mcp-to-work-with-jira

An open-source Model Context Protocol (MCP) server to connect AI agents with your Atlassian Jira instance.

This library allows your AI agent to:

• Search Tickets: Find issues using JQL (e.g., "bugs assigned to me").
• Read Tickets: Get detailed information about specific issues.
• Utilize Skills: Use built-in prompts like summarize_ticket to generate standard reports.

Installation

You can run this directly using npx within your agent's configuration (e.g., in mcp.json or claude_desktop_config.json).

Prerequisites

You need a Jira API Token. Get it here.

Configuration

Add this to your MCP settings file:

{
  "mcpServers": {
    "jira": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-to-work-with-jira",
        "--jira-host=https://your-domain.atlassian.net",
        "[email protected]",
        "--jira-token=your-jira-api-token"
      ]
    }
  }
}

Features

Intelligent Agent Prompts (Skills)

This MCP server provides powerful, multi-phase agent workflows that act as specialized team members:

Features

Intelligent Agent Capabilities

This library provides a set of powerful, interconnected capabilities that work together to streamline your development workflow:

1. Analyze Architecture

• Pattern Discovery: Automatically detects architectural patterns (MFE, Microservices, Pub/Sub, Monorepo) across your repositories.
• Strategic Design: Generates comprehensive High-Level Design (HLD) documents with strategic recommendations and battle-tested improvements.
• Workflow Integration: Designed to work seamlessly with the mcp-to-work-with-github library to discover and analyze repositories accurately.

2. Create Sub-tasks

• Automated Breakdown: INTELLIGENTLY breaks down the High-Level Design into actionable implementation sub-tasks.
• Context Aware: Ensures each sub-task has the necessary context from the architectural analysis.
• Jira Integration: Directly creates sub-tasks in Jira linked to the parent feature ticket.

3. Implement Feature

• Adaptive Workflow: Adapts to ANY tech stack and discovered testing standards (e.g., Browser testing vs Unit tests).
• Cross-Task Awareness: Smartly explores sibling sub-tasks (e.g., checking Backend tasks while working on Frontend) to reuse schemas, types, and endpoints.
• Design Discovery: Automatically finds and analyzes design files (Figma, wireframes) from Jira attachments to guide implementation.

4. Fix Bug

• Root Cause Analysis: Deep dives into the codebase to understand WHY a bug exists, not just where.
• Impact Analysis: proactively checks for similar bugs and potential regressions across the system.
• Adaptive Testing: Automatically discovers the correct testing approach for the specific repository to verify fixes.

5. Generate Rules

• Codebase Analysis: Deeply analyzes your workspace structure, tech stack, and architectural patterns.
• Knowledge Extraction: Extracts actual coding standards, naming conventions, and common workflows directly from your existing code.
• AI-Native Documentation: Generates a concise, practical RULES.md file that serves as the de-facto standard for AI agents to understand and contribute to your codebase effectively.

Core Tools

• jira_search: Search for tickets using JQL.
• jira_get_issue: Get detailed ticket info (including sub-tasks, attachments, and parent links).
• jira_add_comment: specific tool to add comment to the ticket.
• jira_create_subtask: Create sub-tasks for implementation breakdown.

Development

This project follows strict Agent Rules defined in RULES.md.

1. Clone: git clone ...
2. Install: npm install
3. Build: npm run build
4. Dev: npm run start -- --jira-host=...

License

MIT