Introduction

AWS Transform for mainframe is an agentic AI service that helps organizations analyze their legacy mainframe applications and generate both technical and business documentation. One of its key capabilities is generating Business Rules Extraction (BRE) outputs that capture the essential business logic and program flow embedded within mainframe application.

The extracted business logic and documentation can be used as input for migration and modernization the application. However, translating these technical BRE outputs into actionable user stories and comprehensive documentation can be time-consuming and error-prone when done manually. This article demonstrates how to leverage Amazon Q CLI with Model Context Protocol (MCP) servers to automate the generation of JIRA user stories and Confluence documentation directly from BRE outputs.

By the end of this walkthrough, you'll have a streamlined process that transforms complex mainframe business rules into well-structured user stories and documentation, accelerating your modernization efforts and improving stakeholder communication.

---

1. Create BRE documentation:

AWS Transform is available in your AWS account free of charge. Follow the steps here to submit the job to generate the business logic.

You may also refer the article to understand the outputs generated in BRE.

---

2. Setup Amazon Q CLI:

Amazon Q Developer is a generative artificial intelligence (AI) powered conversational assistant that can help you understand, build, extend, and operate AWS applications. We will use the CLI version to generate user stories and upload to JIRA and Confluence in an automated way.

Follow the steps here to install and configure Amazon Q CLI.

Before setting up the MCP server in Amazon Q CLI, we will need the API access token

---

3. Create Access token:

To enable automated interaction with JIRA and Confluence, you'll need API tokens.

In this demonstration, I am using personal Atlassian account and have admin access. Some of the operations need administrator access, so reach out to your organization Atlassian administrator for performing the tasks.

Generate API Token:

• Log into your JIRA or Confluence instance
• Go to Account Settings → Security → API tokens

• Click "Create API token"

• Provide a label (e.g., "amazon_q_cli_mcp")

• Copy and securely store the generated token

---

4. Create JIRA project and Confluence space

You can either use your existing JIRA project and Confluence space or create new ones.

Created a new JIRA project with name Mainframe Modernization -

Created a new Confluence space with name Mainframe Modernization -

---

5. Configure Amazon Q CLI MCP server:

• Model Context Protocol (MCP) servers extend Q CLI's capabilities. We'll configure servers for JIRA and Confluence integration.

• First, create the virtual environment: Use the venv module, to create a new virtual environment. Replace venv with your desired environment name (e.g., env, .venv).

python3 -m venv venv

• Activate the virtual environment: After creation, activate the environment. The activation command varies slightly depending on your operating system:

macOS/Linux:

source venv/bin/activate

Windows:

venv\Scripts\activate.bat

• In this demonstration we will use uvx command to invoke the MCP server. However, you may also use docker installation. Install uv using brew or pip command.

On macOS:

brew install uv

• Configure one or more MCP servers in a file called mcp.json. Store the configuration in home directory (e.g. ~/.aws/amazonq/mcp.json) and it is applied to all projects on the machine.

# Create MCP server configuration
cat > ~/.aws/amazonq/mcp.json << EOF
{
  "mcpServers": {
    "mcp-atlassian": {
      "command": "uvx",
      "args": [
        "mcp-atlassian",
        "--confluence-url=https://your-domain.atlassian.net",
        "--confluence-username=your.email@domain.com",
        "--confluence-token=your_api_token",
        "--jira-url=https://your-domain.atlassian.net",
        "--jira-username=your.email@domain.com",
        "--jira-token=your_api_token"
      ]
    }
  }
}
EOF

Note: Replace the url, username and token in the above configuration.

Test MCP server connectivity:

• Type q on command line to start Amazon Q CLI. At the start it will show a message - mcp_atlassian loaded to indicate the mcp server is connected successfully.

• Next, type /tools to see the list of tools activated. The built-in tools like fs_read, fs_write are activated by default in Amazon Q CLI. Under mcp_atlassian (MCP) you will see all the tool for JIRA and Confluence.

---

6. Generate User Stories and documentation:

• Refer to the article on how to generate and download the BRE outputs from AWS transform.

• Unzip the BRE output file and store it on the same machine where you have the Amazon Q CLI with MCP configured. The folder should be accessible to Amazon Q CLI.

• You can use natural language to converse with the agent in Amazon Q CLI. When chatting with the agent make sure to provide a context on what you are trying to achieve. Be very specific and detail on the ask.

• Example prompts for generating content. Modify as appropriate to your need.

Sample prompt for Confluence Documentation:

Create comprehensive Confluence documentation from the BRE outputs in ./bre-outputs/ with the following structure:

1. Business Rules Overview
2. Rule Definitions and Logic
3. Architecture diagram
4. API contracts
5. UI mockups
6. ..

Format the output in Confluence markup language and include diagrams where appropriate.

Sample prompt for User Stories creation:

Analyze the BRE files at ./bre-outputs/xx-xx.html and generate JIRA EPIC and user stories following this template and save them in locally:

Title: As a [user type], I want [functionality] so that [benefit]
Description: [Detailed description from BRE]
Acceptance Criteria:
- [Criterion 1 based on business rule conditions]
- [Criterion 2 based on business rule actions]
- [Additional criteria as needed]

Priority: [High/Medium/Low based on business rule complexity]
Story Points: [Estimate based on rule complexity]

---

7. Upload to Confluence and JIRA:

• With the MCP servers configured, you can directly upload generated content to JIRA and Confluence.

Example prompt to upload User Stories to JIRA:

Use the jira-integration MCP server to create user stories in project [PROJECT-KEY] with the generated content. Set the issue type to "Story" and assign appropriate labels like "mainframe-modernization".

Example prompt to upload Documentation to Confluence:

Use the confluence-integration MCP server to create a new page in space [SPACE-KEY] titled "Mainframe Business Rules Documentation" with the generated Confluence markup content.

• The MCP servers will handle the API calls and provide feedback on successful uploads, including links to the created items in JIRA and Confluence.

Documentation uploaded to Confluence

EPIC and User Stories created in JIRA:

---

This automated approach significantly reduces the manual effort required to translate technical BRE outputs into actionable development artifacts, ensuring consistency and completeness in your mainframe modernization documentation.

You may further integrate with github or other repository via MCP servers. Even automate code generation by pulling stories from JIRA backlog using MCP server, build and test locally, push the code to github using MCP server for more end-to-end modernization experience of legacy application.

---

References

[1] Setup and submit AWS Transform job

[2] BRE Output explained

[3] AWS Transform for mainframe

[4] Atlassian MCP