Skip to content

blueprint.md

Latest commit

 

History

History
275 lines (184 loc) · 5.46 KB

blueprint.md

File metadata and controls

275 lines (184 loc) · 5.46 KB
Raw

SwiftGuard – Blueprint

1. Project Overview

Project Name: SwiftGuard

Tagline: LLM-driven contract testing framework for SWIFT microservices using Playwright.

Purpose: SwiftGuard validates SWIFT message–based microservice contracts by combining:

  • LLM-generated SWIFT messages
  • Schema-based contract validation
  • HTML-only UI rendering
  • Playwright-driven API + UI verification

The project is designed to run locally without Docker, making it lightweight, interview-ready, and easy to demonstrate.

2. Problem Statement

In banking and fintech systems:

  • Microservices exchange SWIFT messages (e.g., MT103)
  • Contract breaks lead to financial and regulatory risk
  • Test data is hard to generate and maintain
  • UI validation is often disconnected from backend contract tests

Traditional contract testing tools do not:

  • Support SWIFT message formats natively
  • Use AI to generate realistic test messages
  • Validate backend contracts and UI rendering together

SwiftGuard addresses these gaps.

3. Goals & Non-Goals

Goals

  • Validate SWIFT MT message contracts end-to-end
  • Use LLMs to generate realistic SWIFT messages dynamically
  • Perform contract validation without Docker
  • Use Playwright for both API and UI verification
  • Keep UI simulation minimal using pure HTML

Non-Goals

  • Full SWIFT network simulation
  • Production-grade SWIFT parsing engine
  • UI frameworks like React/Angular

4. Target Users

  • SDET / QA Automation Engineers
  • Banking & FinTech QA teams
  • Engineers testing message-driven microservices
  • Interviewers evaluating contract-testing expertise

5. Supported Message Types (Phase 1)

  • SWIFT MT103 – Customer Credit Transfer

Future extensibility:

  • MT202
  • MT940
  • ISO 20022 messages

6. High-Level Architecture

LLM (Gemini / Free LLM)
        ↓
SWIFT Message Generator
        ↓
SWIFT Parser
        ↓
Contract Validator (JSON Schema)
        ↓
Local Microservice (Node.js)
        ↓
HTML UI Renderer
        ↓
Playwright Contract Tests

7. Technology Stack

Layer Technology
LLM Gemini API / OpenRouter (free models)
Runtime Node.js
Contract Validation JSON Schema + AJV
UI Plain HTML + JavaScript
Test Framework Playwright
API Native Node HTTP server

8. LLM Integration Design

Responsibilities

  • Generate valid SWIFT MT103 messages
  • Generate negative/invalid variants for testing

Prompt Strategy

  • Enforce mandatory SWIFT fields
  • Restrict output to SWIFT format only
  • No explanations or markdown

Example prompt:

Generate a valid SWIFT MT103 message with mandatory fields only. Output strictly in SWIFT format.

9. Contract Definition Strategy

Contract Components

  1. SWIFT Syntax Rules
  2. Mandatory Field Validation
  3. Schema Enforcement

Example Mandatory Fields (MT103)

  • :20: Transaction Reference
  • :23B: Bank Operation Code
  • :32A: Value Date/Currency/Amount
  • :50K: Ordering Customer
  • :59: Beneficiary Customer
  • :71A: Charges

JSON Schema is used as the formal contract definition.

10. Local Microservice Design

Responsibilities

  • Accept SWIFT message input
  • Parse message into structured JSON
  • Validate against schema
  • Return success or error response

API Endpoints

  • POST /swift – Submit SWIFT message
  • GET /swift – Retrieve last processed message

No containerization or orchestration is used.

11. UI Simulation Strategy

Constraints

  • HTML-only
  • No UI frameworks
  • Minimal JavaScript

Purpose

  • Display parsed SWIFT fields
  • Act as UI contract consumer

UI rendering must match backend contract output exactly.

12. Playwright Testing Strategy

Test Layers

  1. Contract Validation Tests
  2. API Tests
  3. UI Rendering Tests

Playwright Responsibilities

  • Submit LLM-generated SWIFT messages
  • Validate API response
  • Verify UI displays correct contract data
  • Assert failures for invalid contracts

Playwright acts as the single end-to-end validation tool.

13. Test Scenarios

Positive Scenarios

  • Valid MT103 message accepted
  • UI renders all mandatory fields

Negative Scenarios

  • Missing mandatory fields
  • Invalid SWIFT tag formats
  • Corrupted LLM output

14. Folder Structure

SwiftGuard/
├── llm/
│   └── swiftGenerator.js
├── contracts/
│   └── mt103.schema.json
├── parser/
│   └── swiftParser.js
├── service/
│   └── swiftService.js
├── ui/
│   └── index.html
├── tests/
│   └── swiftContract.spec.ts
├── playwright.config.ts
└── README.md

15. Key Outcomes

  • Demonstrates contract testing expertise
  • Showcases AI-assisted test data generation
  • Validates backend and UI contracts together
  • Aligns with real-world banking QA challenges

16. Resume Value

This project demonstrates:

  • Advanced contract testing
  • Message-driven microservices validation
  • Playwright beyond UI testing
  • Practical AI integration in QA

Ideal for Senior SDET and FinTech QA roles.

17. Future Enhancements

  • Support ISO 20022 XML
  • Add contract versioning
  • Failure explanation using LLM
  • CI execution without Docker

End of Blueprint