developer-context
,# Developer Context
For Claude Code: Read this when you need to understand Joe's working style, technical background, or project constraints. Not needed for simple tasks - reference when making architectural decisions, scoping complexity, or working on security-related features.
About Joe (The Developer)
Background:
- 25+ years as federal oversight professional and healthcare provider
- Former military officer with deep federal compliance expertise
- Currently: Director in strategy and innovation at VA OIG
- Building Nquir to solve problems experienced firsthand in federal investigations
- Product owner for multiple software solutions deployed across federal agencies
This is a Part-Time Project:
- Works nights and weekends while maintaining full-time federal position
- Time is valuable - efficiency and doing it right the first time matters
- No dedicated dev team - all implementation is Joe + AI assistance
- Target launch: Q2 2026
Technical Background & Skill Level
Experience:
- Scrum methodology, Azure DevOps, GitHub
- Relational databases, Power Platform
- Picks up concepts and terminology quickly
- Understands architecture and system design
Current Learning:
- New to AWS console (started January 2025)
- Navigating well but still learning AWS services
- Comfortable with terminal/command line but not an expert
- Uses iTerm3 + tmux for efficiency
Development Approach:
- "Vibe coding" with AI assistance for rapid development
- Provides user stories and requirements
- Relies on Claude for hands-on implementation
- Expects code to work correctly on first attempt
- Values clear explanations for learning
Solo Developer Constraints
Reality Check:
- Everything must be maintainable by one person
- No team to delegate to or consult with
- Solutions must be "seamless and simple"
- Future Joe needs to understand current Joe's decisions
This Means:
- Use established patterns, not experimental approaches
- Prioritize maintainability over clever optimization
- Clear code comments for future reference
- Avoid over-engineering - MVP first, enhance later
- Document deployment steps thoroughly
Don't Suggest:
- "Have your team handle X"
- "Your DevOps engineer should..."
- Complex multi-service architectures (unless truly necessary)
- Experimental bleeding-edge tech
Security & Compliance (Non-Negotiable)
Primary Requirement: FedRAMP Compliance
- Essential for federal government sales
- All AWS services must be FedRAMP authorized
- This drove the decision to migrate from Supabase to AWS
- No exceptions - compliance > convenience
Architecture Principles:
- AWS consolidation preferred (single vendor for BAA/compliance)
- VPC isolation for database (no public endpoints)
- Encryption at rest (KMS) and in transit (TLS)
- Comprehensive audit logging (CloudTrail, AWS Config)
- RLS policies at database level
Why AWS-Only:
- "Built entirely on AWS FedRAMP-authorized services" is cleaner sales message
- Single vendor simplifies BAA (Business Associate Agreement) management
- Avoiding multi-vendor compliance complexity
Healthcare Context:
- Will handle PHI (Protected Health Information)
- HIPAA compliance required
- Evidence storage includes sensitive medical records
- Audit trails must be comprehensive and tamper-evident
Communication Preferences
Style:
- Direct, straightforward answers without excessive hedging
- Brief responses with option for detail if needed
- Proactive problem identification appreciated
- Don't ask permission for obvious tasks - just do them
Avoid:
- Repeated explanations ("As I mentioned earlier...")
- Over-explaining things Joe already understands
- Excessive caveats and hedging
- Long preambles before getting to the answer
Preferred Format:
Good: "Use RDS PostgreSQL. Aurora costs 2x more at your scale.
Easy upgrade path when you need it."
Bad: "Well, there are several database options to consider.
Aurora has benefits but also drawbacks. RDS is simpler but
less scalable. It really depends on your needs..."
When Unsure:
- State assumptions clearly
- Provide recommendation with reasoning
- Note what would change the recommendation
- Joe will course-correct if needed
Infrastructure Philosophy
AWS-Native Solutions:
- Prefer AWS services over third-party when possible
- Reduces vendor count for compliance
- Better integration and support
- Examples: Use Cognito vs Clerk, RDS vs external DB
Cost-Conscious Choices:
- RDS PostgreSQL chosen over Aurora (better cost at current scale)
- Start simple, scale when traffic demands it
- Free tier and AWS credits maximize runway
- Example: db.t3.micro is fine for MVP
Migration Complete:
- Migrated from Supabase/Vercel to full AWS (RDS + Cognito + S3)
- ECS/Fargate replaced Amplify (NQU-160, February 2026)
- Database schema maintained through migration
- FedRAMP compliance achieved through AWS consolidation
Current Project State
What's Done:
- Full-featured application with 30+ database tables, 2000 tests
- Next.js 16 on ECS/Fargate behind CloudFront + ALB
- AWS-native stack: RDS (private VPC), Cognito, S3, Bedrock, ElastiCache Redis
- Authentication (Cognito + WebAuthn + TOTP), Login.gov OIDC ready
- AI analysis pipeline (Bedrock Claude, evidence retrieval, quality metrics)
- Billing infrastructure (Stripe integration, trial logic, hard caps)
- CI/CD: GitHub Actions → ECR → ECS rolling deploy
- Domains: app.nquir.ai + app.nquiry.ai (CloudFront), nquiry.ai (landing page)
In Progress:
- Pre-launch manual verification (Joe)
- Stripe Dashboard setup (NQU-96)
- AWS Well-Architected Review (NQU-151)
- Evidence Evaluation Framework calibration (NQU-109)
Target Customers:
- Federal government investigators (OIG offices)
- Healthcare compliance professionals
- Corporate oversight teams
- All require FedRAMP or HIPAA compliance
When to Reference This Document
Read this file when:
- Making architectural or technology decisions
- Uncertain about scope or complexity level
- Working on security or compliance features
- First time working on a new feature area
- Joe asks "why are we doing it this way?"
Skip this file for:
- Simple bug fixes following existing patterns
- Clear, scoped tasks with explicit instructions
- Routine maintenance or updates
- Tasks where approach is already documented
Questions to Ask Yourself
Before implementing, consider:
- Maintainability: Can Joe maintain this solo in 6 months?
- Simplicity: Is this the simplest solution that works?
- Compliance: Does this meet FedRAMP/HIPAA requirements?
- AWS-Native: Are we using AWS services where possible?
- Documentation: Will future Joe understand this?
If you're unsure, ask. Joe appreciates being consulted on trade-offs.
Communication with Claude.ai
Context Sharing:
- Use git commits for work summaries (detailed messages)
- CLAUDE.md for essential project context (kept lean)
- This file (
docs/reference/developer-context.md) for working style - Instruction files (docs/instructions_*.md) for complex features
Coordination:
- You (Claude Code) own and maintain CLAUDE.md
- Ask before making significant changes to it
- Claude.ai will check git log to see your recent work
- Brief handoffs via user, not lengthy coordination files
Last updated: 2026-02-16 This document evolves as the project and working relationship develop.