Jira MCP Server

A Model Context Protocol (MCP) server for Jira Cloud, built with Rust. This server enables AI assistants to manage issues, sprints, and boards directly through a standardized interface.

Features

• 🚀 Full Jira API Support: Management of Projects, Issues, Sprints, Boards, and Backlog.
• 🧠 Smart Context & Instructions: Automatic ADF conversion and hardcoded AI directives for zero-hallucination project metadata validation.
• ⚡ High Performance & Caching: Fast native Rust implementation with thread-safe in-memory caching for Issue Types and Custom Fields.
• 🔍 Field Filtering: Reduces token usage (70-90%) via smart filters (minimal, basic, standard, detailed).
• 🌐 Universal AI Compatibility: Native JSON schema sanitization ensures seamless integration with Claude, OpenAI, and Gemini.
• 🎯 Automatic Detection: Identifies custom fields like "Story Points" without manual configuration.

Installation

Prerequisites

• Rust 1.80 or higher.
• A Jira Cloud account and an API Token.

Build

git clone https://github.com/rafael-arreola/mcp-jira-rs
cd mcp-jira-rs
cargo build --release

The binary will be available at target/release/jira-rs.

Configuration

Environment Variables

The server requires the following environment variables:

JIRA_WORKSPACE="your-subdomain"      # e.g., "mycompany" for mycompany.atlassian.net
JIRA_USERNAME="your@email.com"        # Your Atlassian account email
JIRA_TOKEN="your-api-token"           # Generated at id.atlassian.com

Claude Desktop Configuration

Add this to your claude_desktop_config.json file:

{
  "mcpServers": {
    "jira": {
      "command": "npx",
      "args": ["-y", "@rafael-arreola/jira-rs@latest"],
      "env": {
        "JIRA_WORKSPACE": "your-subdomain",
        "JIRA_USERNAME": "your@email.com",
        "JIRA_TOKEN": "your-api-token"
      }
    }
  }
}

Available Tools (23)

🎫 Issue Management

💬 Content and Links

🏃 Agile Operations

🔍 Discovery

Usage Examples

Discover project metadata before creating issues

To avoid hallucinations and ensure you are using the correct issueType names (e.g., "Story", "Bug") and required fields for a specific project, use project_get_metadata first:

{
  "projectKey": "PROJ"
}

Retrieve an issue while saving tokens

To avoid saturating the AI context, use the filter parameter:

{
  "issueKey": "PROJ-123",
  "filter": "basic"
}

Available presets: minimal, basic, standard, detailed.

Create an issue with story points

Use storyPoints for Classic (Company-managed) projects and storyPointEstimate for Next-Gen (Team-managed) projects.

{
  "projectKey": "PROJ",
  "summary": "Implement OAuth2 authentication",
  "issueType": "Story",
  "storyPoints": 5,
  "storyPointEstimate": 3
}

Set Story Points

Use issue_set_story_points to update the estimation of an existing issue. The tool automatically detects if the project uses "Story Points" or "Story point estimate".

{
  "issueKey": "PROJ-123",
  "storyPoints": 8
}

Change Issue Type

Use issue_edit_details to convert an issue to a different type (e.g., from Task to Bug):

{
  "issueKey": "PROJ-123",
  "issueType": "Bug"
}

Link an existing Story to an Epic

Use issue_set_parent to organize Stories under Epics:

{
  "issueKey": "PROJ-123",
  "parentKey": "PROJ-100"
}

To remove the link, pass an empty parentKey:

{
  "issueKey": "PROJ-123",
  "parentKey": ""
}

Troubleshooting

• Error 401/403: Verify that JIRA_TOKEN is an API Token and not your personal password.
• Fields not found: If a custom field does not update, use fields_list to find its actual ID (e.g., customfield_10016).

Contributing

If you wish to add a tool, please add the corresponding DTO in src/domains/ and the implementation in src/jira.rs using the #[rmcp::tool] macro.