Technical Documentation
This section provides a comprehensive technical reference for the SisterShield platform. It is intended for developers, maintainers, and evaluators who need to understand how the system is built, how its features work, and how to set up a development environment.
Quick Reference
| Technology | Version | Purpose |
|---|---|---|
| Next.js (App Router) | 14.2.21 | Full-stack React framework |
| TypeScript | 5.7.2 | Type-safe language |
| React | 18.3.1 | UI component library |
| PostgreSQL | — | Relational database |
| Prisma | 6.2.1 | ORM and database toolkit |
| NextAuth | 4.24.11 | Authentication (JWT strategy) |
| next-intl | 3.26.3 | Internationalization (ko-KR, en-US) |
| Tailwind CSS | 3.4.17 | Utility-first CSS framework |
| shadcn/ui (Radix UI) | — | Accessible component primitives |
| Lucide React | 0.469.0 | Icon library |
| Anthropic SDK | 0.39.0 | Claude Sonnet LLM integration |
| OpenAI SDK | 4.77.0 | GPT-4o and DALL-E 3 integration |
| Playwright | 1.40.0 | End-to-end testing |
| Zod | 3.24.1 | Runtime schema validation |
| bcryptjs | 2.4.3 | Password hashing |
Section Map
System Architecture Overview
High-level architecture, component diagram, data flow, and how the major subsystems connect.
Key Application Flows
Sequence diagrams for authentication, course creation, course play, submission review, and the Quick Exit safety mechanism.
Feature Map
A table mapping every major feature to its code location, description, and implementation status.
How to Use SisterShield
Role-based user guide covering Student, Teacher, and Admin workflows.
Stack & Setup Guide
Prerequisites, installation steps, environment variables, and development commands.
Privacy & Safety Controls
Data lifecycle, authentication details, Quick Exit behavior, AI safety constraints, and privacy considerations.
Semantic Design Tokens
The token architecture, CSS custom properties, Tailwind configuration, CTA styling rules, and anti-creep guidelines.
AI Prompting & Disclosure
How AI is used across the platform, provider abstraction, safety controls, and ethical disclosure.
Course Generator LLM Prompt Spec
The full system prompt, reference standards, output structure, few-shot examples, and safety strategy for AI-generated courses.
Source Code Layout
src/ app/ [locale]/ (auth)/ Login, Register pages (dashboard)/ Dashboard, Courses, Submissions, Progress, Settings api/ REST API routes components/ ui/ shadcn/ui components layout/ Header, Sidebar, Footer, LocaleSwitcher twine/ TwinePlayer (iframe renderer) hero/ Landing page components safety/ QuickExit, GetHelp courses/ CourseCard, ImagePromptPanel lib/ prisma.ts Database client singleton auth.ts NextAuth configuration analytics.ts Event tracking storage/ File storage abstraction (local, future S3) twine/ Parser, compiler, injector, builder, renderer, error detection i18n/ Locale config, navigation, request handling llm/ LLM client, story generation, translation, error fixing, image prompts hooks/ useProgress and other custom hooks messages/ en-US.json, ko-KR.json (translation strings)prisma/ schema.prisma Database schema seed.ts Seed script for test data