@leilashellroot/jira-data-center-mcp

MCP server for Jira Server/Data Center. Connects AI assistants to self-hosted Jira instances.

Features

• Issue management — get, create, update, delete, search with JQL
• Comments — add, edit, delete, list
• Transitions — list available transitions, transition issues
• WebLinks / Remote links — read and manage external URLs (GitLab MRs, Confluence pages, docs)
• Issue links — create and remove links between issues (Blocks, Relates, etc.)
• Agile — boards, sprints, sprint issues
• Worklogs — get and add worklog entries
• Changelog — view issue change history
• Attachments — get metadata, add files, delete
• jira_get_issue_context — single call that returns everything: summary, description, comments, attachments, issuelinks, weblinks, changelog, parent, subtasks

Installation

npm install -g @leilashellroot/jira-data-center-mcp

Configuration

The simplest configuration uses the existing environment variables:

export JIRA_BASE_URL=https://jira.example.com
export JIRA_PAT=your-personal-access-token

The upstream-compatible names are also supported:

export JIRA_HOST=jira.example.com
export JIRA_API_TOKEN=your-personal-access-token

JIRA_HOST assumes HTTPS. Use JIRA_API_BASE_PATH for a full URL such as https://jira.example.com/rest when Jira is hosted below a context path.

Setup CLI

The setup command validates credentials and stores configuration in ~/.atlassian-dc-mcp/jira.env with mode 0600 on POSIX systems:

npx @leilashellroot/jira-data-center-mcp setup

It also supports scripted setup:

npx @leilashellroot/jira-data-center-mcp setup --non-interactive \
  --host jira.example.com \
  --token "$JIRA_TOKEN"

Configuration precedence is process environment, ATLASSIAN_DC_MCP_CONFIG_FILE or the current directory .env, the home configuration file, then the macOS Keychain token when available. Process environment variables always win.

JIRA_DEFAULT_PAGE_SIZE defaults to 25. Set ATLASSIAN_DC_MCP_REQUEST_TIMEOUT_MS to change the default 30-second Jira request timeout.

Generate a PAT in Jira: Profile → Personal Access Tokens → Create token.

Usage

Direct

npx @leilashellroot/jira-data-center-mcp

Claude Desktop

{
  "mcpServers": {
    "jira": {
      "command": "npx",
      "args": ["-y", "@leilashellroot/jira-data-center-mcp"],
      "env": {
        "JIRA_BASE_URL": "https://jira.example.com",
        "JIRA_PAT": "your-token"
      }
    }
  }
}

OpenCode / Cursor

{
  "mcp": {
    "jira": {
      "type": "local",
      "command": ["npx", "-y", "@leilashellroot/jira-data-center-mcp"],
      "environment": {
        "JIRA_BASE_URL": "https://jira.example.com",
        "JIRA_PAT": "your-token"
      },
      "enabled": true
    }
  }
}

Remote SSE

Start the server with the legacy SSE transport:

MCP_TRANSPORT=sse MCP_HOST=0.0.0.0 MCP_PORT=3000 npx @leilashellroot/jira-data-center-mcp

Connect an SSE-compatible MCP client to http://localhost:3000/sse. The client will receive the POST endpoint from the SSE endpoint event.

Remote Streamable HTTP

Start the current MCP remote transport:

MCP_TRANSPORT=streamable-http MCP_HOST=0.0.0.0 MCP_PORT=3000 npx @leilashellroot/jira-data-center-mcp

Connect a Streamable HTTP MCP client to http://localhost:3000/mcp.

For OpenCode or Cursor, use a remote server entry instead of a local command:

{
  "mcp": {
    "jira": {
      "type": "remote",
      "url": "http://localhost:3000/mcp",
      "enabled": true
    }
  }
}

Remote configuration variables:

| Variable | Default | Description | |---|---|---| | MCP_TRANSPORT | stdio | stdio, sse, or streamable-http | | MCP_HOST | 127.0.0.1 | HTTP bind host; use 0.0.0.0 for network access | | MCP_PORT | 3000 | HTTP port | | MCP_HTTP_PATH | /mcp | Streamable HTTP endpoint | | MCP_SSE_PATH | /sse | SSE connection endpoint | | MCP_MESSAGES_PATH | /messages | SSE message POST endpoint | | MCP_ALLOWED_HOSTS | unset | Comma-separated Host values for DNS rebinding protection | | MCP_ALLOWED_ORIGINS | unset | Comma-separated Origin values for DNS rebinding protection | | MCP_CORS_ORIGIN | unset | Optional Access-Control-Allow-Origin value |

When binding beyond localhost, put the server behind authentication and HTTPS. The Jira PAT is used by the server for every connected client.

Tools

| Tool | Description | |---|---| | jira_get_issue | Get issue details with optional field projections and expansions | | jira_get_issue_context | Get comprehensive context in one call | | jira_search_issues | Search with JQL, pagination, field projections, and expansions | | jira_create_issue | Create a new issue | | jira_update_issue | Update an existing issue | | jira_delete_issue | Delete an issue | | jira_add_comment | Add a comment | | jira_edit_comment | Edit a comment | | jira_delete_comment | Delete a comment | | jira_get_comments | List comments with pagination and rendered-body expansion | | jira_get_projects | List projects | | jira_get_project | Get project details | | jira_get_issue_types | Get issue types for a project | | jira_get_project_versions | Get project versions | | jira_assign_issue | Assign to a user | | jira_get_current_user | Get authenticated user | | jira_search_users | Search users | | jira_get_transitions | List available transitions | | jira_transition_issue | Transition to a new status with fields and update operations | | jira_get_worklogs | Get worklog entries | | jira_add_worklog | Add a worklog | | jira_get_changelog | View change history | | jira_get_attachment | Get attachment metadata | | jira_add_attachment | Add a file attachment | | jira_delete_attachment | Delete an attachment | | jira_get_link_types | List issue link types | | jira_create_issue_link | Link two issues | | jira_remove_issue_link | Remove issue link | | jira_get_remote_links | Get weblinks/remote links | | jira_create_remote_link | Create a weblink | | jira_delete_remote_link | Delete a weblink | | jira_get_agile_boards | List agile boards | | jira_get_sprints | List sprints | | jira_get_sprint_issues | Get sprint issues | | jira_get_board_issues | Get board issues | | jira_search_fields | Search custom fields | | jira_get_field_options | Get custom field options | | jira_get_components | Get project components | | jira_get_create_meta | Get issue creation metadata | | jira_get_dev_status | Get development summary or detailed pull requests, repositories, and branches | | jira_get_filter | Resolve filter ID to JQL, name, view URL | | jira_get_filter_issues | Get filter metadata + matching issues |

Requirements

• Node.js >= 18
• Jira Server/Data Center 7.0+

Acknowledgements

This project incorporates feature ideas and workflows from the community-maintained Atlassian Data Center MCP project by b1ff. In particular, the upstream-compatible configuration sources and setup CLI, configurable page sizes and field projections, detailed development-status queries, flexible transition payloads, request timeouts, and credential validation were adapted from that project. The upstream project remains independently licensed under the MIT License.

Enhanced tool parameters

• jira_search_issues: startAt, maxResults, fields, and expand are optional.
• jira_get_issue: fields and expand are optional; compact fields are returned by default to reduce response size.
• jira_get_comments: startAt, maxResults, and expand are optional.
• jira_update_issue: supports issueTypeId and Jira update operations in addition to the existing standard fields.
• jira_transition_issue: supports transition-screen fields and arbitrary customFields/update payloads.
• jira_get_dev_status: set detail to true and choose dataType (pullrequest, repository, or branch) and applicationType (stash, bitbucket, github, or githube).

License

MIT