proposal-system/CLAUDE.md

77 lines
3.6 KiB
Markdown
Raw Normal View History

# Proposal System - Claude Code Project Memory
## Project Overview
Proposal management platform for Sea Haven Industries. Dispatchers submit service requests, AI generates draft line items via Bedrock RAG, admins review/approve in a pricing workspace, system produces branded PDFs.
## Architecture
- **api/**: .NET 8 API, EF Core, PostgreSQL, Cognito JWT auth
- **web/**: React 19 + MUI v7 SPA, Vite, CloudFront + S3
- **mobile/**: React Native 0.85 iOS app, offline-capable, Hermes
- **lambdas/**: Python 3.12 Lambdas (ARM64): pdf-extract, pdf-generate, library-ingest, suggestions, oss-index-creator
- **infra/**: CDK TypeScript (foundation-stack, compute-stack, frontend-stack)
- **shared/**: Shared TypeScript API contracts
- **scripts/**: Local dev helpers
## Auth Model
External: Cognito JWT via API Gateway (web + mobile client IDs, groups: dispatchers/admins/sysadmins)
Internal: Lambdas → .NET Function URL with Secrets Manager API key via custom middleware
## Request Flow
1. Dispatcher submits proposal (web/mobile) → InReview (no Draft stage)
2. Bedrock RAG suggests line items from pricing library
3. Admin reviews in workspace, edits line items, approves
4. PDF generation queued via SQS → Python Lambda → branded PDF → S3
5. State machine: InReview → Approved → Sent → Revised
## Infrastructure
AWS us-east-1, RDS PostgreSQL 15, S3, SQS+DLQ, Cognito+Google OAuth, OpenSearch Serverless, Bedrock KB, GitHub Actions OIDC, CloudFront+S3 OAC
## Agent Delegation Rules
### For audit and hardening work, use the Explore-Plan-Execute pipeline:
1. Spawn specialist subagents for parallel investigation (api-security, web-audit, mobile-audit, lambda-pipeline, infra-cicd, qa-testing)
2. Consolidate findings into AUDIT-REPORT.md before implementing
3. Prioritize: Critical > High > Medium > Low
4. Implement fixes in logical phases, commit after each phase
5. Use separate git worktrees/branches for parallel implementation where safe
### Working Rules
- Never commit secrets, credentials, .env files, or local artifacts
- If secrets found in code: document, remove safely, ensure proper config mechanism
- Preserve existing business logic unless broken, insecure, or contradicted
- Run lint/typecheck/build/test after each phase
- If context reaches 65%, pause, commit, update AUDIT-REPORT.md with HANDOFF ADDENDUM
### Severity Levels
- **Critical**: security/data exposure/auth bypass/data corruption
- **High**: broken core workflow, deployment blocker, missing authz, invalid infra
- **Medium**: reliability, validation, logging, test gaps
- **Low**: cleanup, DX, docs, polish
## Audit Status
AUDIT-REPORT.md completed 2026-05-27. 5 Critical, 36 High, 75+ Medium, 60+ Low findings. Phase 1-5 complete: all Critical/High fixed, 17 Medium fixed, CI runs 108 tests.
### Critical Findings (fix first)
- **API-C1**: InternalApiKeyMiddleware applies globally, bypasses JWT on any route
- **API-C2**: JWT signature validation skipped when Authority is empty
- **WEB-C1**: JWT stored in localStorage (XSS token theft)
- **LAM-C1 / INF-H1**: Function URL authType NONE, publicly accessible
- **QA-C1**: Zero test coverage, no test projects, CI runs no tests
### Remediation Conventions
- Reference finding IDs (API-C1, WEB-H3, LAM-H4, etc.) in commit messages and code comments
- Format: `// Fix: API-C1 — scope internal key to /internal/ paths`
- Update AUDIT-REPORT.md after each phase: mark fixed findings, note deferred items
- Parallel-safe worktree splits: api/ changes, web/ changes, and infra/ changes don't conflict
- Verify after each phase: `dotnet build` (api), `npx tsc --noEmit` (web), `npx cdk synth` (infra)