Recommandé)
- Playwright — E2E testing
- Sequential Thinking — Raisonnement structuré
- Aucun pour l'instant
Phase 1 — Scaffolding
- Initialiser avec le CLI officiel du framework choisi
- Créer la structure hexagonale :
src/core/entities/ src/core/usecases/ src/core/ports/ src/adapters/db/ src/adapters/api/ src/adapters/auth/ src/shared/ src/config/ - Configurer imports absolus (
@/), TypeScript strict, ESLint, Prettier - Créer
.env.exampleavec toutes les variables - Initialiser git +
.gitignore - Installer les dépendances extras choisies (Q9)
Phase 2 — Fondations SaaS
- Schema DB de base :
users(id UUID v7, email, name, role, avatarUrl, createdAt, updatedAt, deletedAt)organizations(id, name, slug unique, plan, createdAt, updatedAt, deletedAt)memberships(userId, organizationId, role enum)
- Provider auth choisi + middleware protection
- Logger structuré (pino)
- Pattern réponse API :
{ data, error, meta } - Error types de base :
NotFoundError,UnauthorizedError,ValidationError
Phase 3 — DX Setup
- Scripts :
dev,build,test,lint,db:push,db:studio,db:seed - Configurer vitest avec coverage
- Tests de base : health check, auth flow
- Seed de dev avec données de test
Phase 4 — Finalisation
- Installer MCPs choisis (Q10) dans
.claude/mcp/config.json - Sauver stack dans
.claude/memory/stack.md - Mettre à jour
.claude/memory/session.md - Créer ADR 0001 documentant les choix techniques
- Afficher récap : stack, structure, prochaines étapes
Post-Init — Suggestions
Proposer à l'utilisateur :
- "Voulez-vous que je génère une landing page ? →
/ui landing" - "Voulez-vous configurer le CI/CD ? →
/deploy-check" - "Voulez-vous ajouter une feature ? →
/feature [description]" - "Voulez-vous un PRD complet ? →
/prd brainstorm"
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
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.

