Penthera
PEN Quality Agent

Your intelligent quality engineering companion — PEN Quality Agent analyzes documents, generates test cases, manages test execution, and drives quality decisions so your team ships with confidence.

Document Analysis
Test Management
Test Generation
Quality Management
4
Core Functions
AI
Powered by Claude
E2E
Full Lifecycle
24/7
Quality at Scale

Core Functions

Four integrated capabilities that power the full quality engineering lifecycle — from requirement analysis to release readiness.

Document Analysis

AI-driven parsing and understanding of requirements, user stories, and specifications from Jira, Confluence, and other sources to extract testable acceptance criteria.

Jira Integration Confluence Sync Requirement Parsing Acceptance Criteria

Test Generation

Automatically generates comprehensive test cases — including edge cases, negative tests, and boundary conditions — from analyzed requirements using Claude AI.

Claude AI Edge Cases Negative Tests Auto Priority

Test Management

Centralized tracking of test suites, execution status, and coverage across sprints — with real-time dashboards synced to Jira boards.

Test Suites Execution Tracking Sprint Coverage Real-time Sync

Quality Management

End-to-end quality oversight with metrics, trend analysis, and reporting to ensure release readiness and continuous improvement.

Quality Metrics Trend Analysis Release Readiness CI/CD Reports

Quality Engineering Lifecycle

Analyze Requirements
Generate Test Cases
Manage Execution
Assure Quality

Integration Architecture

How Claude AI orchestrates Playwright MCP for real-time browser exploration, generates Penthera test scripts, and pushes results to ReportPortal.

Claude AI Agent

Central intelligence that understands requirements, plans test strategy, generates code, and self-heals failures.

  • Natural Language Understanding
  • Test Strategy & Planning
  • Code Generation & Analysis
  • Self-Healing & Recovery

Playwright MCP

Bridge between Claude and the browser — enables real-time exploration, interaction, and validation via MCP protocol.

  • browser_navigate / snapshot
  • browser_click / fill_form
  • browser_take_screenshot
  • browser_evaluate / network

Penthera Framework

Java + Playwright test automation framework with TestNG, Page Objects, data-driven testing, and thread-safe execution.

  • TestNG Suites & DataProviders
  • Page Object Model + ObjectManager
  • ThreadLocal Parallel Execution
  • REST-Assured API Testing

ReportPortal

Centralized test reporting with real-time dashboards, ML-powered failure analysis, and historical trend tracking.

  • TestNG Listener Auto-Report
  • Real-time Dashboards
  • ML Failure Classification
  • Historical Trends & Analytics

Penthera Framework Layers

A layered architecture designed for maintainability, thread-safety, and parallel execution.

Test Layer

Orchestration & Data

TestNG Suites
smoke, promotion, voucher, login
BaseTest
@Before/@After hooks, lifecycle
Test Scripts
Payment, Login, Display, API
DataProviders
JSON + Excel per environment

Object Layer

Page Objects & APIs

ObjectManager
Composition root, ThreadLocal
Page Objects
Home, Login, Payment, TPP, Result
API Wrappers
PaymentApi, PromotionApi, ZingCardApi, CommonApis

Core Layer

Drivers & Services

PlaywrightFactory
ThreadLocal browser mgmt
REST-Assured
HTTP client
Services
Business logic

Support Layer

Reporting, Helpers, Context

ExtentReports
HTML reports + screenshots
13 Helpers
JSON, Excel, CSV, Date, String...
TestContext
Thread-safe test data storage
Models & POJOs
Request/Response mapping

End-to-End Workflow

The 9-step AI-driven test lifecycle — from natural language request to self-healing execution.

1

User Request

Natural language test requirement — "Test login with phone number on VNG Webshop"

Claude AI
2

Test Planning

Claude analyzes Penthera codebase, identifies Page Objects, DataProviders, and existing patterns.

Claude AI
3

MCP Exploration

Navigate to target URL, take accessibility snapshots, identify page elements and locators.

Playwright MCP
4

MCP Interaction

Click buttons, fill forms, interact with the real browser to validate flows before coding.

Playwright MCP
5

MCP Validation

Take screenshots, evaluate JavaScript, monitor network requests to verify expected behavior.

Playwright MCP
6

Code Generation

Claude generates Penthera-compatible test scripts, page objects, and test data files.

Claude + Penthera
7

Test Execution

mvn clean test — TestNG runs suite with parallel threads, Playwright drives browsers.

Penthera
ReportPortal
8

Artifact Collection

Reports (HTML + Excel), screenshots, video recordings, traces, and logs all generated and sent to ReportPortal.

ReportPortal
9

Analysis & Self-Healing

Claude reviews results from ReportPortal, fixes broken locators, updates scripts — continuous loop until pass.

Claude AI
Continuous feedback loop — Steps 7-9 repeat until all tests pass

Technology Stack

The libraries and tools powering the integration.

1.26
Playwright
7.8
TestNG
5.6
ReportPortal Agent
5.4
REST-Assured
5.1
ExtentReports
17
Java

ReportPortal Configuration

# reportportal.properties
rp.endpoint = https://reportportal.vnggames.net
rp.launch = Demo
rp.project = penthera
# META-INF/services/org.testng.ITestNGListener
com.epam.reportportal.testng.ReportPortalTestNGListener

See the Full Integration Diagram

Interactive visualization of all components, connections, and data flows.

Open Integration Diagram