使用说明

🤖 AI QA Automation Agent — Playwright MCP + JIRA MCP

A fully autonomous, AI-powered QA Automation Framework that generates test cases, executes them in a real browser via Playwright MCP, captures evidence, and raises JIRA bugs — all without writing a single line of test code manually.

📋 Table of Contents

Overview
Test Design Method — RICE POT
Architecture
Project Structure
Tech Stack
Prerequisites
Setup & Installation
How to Run
Test Cases
Execution Results
JIRA Integration
Generated Artifacts
Special Defect Scenario
MCP Configuration

🎯 Overview

This framework is driven entirely by an AI QA Automation Agent operating through two MCP (Model Context Protocol) servers:

MCP Server	Role
🎭 Playwright MCP	Browser automation — navigate, interact, screenshot
🔵 JIRA MCP	Bug management — create tickets, attach evidence

What the Agent Does Automatically:

📝 Generates test cases using the RICE POT design method
📊 Exports test cases to Excel (.xlsx)
🌐 Executes tests live in Chromium via Playwright MCP
✅❌ Marks results PASS or FAIL
📸 Captures screenshots for failed tests
📁 Collects metadata & logs (browser, OS, network, console)
🖥️ Generates an HTML report with KPI dashboard
🗄️ Stores all artifacts in a structured folder
🐛 Creates JIRA bug tickets for failures
📎 Attaches evidence to the JIRA ticket

🧪 Test Design Method — RICE POT

Every test case is designed using the RICE POT framework:

Letter	Meaning	Description
R	Requirement	The feature or user story being tested
I	Inputs	Data entered into the system
C	Conditions	The state of the system before/during execution
E	Expected Result	What the system should do
P	Preconditions	What must be true before running the test
O	Output	Actual observable output
T	Test Steps	Step-by-step instructions for reproduction

🏗️ Architecture

┌─────────────────────────────────────────────────────┐
│              AI QA Automation Agent                 │
│           (SKILL.md instruction set)                │
└──────────────┬──────────────────────┬───────────────┘
               │                      │
     ┌─────────▼──────────┐  ┌────────▼────────┐
     │   Playwright MCP   │  │    JIRA MCP     │
     │  Browser Control   │  │  Bug Reporting  │
     └─────────┬──────────┘  └────────┬────────┘
               │                      │
     ┌─────────▼──────────┐  ┌────────▼────────────────┐
     │  https://freecrm   │  │  myqatask.atlassian.net  │
     │  (Target App)      │  │  Project: SCRUM          │
     └────────────────────┘  └─────────────────────────┘

📁 Project Structure

MCP - E2E Framework/
│
├── 📄 SKILL.md                      ← AI Agent instruction set (10 steps)
├── 📄 README.md                     ← This file
│
├── 🔧 playwright.config.ts          ← Playwright configuration
├── 🔧 package.json                  ← Node.js dependencies
│
├── 🧪 tests/
│   └── freecrm.spec.ts              ← Playwright test spec (6 test cases)
│
├── 📜 generate_excel.js             ← Step 2: Generate Excel test cases
├── 📜 build_artifacts.js            ← Step 6-8: Generate logs, report & summary
├── 📜 create_jira_bug.js            ← Step 9: Create JIRA bug via REST API
├── 📜 attach_jira_evidence.js       ← Step 10: Attach evidence to JIRA ticket
│
└── 📦 freecrm-test-results/         ← All generated artifacts
    ├── 📊 freecrm_login_testcases.xlsx     ← Test cases (Step 2)
    ├── 🌐 test_execution_report.html       ← HTML report (Step 7)
    ├── 📋 execution_summary.json           ← Machine-readable results (Step 8)
    ├── 📝 conversation_history.txt         ← Execution session log
    ├── 📝 prompt_used.txt                  ← Skill prompt reference
    │
    ├── 📸 screenshots/                     ← Screenshots (Step 5)
    │   ├── TC_LOGIN_01_screenshot.png
    │   ├── TC_LOGIN_02_screenshot.png
    │   ├── TC_LOGIN_03_failure_screenshot.png  ⚠️ Bug Evidence
    │   ├── TC_LOGIN_04_screenshot.png
    │   ├── TC_LOGIN_05_screenshot.png
    │   ├── TC_LOGIN_06_screenshot.png
    │   └── test_execution_report_fullpage.png
    │
    └── 📁 logs/                            ← Execution logs (Step 6)
        ├── metadata.json
        ├── playwright_execution.log
        ├── network.log
        └── console.log

🛠️ Tech Stack

Technology	Version	Purpose
Node.js	18+	Runtime environment
TypeScript	Latest	Test spec language
Playwright	1.51.x	Browser automation
@playwright/mcp	Latest	Playwright MCP server
@nexus2520/jira-mcp-server	Latest	JIRA MCP server
xlsx	Latest	Excel file generation
form-data	Latest	JIRA attachment upload

✅ Prerequisites

Node.js v18 or higher
Antigravity (AI agent with MCP support)
JIRA account with API token
MCP configuration file set up (see MCP Configuration)

⚙️ Setup & Installation

# 1. Clone or open the project
cd "MCP - E2E Framework"

# 2. Install dependencies
npm install

# 3. Install Playwright browser
npx playwright install chromium

# 4. Install additional packages
npm install xlsx form-data

🚀 How to Run

Run All 10 Steps via AI Agent

Simply open SKILL.md and execute it through the Antigravity AI agent with Playwright MCP and JIRA MCP configured.

Run Individual Steps Manually

# Step 2 — Generate Excel test cases
node generate_excel.js

# Step 3 — Execute Playwright tests
npx playwright test

# Step 7-8 — Build HTML report + logs + summary
node build_artifacts.js

# Step 9 — Create JIRA bug
node create_jira_bug.js

# Step 10 — Attach evidence to JIRA
node attach_jira_evidence.js

🧪 Test Cases

TC ID	Title	Type	Priority	Severity
TC_LOGIN_01	Valid Login	✅ Positive	High	Critical
TC_LOGIN_02	Valid Email Case Sensitivity	✅ Positive	Medium	Major
TC_LOGIN_03	Invalid Email & Password	❌ Negative	High	Critical
TC_LOGIN_04	Invalid Email Format	❌ Negative	Medium	Minor
TC_LOGIN_05	Arabic/Chinese Characters	❌ Negative	Low	Minor
TC_LOGIN_06	Dummy Login Credentials	❌ Negative	Low	Minor

📊 Execution Results

Metric	Value
Total Tests	6
Passed	5 ✅
Failed	1 ❌
Pass Rate	83%
Browser	Chromium (Playwright MCP)
Target URL	https://ui.freecrm.com
Execution Date	2026-03-15

🐛 Special Defect Scenario

Performance Bug — TC_LOGIN_03

Field	Details
JIRA Ticket	SCRUM-25
Bug Title	Login error message delayed by 3 seconds for invalid credentials
Priority	🔴 High
Labels	`Performance-Bug` `Login` `FreeCRM` `Playwright-MCP` `P1`
Steps	Navigate → Enter invalid credentials → Click Login
Expected	Error message within < 1 second
Actual	Error message appears after ~3 seconds
Evidence	Screenshot + HTML report + logs attached to JIRA

🔵 JIRA Integration

The agent automatically creates a JIRA bug for every failed test case.

Bug Details for SCRUM-25:

Project: My Automation QA Team (SCRUM)
Issue Type: Bug
Rich ADF description with steps, expected, actual, environment
Attachments uploaded:
- TC_LOGIN_03_failure_screenshot.png
- test_execution_report.html
- execution_summary.json
- playwright_execution.log

🗄️ Generated Artifacts

Artifact	Format	Description
`freecrm_login_testcases.xlsx`	Excel	All 6 test cases with RICE POT fields
`test_execution_report.html`	HTML	Premium dark-themed dashboard report
`execution_summary.json`	JSON	Machine-readable pass/fail summary
`screenshots/*.png`	PNG	One screenshot per test case
`logs/playwright_execution.log`	Text	Full step-by-step execution trace
`logs/metadata.json`	JSON	Browser, OS, timestamp, user agent
`logs/network.log`	Text	All HTTP requests & responses
`logs/console.log`	Text	Browser console output

⚙️ MCP Configuration

Configure your mcp_config.json (at ~/.gemini/antigravity/mcp_config.json):

{
    "mcpServers": {
        "playwright": {
            "command": "npx",
            "args": ["-y", "@playwright/mcp@latest"]
        },
        "jira": {
            "command": "npx",
            "args": [
                "-y",
                "@nexus2520/jira-mcp-server@latest",
                "--url",
                "https://YOUR-DOMAIN.atlassian.net"
            ],
            "env": {
                "JIRA_EMAIL": "your-email@example.com",
                "JIRA_API_TOKEN": "YOUR_JIRA_API_TOKEN"
            }
        }
    }
}

🔐 Security Note: Never commit your actual JIRA API token to version control. Use environment variables or a secrets manager.

📄 License

This project is for demonstration and educational purposes as part of the AI QA Automation Agent skill framework.

Built with ❤️ by AI QA Automation Agent & Bhaumik Gohel Powered by Playwright MCP + JIRA MCP