Test Gen

🧪 Test Generation Assistant

🎯 Overview

AI-assisted test generation and execution system that automates the complete testing workflow:

• Test Case Generation: Generate comprehensive test cases from JIRA tickets
• E2E Test Code Generation: Create executable E2E test code
• Test Execution: Run E2E tests and capture results
• PR Submission: Automatically create pull requests for generated test code
• JIRA Integration: Update JIRA tickets with test status and results

This system uses multiple specialized agents to handle each phase of the testing process, ensuring consistent and thorough test coverage for OpenShift components.

📋 Prerequisites

Before you can use this test generation system, you need to set up the following integrations:

1. Install Claude Code, Cursor, or Gemini

• Claude Code access: Follow Claude Code Install Instructions
• Cursor access: Download and install Cursor for an IDE-based experience with AI assistance
• Gemini access: Refer to Gemini documentation for AI assistance
• Important: Choose Claude Code for command-line interaction, Cursor for a full IDE experience, or Gemini for alternative AI assistance. All can be used to run the test generation system.

1.1 Install as Claude Plugin (Recommended for Teams) ⭐

ONE-COMMAND INSTALLATION - The simplest way to share with your team:

claude plugin marketplace add YOUR_USERNAME/test-generation-assistant

That's it! The plugin is now available in Claude Code with all agents and workflows.

What's included:

• ✅ All test generation agents
• ✅ Workflow orchestrator
• ✅ Configuration templates
• ✅ Automatic updates
• ✅ Team collaboration support

Complete Guide: PLUGIN_INSTALLATION.md

---

2. JIRA MCP Access

• Apply for jira-mcp-snowflake token: Refer to Jira-MCP-Snowflake-Token-Guide

• Config jira-mcp-snowflake mcp server:

  claude mcp add jira-mcp-snowflake https://jira-mcp-snowflake.mcp-playground-poc.devshift.net/sse --transport sse -H "X-Snowflake-Token: your_token_here"
  

3. DeepWiki MCP Connection

claude mcp add -s user -t http deepwiki https://mcp.deepwiki.com/mcp

4. Prepare Test Generation Rules

• Purpose: Configure rules to guide AI in generating realistic and executable test cases
• Location: config/rules/
• Content: Component-specific testing guidelines, validation criteria, and best practices

5. Configure GitHub CLI (gh)

• Purpose: Enable PR creation and GitHub repository operations
• Installation:

  # macOS
  brew install gh
  
  # Linux/Windows
  # Follow: https://cli.github.com/manual/installation
  

6. Environment Configuration

Purpose: Configure environment variables for GitHub, JIRA, and OpenShift access

Quick Setup:

# Copy example environment file
cp .env.example .env

# Edit with your credentials
nano .env

Required Variables:

• GITHUB_FORK_OWNER - Your GitHub username (for E2E test generation)
• JIRA_SERVER - JIRA server URL (default: https://issues.redhat.com)
• JIRA_TOKEN - Your JIRA authentication token

For E2E Testing:

• KUBECONFIG - Path to your OpenShift cluster kubeconfig
• TEST_PLATFORM - Target platform (AWS, Azure, GCP, etc.)

See: ENV_SETUP.md for detailed environment setup guide

7. Execution Method

CRITICAL: All agents must be executed by reading agent YAML configurations directly.

Execution Pattern:

1. Read CLAUDE.md (for Claude Code), GEMINI.md (for Gemini), or Cursor documentation (see Cursor Configuration in Prerequisites) for agent instructions
2. Read specific agent YAML config (e.g., config/agents/test_case_generation.yaml)
3. Execute each step manually using available tools
4. Verify each step before proceeding

🎛️ Workflow Orchestrator (PRIMARY ENTRY POINT)

MANDATORY: The workflow orchestrator is the PRIMARY entry point for ALL test generation requests.

CRITICAL: Before executing ANY user request, AI MUST:

1. Read config/agents/workflow_orchestrator.yaml FIRST
2. Let orchestrator identify workflow automatically
3. Execute agents in correct sequence

What is it?

A top-level agent that automatically:

• ✅ Parses user requests and identifies intent
• ✅ Matches requests to predefined workflows
• ✅ Validates prerequisites automatically
• ✅ Executes agents in correct sequence
• ✅ Handles REGENERATE mode
• ✅ Generates comprehensive execution reports

Configuration

Location: config/agents/workflow_orchestrator.yaml
Guide: config/agents/WORKFLOW_ORCHESTRATOR_GUIDE.md

Rule Keyword Enforcement

When AI reads any configuration or guideline files:

• MANDATORY = ABSOLUTE REQUIREMENT (must be executed without exception)
• NEVER = ABSOLUTE PROHIBITION (must never be violated)
• CRITICAL = STOP and verify (must pause before proceeding)
• FORBIDDEN = ABSOLUTE PROHIBITION (same as NEVER)
• Violation of any keyword = Complete execution failure, must restart

Supported Workflows

1. full_flow - Test Case → E2E Code → Execution
2. test_case_only - Test Case Generation
3. e2e_generation - E2E Test Generation
4. test_execution - Test Execution Only
5. pr_submission - PR Submission
6. e2e_and_run - E2E Generation + Execution

Quick Start

The orchestrator automatically identifies your intent from natural language:

# Complete flow (3 agents executed automatically)
"Generate test cases and run them for HIVE-2883"

# Single agent via orchestrator
"Create test case for HIVE-2883"

# Regenerate mode (skips prerequisite checks)
"re-create e2e test for HIVE-2883"

See config/agents/WORKFLOW_ORCHESTRATOR_GUIDE.md for detailed usage instructions.

---

🚀 Agents

1. Generate Test Cases

Purpose: Generate manual test cases from JIRA tickets
Agent Config: config/agents/test_case_generation.yaml
Input: JIRA issue key (e.g., "HIVE-2883")
Output:

• test_requirements_output.yaml - Detailed requirements analysis
• test_strategy.yaml - Test strategy based on requirements
• {JIRA_KEY}_test_case.md - Executable test cases in markdown format

Example Usage:

"Create test case for HIVE-2883"
"Generate test case for JIRA issue HIVE-2883"
"Generate test cases and run them for HIVE-2883"  # Full flow

Features:

• Systematic thinking framework (4 mandatory phases)
• JIRA data extraction and analysis
• Architecture analysis with DeepWiki integration
• User-reality driven test scenarios
• Quantitative validation methods

2. Generate E2E Test Code

Purpose: Generate E2E test code based on existing test cases
Agent Config: config/agents/e2e_test_generation_openshift_private.yaml
Output: E2E test code integrated into openshift-tests-private repository

2.1 Execute E2E Tests

Purpose: Execute generated E2E tests and capture results
Agent Config: config/agents/test-executor.yaml
Output: Test execution results and reports

3. Submit E2E Pull Request

Purpose: Create PR for generated E2E tests
Tools: Git, GitHub CLI (gh)
Output: Created pull request with E2E test code

4. Update JIRA QE Comment [In Progress]

Agent Dependencies

• test-executor requires completion of e2e_test_generation_openshift_private
• PR submission requires completion of e2e_test_generation_openshift_private (and optionally test-executor)
• JIRA QE comment update can be executed after any agent completion

🚀 Usage Options

Option 1: Claude Code Usage

Quick Start

1. Setup Prerequisites (see above sections):

   • Install Claude Code
   • Configure JIRA MCP access
   • Configure DeepWiki MCP
   • Configure GitHub CLI (gh)

2. Open Claude Code:

   claude
   

3. Navigate to project directory:

   cd /path/to/.ai_testgen
   

4. Execute agents directly:

   # Generate test cases
   "Create test case for HIVE-2883"
   
   # Generate E2E code
   "Generate E2E case for HIVE-2883"
   
   # Execute E2E tests
   "Run E2E tests for HIVE-2883"
   
   # Complete end-to-end flow (executes all 3 agents in sequence)
   "Generate test cases and run them for HIVE-2883"
   # This automatically executes: test_case_generation → e2e_test_generation_openshift_private → test-executor
   
   # Create PR
   "Create PR for HIVE-2883 E2E tests"
   
   # Add QE comment
   "Add QE comment to HIVE-2883"
   

How It Works

1. Claude reads CLAUDE.md - Gets agent instructions
2. Reads agent YAML configs - Gets specific execution steps
3. Executes steps - Uses available tools (JIRA MCP, WebFetch, Bash, etc.)
4. Verifies outputs - Ensures each step completes successfully
5. Generates artifacts - Creates test cases, E2E code, PRs, etc.

No API Server Needed

• Everything runs through Claude Code directly
• No need for separate microservice or Docker containers
• All agents executed via natural language commands
• Real-time interaction and feedback

Option 2: Cursor IDE Usage

Quick Start

1. Setup Prerequisites (see above sections):

   • Install Cursor IDE
   • Configure JIRA MCP access (if using Cursor's AI features)
   • Configure DeepWiki MCP
   • Configure GitHub CLI (gh)

2. Open Cursor IDE:

   cursor /path/to/ai_testgen
   

3. Use Cursor's AI Chat:

   • Open Cursor's AI chat panel (Cmd+L or Ctrl+L)
   • Execute test generation commands directly in the chat
   • Use Cursor's integrated terminal for command execution

4. Execute agents directly:

   # Generate test cases
   "Create test case for HIVE-2883"
   
   # Generate E2E code
   "Generate E2E case for HIVE-2883"
   
   # Execute E2E tests
   "Run E2E tests for HIVE-2883"
   
   # Complete end-to-end flow (executes all 3 agents in sequence)
   "Generate test cases and run them for HIVE-2883"
   # This automatically executes: test_case_generation → e2e_test_generation_openshift_private → test-executor
   
   # Create PR
   "Create PR for HIVE-2883 E2E tests"
   
   # Add QE comment
   "Add QE comment to HIVE-2883"
   

Cursor Advantages

• Integrated Development Environment: Full IDE experience with AI assistance
• Code Navigation: Easy file browsing and code navigation
• Terminal Integration: Built-in terminal for command execution
• AI-Powered Editing: Smart code completion and refactoring
• Project Management: Better project structure visualization

🔗 Integrations

JIRA Integration

• MCP Endpoint: jira-mcp-snowflake
• Purpose: Extract ticket requirements and details

DeepWiki Integration

• MCP Endpoint: DeepWiki MCP
• Purpose: Code change analysis and testing implications

GitHub Integration

• Repository Management: Automatic branch creation
• File Operations: Direct file creation in E2E repositories

📚 Documentation

For End Users

• QUICK_START.md - Quick start guide for team members
• PLUGIN_INSTALLATION.md - Detailed installation instructions
• EXAMPLE_USAGE.md - Usage examples and best practices

For Plugin Maintainers

• DISTRIBUTION_GUIDE.md - How to distribute the plugin to your team
• CLAUDE.md - Complete agent instructions and execution guide
• WORKFLOW_ORCHESTRATOR_GUIDE.md - Workflow orchestrator documentation
• STARTUP_CHECKLIST.md - Startup checklist for agents