Technical Architecture Skill

Provides comprehensive guidance on technical architecture and implementation decisions for full-stack applications, from system design to technology selection.

Quick Decision Guide

What do you need?

code

System architecture / Tech stack selection
  → Use: THIS skill (overall architecture, technology choices)

Security architecture / Threat modeling
  → Use: security skill (security patterns, compliance, risk)

Code review / Code quality
  → Use: code-review skill (line-by-line review, vulnerabilities)

CI/CD pipeline / Infrastructure / Deployment
  → Use: devops skill (pipelines, containers, monitoring)

UI/UX design / Wireframes / Mockups
  → Use: ux skill (user interface design, user flows)

Product planning / Feature prioritization
  → Use: product-owner skill (roadmap, user stories, backlog)

Scope of This Skill

This Skill Covers (Technical Architecture)

•✅ Overall system architecture and service design
•✅ Component architecture and code organization
•✅ API design (REST, GraphQL, tRPC)
•✅ Database schema and data modeling
•✅ State management patterns
•✅ Technology stack selection
•✅ Performance optimization strategy
•✅ Caching architecture
•✅ Scalability planning
•✅ Third-party service integration
•✅ Frontend and backend patterns

Use Other Skills For

•❌ Security architecture → Use security skill
•❌ Code security review → Use code-review skill
•❌ CI/CD pipelines → Use devops skill
•❌ Infrastructure setup → Use devops skill
•❌ UI/UX design → Use ux skill
•❌ Product planning → Use product-owner skill

This Skill Works With Other Skills

•security skill → Technical arch provides structure, security validates it
•code-review skill → Technical arch defines patterns, code-review ensures compliance
•devops skill → Technical arch chooses stack, devops deploys it
•ux skill → UX defines user needs, technical arch implements them

Pre-Planning Checklist

Before starting architectural planning, gather:

Requirements

• What problem are we solving?
• Who are the users? (Students, teachers, admins, public)
• What are the core features?
• What's the expected scale? (users, data volume, traffic)
• What's the timeline? (MVP in weeks vs production in months)

Constraints

• Budget limitations?
• Team skills and size?
• Performance requirements? (response time, throughput)
• Compliance requirements? (GDPR, HIPAA, etc.)
• Integration requirements? (existing systems, APIs)

Context

• Starting from scratch or evolving existing system?
• Mobile, web, or both?
• Real-time features needed?
• Offline support needed?
• Global users or regional?

References to Load

• Frontend decisions → references/frontend-patterns.md
• Backend decisions → references/backend-patterns.md
• Database design → references/database-design.md
• Deployment/hosting → references/deployment-systems.md

Architecture Decision Framework

1. Understand Requirements

Start with the pre-planning checklist above.

2. Choose Architecture Pattern

Based on your stage and needs, select the appropriate pattern:

For MVPs / New Products:

•Pattern 1: Serverless Full-Stack (fast iteration, low maintenance)

For Growing Products:

•Pattern 3: Hybrid (pragmatic middle ground)

For Established Products:

•Pattern 2: Containerized Microservices (full control, scale)

See "Common Architecture Patterns" section below for details.

3. Select Technologies

Consult reference files for detailed guidance:

•Frontend: references/frontend-patterns.md
•Backend: references/backend-patterns.md
•Database: references/database-design.md
•Deployment: references/deployment-systems.md

4. Validate with Other Skills

Security validation:

•Use security skill to threat model your architecture
•Review authentication/authorization patterns
•Validate data protection approach

Code organization validation:

•Use code-review skill to establish coding standards
•Define maintainability criteria

Deployment validation:

•Use devops skill to design CI/CD pipeline
•Plan infrastructure and monitoring

5. Evaluate Tradeoffs

Every architecture decision involves tradeoffs:

Complexity vs Flexibility:

•Simple solutions are easier to maintain
•Complex solutions offer more flexibility
•Rule: Choose simplest solution that meets requirements

Performance vs Development Speed:

•Optimizations take time to implement
•Premature optimization wastes effort
•Rule: Optimize based on actual bottlenecks, not assumptions

Cost vs Scale:

•Serverless: Low cost at low scale, expensive at high scale
•Dedicated servers: Higher baseline cost, cheaper at scale
•Rule: Choose based on expected traffic patterns

Vendor Lock-in vs Convenience:

•Managed services (Vercel, Supabase) are convenient but lock you in
•Self-hosted solutions give flexibility but require maintenance
•Rule: Consider long-term strategic value and exit cost

Build vs Buy:

•Building gives control but takes time
•Buying gets you features fast but adds dependencies
•Rule: Buy for non-core features, build for competitive advantages

Common Architecture Patterns

Pattern 1: Serverless Full-Stack (Recommended for MVPs)

Stack:

code

Frontend:    Next.js on Vercel
Backend:     Next.js API Routes (serverless functions)
Database:    Supabase (PostgreSQL) or PlanetScale (MySQL)
Cache:       Upstash Redis
File Storage: Vercel Blob or S3
Auth:        NextAuth.js or Clerk

When to use:

•✅ New products/MVPs
•✅ Small to medium scale (< 1M requests/month)
•✅ Want fast iteration
•✅ Limited DevOps resources
•✅ Unpredictable traffic

Pros:

•Fast to build and deploy (hours, not days)
•Scales automatically (0 → 10k users without changes)
•Low maintenance (no servers to manage)
•Pay only for usage (can start at $0/month)
•Excellent developer experience

Cons:

•Can get expensive at high scale (> $500/month at scale)
•Some vendor lock-in (but manageable)
•Cold start latency (200-500ms first request)
•Limited control over infrastructure

Migration path: When you outgrow this, move to Pattern 3 (Hybrid)

Pattern 2: Containerized Microservices (For Scale)

Stack:

code

Frontend:    Next.js in containers (ECS/Kubernetes)
Backend:     Node.js services in containers
Database:    RDS PostgreSQL with read replicas
Cache:       ElastiCache Redis
File Storage: S3
Auth:        Custom JWT implementation
Message Queue: SQS or RabbitMQ

When to use:

•✅ Proven product with growth (> 100k users)
•✅ Need more control over infrastructure
•✅ Complex business logic requiring service isolation
•✅ Team has DevOps skills
•✅ Predictable high traffic

Pros:

•Full control over infrastructure
•Better cost at scale ($500-2k/month for high traffic)
•Service isolation (can scale services independently)
•Technology flexibility (mix languages)
•No cold starts

Cons:

•More complex setup (weeks of infrastructure work)
•Requires DevOps expertise
•Higher baseline costs (minimum ~$200/month)
•More maintenance burden

Migration path: Usually don't migrate away, just add more services

Pattern 3: Hybrid (Pragmatic Middle Ground)

Stack:

code

Frontend:    Next.js on Vercel (fast deploys, edge)
Backend:     Dedicated API server on Railway/Render/Fly.io
Database:    Managed PostgreSQL (Supabase/Neon/RDS)
Cache:       Upstash Redis or managed Redis
File Storage: S3 or Cloudflare R2
Auth:        NextAuth.js or custom

When to use:

•✅ Outgrowing serverless costs (> $300/month)
•✅ Not ready for full container complexity
•✅ Need more backend flexibility (long-running tasks)
•✅ Want good DX with more control
•✅ Growing team (2-5 developers)

Pros:

•Better cost than pure serverless at scale
•More backend control (background jobs, websockets)
•Frontend still has great DX (Vercel)
•Easier than full microservices
•Good balance of control and convenience

Cons:

•More complex than pure serverless
•Need to manage API server
•Not as automated as Pattern 1
•Not as flexible as Pattern 2

Migration path: Can move to Pattern 2 when you need service isolation

Technology Selection Guidelines

Frontend Framework

Next.js (Recommended):

•✅ Full-stack capabilities (API routes, server actions)
•✅ Excellent performance (SSR, ISR, SSG)
•✅ Great DX with Vercel
•✅ Large ecosystem and community
•✅ Built-in optimizations (images, fonts, code splitting)
•Use for: Most web applications, SEO-critical sites

React SPA:

•✅ Simple deployment (static files)
•✅ Full client-side control
•✅ Works well with any backend
•❌ Worse SEO (unless prerendering)
•❌ Slower initial load
•Use for: Admin dashboards, internal tools, web apps where SEO doesn't matter

React Native (Mobile):

•✅ Code sharing with web (if using React)
•✅ Native performance
•✅ Single codebase for iOS + Android
•❌ More complex than web
•❌ App store deployment overhead
•Use for: Mobile apps (student/guardian apps for educational app)

Backend Language/Runtime

Node.js/TypeScript (Recommended):

•✅ Share code and types with frontend
•✅ Great ecosystem (npm packages)
•✅ Fast development (familiar to frontend devs)
•✅ Good for I/O-heavy tasks (APIs, databases)
•✅ Async by default
•Use for: Most applications, especially full-stack TypeScript

Python:

•✅ Excellent for ML/AI workloads
•✅ Great data processing libraries (pandas, numpy)
•✅ Good for scripts and background jobs
•❌ Harder to share code with frontend
•❌ Slower for I/O-heavy tasks
•Use for: ML pipelines, data processing, computer vision, background jobs

When to use both:

•Node.js for API server (user-facing)
•Python for background processing (ML, data analysis)

Database

PostgreSQL (Recommended for most apps):

•✅ Excellent for relational data
•✅ ACID transactions (data integrity)
•✅ Rich query capabilities (joins, aggregations)
•✅ JSON support for flexibility
•✅ Great performance with proper indexing
•✅ Mature ecosystem
•Use for: Educational apps, CRUDs, any structured data

MongoDB:

•✅ Flexible schema (good for rapid prototyping)
•✅ Fast writes
•✅ Good for unstructured data
•❌ Weaker consistency guarantees
•❌ No joins (application-level only)
•Use for: Rapid prototyping, analytics, logs, event data

Decision matrix:

•Structured data with relationships? → PostgreSQL
•Need ACID transactions? → PostgreSQL
•Flexible/evolving schema? → MongoDB or PostgreSQL (with JSONB)
•Heavy analytics? → PostgreSQL (with indexing) or ClickHouse

For detailed schema design, consult references/database-design.md

State Management

React Query/SWR (Recommended for server state):

•✅ Handles 90% of state needs
•✅ Automatic caching and refetching
•✅ Loading/error states handled
•✅ Less boilerplate than Redux
•✅ Optimistic updates
•Use for: API data, server state, any data from backend

Zustand (Recommended for client state):

•✅ Simple API
•✅ Small bundle size (<1KB)
•✅ No boilerplate
•✅ TypeScript support
•✅ No context provider needed
•Use for: Global UI state, user preferences, theme, modals

Context API (Built-in):

•✅ No dependencies
•✅ Good for simple state
•❌ Can cause unnecessary re-renders
•❌ Verbose for complex state
•Use for: Simple global state (theme, user info)

Redux:

•✅ Powerful debugging (DevTools)
•✅ Middleware ecosystem
•✅ Time-travel debugging
•❌ More boilerplate
•❌ Steeper learning curve
•Use for: Very complex state, need time-travel debugging

Rule of thumb:

•Server data → React Query/SWR
•Global UI state → Zustand
•Simple context → Context API
•Complex state machine → Redux or XState

Architecture Review Checklist

When reviewing or planning architecture, verify:

System Design

• Architecture pattern chosen (serverless/hybrid/containers)?
• Frontend and backend clearly separated?
• API design follows REST/GraphQL best practices?
• Database schema normalized appropriately?
• State management approach defined?
• Authentication strategy chosen?
• File storage solution selected?

Scalability

• Can handle 10x current load?
• Database properly indexed?
• Caching strategy in place?
• No N+1 query problems?
• Background jobs for heavy processing?
• CDN for static assets?

Performance

• Images optimized?
• Code splitting implemented?
• Lazy loading for heavy components?
• Database queries optimized?
• Caching for expensive operations?
• Bundle size monitored?

Maintainability

• Clear code organization (feature-based)?
• Consistent naming conventions?
• Components focused (single responsibility)?
• TypeScript for type safety?
• Linting and formatting configured?
• Documentation for complex logic?

Developer Experience

• Easy local development setup?
• Environment variables managed properly?
• Hot reload working?
• Fast build times (< 30 seconds)?
• Clear error messages?

Security (High-Level - Delegate to security skill)

• Use security skill for detailed review
• Authentication implemented?
• HTTPS enabled?
• Secrets not in code?
• Input validation at API boundaries?

Testing (High-Level - Delegate to code-review skill)

• Use code-review skill for test strategy
• Critical paths have tests?
• API endpoints tested?
• E2E tests for key flows?

Deployment (High-Level - Delegate to devops skill)

• Use devops skill for CI/CD design
• Automated deployments configured?
• Staging environment available?
• Rollback plan ready?

Best Practices Summary

Code Organization

•Feature-based structure over type-based

code

✅ Good: features/assignments/, features/students/
❌ Bad: components/, hooks/, utils/, services/

•Colocate related code (components, hooks, styles together)
•Keep components small (< 200 lines, single responsibility)
•Use absolute imports (@/components vs ../../../components)
•Separate business logic from UI (custom hooks, service layer)

Performance

•Code splitting for routes (automatic in Next.js)
•React.memo for expensive components
•Optimize images (Next.js Image component, WebP format)
•Database indexes for frequent queries
•Cache expensive operations (React Query, Redis)
•Lazy load heavy components
•Debounce/throttle expensive operations (search, scroll)

API Design

•RESTful conventions (GET/POST/PUT/DELETE, proper status codes)
•Consistent naming (camelCase for JSON, snake_case for DB)
•Versioning (/api/v1/)
•Pagination for collections (limit/offset or cursor-based)
•Error handling (consistent error format)
•Rate limiting (prevent abuse)
•Documentation (OpenAPI/Swagger)

Database Design

•Normalize to avoid redundancy (usually 3NF)
•Index frequently queried fields
•Foreign keys for referential integrity
•Migrations for schema changes
•Soft deletes for important data (deleted_at column)
•Timestamps (created_at, updated_at) on all tables

Integration with Other Skills

With security Skill

code

Technical Skill: Designs system architecture
         ↓
   Identifies integration points
         ↓
security Skill: Reviews security architecture
         ↓
   Identifies security requirements
         ↓
Technical Skill: Incorporates security controls
         ↓
security Skill: Validates implementation

Example workflow:

•You: "Design authentication system for student/teacher app"
•Technical skill: Recommends NextAuth.js with JWT, session strategy
•security skill: Reviews JWT storage, session management, threat model
•Technical skill: Updates architecture based on security recommendations

With code-review Skill

code

Technical Skill: Defines architecture patterns
         ↓
   Sets coding standards
         ↓
code-review Skill: Reviews code against patterns
         ↓
   Identifies deviations and issues
         ↓
Technical Skill: Updates patterns if needed

Example workflow:

•You: "Review this API endpoint implementation"
•Technical skill: Confirms it follows API design patterns
•code-review skill: Reviews code quality, security, maintainability
•Technical skill: Suggests architectural improvements if patterns emerge

With devops Skill

code

Technical Skill: Chooses technology stack
         ↓
   Defines deployment requirements
         ↓
devops Skill: Designs CI/CD pipeline
         ↓
   Sets up infrastructure
         ↓
Technical Skill: Validates architecture works in production

Example workflow:

•You: "Set up deployment for Next.js app with PostgreSQL"
•Technical skill: Recommends Vercel + Supabase (serverless pattern)
•devops skill: Configures CI/CD, environment variables, monitoring
•Technical skill: Confirms deployed architecture matches design

With ux Skill

code

ux Skill: Designs user interface and flows
         ↓
   Identifies data requirements
         ↓
Technical Skill: Designs API and data model
         ↓
   Ensures UI can be implemented efficiently
         ↓
ux Skill: Adjusts design if technical constraints exist

Example workflow:

•You: "Design task breakdown feature for students"
•ux skill: Creates wireframes and user flow
•Technical skill: Designs database schema and API endpoints
•ux skill: Adjusts UI based on data structure (e.g., nested tasks)

Common Architecture Mistakes to Avoid

Over-Engineering

•❌ Microservices for a 2-person team
•❌ Complex caching before measuring performance
•❌ Custom authentication when NextAuth exists
•❌ GraphQL when REST would suffice
•Rule: Start simple, add complexity when needed

Under-Engineering

•❌ No database indexes
•❌ No caching for expensive operations
•❌ No error boundaries
•❌ No loading states
•Rule: Handle the basics (errors, loading, caching)

Wrong Technology Choices

•❌ MongoDB for highly relational data
•❌ Redux for simple global state
•❌ Serverless for compute-heavy workloads
•❌ NoSQL for ACID requirements
•Rule: Choose technologies based on requirements, not hype

Poor Code Organization

•❌ Type-based folders (all components in one folder)
•❌ Deeply nested components (6+ levels)
•❌ 1000+ line files
•❌ Mixing business logic and UI
•Rule: Feature-based, colocate, keep files small

Performance Mistakes

•❌ No code splitting
•❌ Large unoptimized images
•❌ Re-rendering entire lists
•❌ N+1 queries
•Rule: Optimize the common path, measure before optimizing

Security Mistakes (Architecture Level)

•❌ Authentication logic only on frontend
•❌ Direct database access from frontend
•❌ API keys in frontend code
•❌ No rate limiting
•Rule: Use security skill for comprehensive review

Migration Strategies

From Serverless to Hybrid

When: Serverless costs exceed $300-500/month

Approach:

•Keep frontend on Vercel (still cost-effective)
•Move API to Railway/Render/Fly.io
•Move database to managed service with connection pooling
•Add Redis for caching
•Keep file storage on S3/Vercel Blob

Timeline: 1-2 weeks

From Hybrid to Microservices

When: Need service isolation, team > 5 people

Approach:

•Identify service boundaries (auth, payments, notifications)
•Extract one service at a time (start with least critical)
•Add message queue (SQS, RabbitMQ) for async communication
•Set up service mesh (if using Kubernetes)
•Add centralized logging and monitoring

Timeline: 1-3 months

From Monorepo to Multiple Repos

When: Team > 10 people, clear service ownership

Approach:

•Extract shared code to npm packages
•Split services into separate repos
•Set up consistent CI/CD across repos
•Use dependency management (Renovate, Dependabot)

Timeline: 2-4 weeks

Getting Started

New Project Workflow

•Define requirements (use pre-planning checklist)
•Choose architecture pattern (serverless/hybrid/microservices)
•Select technologies (consult reference files)
•Design data model (references/database-design.md)
•Plan API structure (references/backend-patterns.md)
•Design frontend (references/frontend-patterns.md)
•Validate with security skill (threat model, security controls)
•Set up with devops skill (CI/CD, infrastructure)
•Document decisions (ADRs - Architecture Decision Records)

Existing Project Review Workflow

•Understand current architecture (diagram, document)
•Identify pain points (slow, expensive, hard to maintain)
•Assess against checklist (above)
•Identify gaps (missing caching, poor indexes, etc.)
•Propose improvements (prioritized by impact)
•Create migration plan (phased approach)
•Validate with other skills (security, code-review, devops)

Requesting Architectural Guidance

For best results, provide:

Context:

•What you're building (brief description)
•Current state (greenfield vs existing system)
•Team size and skills

Specific Question:

•What decision are you trying to make?
•What options are you considering?
•What's the tradeoff you're evaluating?

Constraints:

•Budget ($ per month)
•Timeline (MVP in 2 weeks vs production in 3 months)
•Scale requirements (expected users, traffic)
•Compliance (GDPR, HIPAA, etc.)

Example:

"I'm building an educational app for 1000 students and 50 teachers. Need task management, calendar, and file uploads. Team is 2 developers with React/Node.js skills. Budget is $100/month. Need MVP in 4 weeks. Should I use serverless or dedicated backend?"

This skill will consult the relevant references and provide context-specific, actionable recommendations.

Reference Files

Load these for detailed guidance:

•references/frontend-patterns.md - React/Next.js patterns, state management, performance
•references/backend-patterns.md - API design, authentication, caching, background jobs
•references/database-design.md - Schema design, indexing, migrations, scaling
•references/deployment-systems.md - Hosting platforms, CI/CD, monitoring

For comprehensive coverage, use with:

•security skill - Security architecture, threat modeling, compliance
•code-review skill - Code quality, testing strategy, maintainability
•devops skill - CI/CD pipelines, infrastructure, deployment
•ux skill - User interface design, user flows, wireframes