Skip to main content
Glama

db3-atlassian-mcp

Lightweight MCP server for Jira, Confluence, and Bitbucket — read, create, and update from your AI IDE.

Install

Step 1: Get an API Token

Go to https://id.atlassian.com/manage-profile/security/api-tokens and create a new token.

Step 2: Create your .env file

mkdir -p ~/.db3-atlassian-mcp

cat > ~/.db3-atlassian-mcp/.env << EOF
JIRA_BASE_URL=https://yourcompany.atlassian.net
JIRA_USER=you@yourcompany.com
JIRA_API_KEY=your-api-token-here

# Optional: Bitbucket Cloud
BITBUCKET_WORKSPACE=your-workspace
BITBUCKET_USER=you@yourcompany.com
BITBUCKET_API_TOKEN=your-bitbucket-api-token

# Optional: Enable additional Bitbucket tool tiers
# ENABLE_BITBUCKET_PR_TOOLS=true
# ENABLE_BITBUCKET_ADMIN_TOOLS=true
EOF

Step 3: Add to your IDE's MCP config

Tip: If you skip Step 2, you can put your credentials directly in the env block below instead of using a .env file. Either approach works.

Kiro

You can ask Kiro to do this for you: "Add db3-atlassian-mcp to my MCP config using uvx with my Atlassian credentials"

Or manually open ~/.kiro/settings/mcp.json and add inside "mcpServers":

"db3.atlassian-mcp": {
  "command": "uvx",
  "args": ["db3-atlassian-mcp@latest"],
  "disabled": false,
  "autoApprove": [],
  "env": {
    "JIRA_BASE_URL": "https://yourcompany.atlassian.net",
    "JIRA_USER": "you@yourcompany.com",
    "JIRA_API_KEY": "your-api-token-here"
  }
}

The env block is optional if you already created a .env file in Step 2.

VS Code / Cursor / Windsurf

Add to .vscode/mcp.json in your workspace (or global settings):

{
  "mcpServers": {
    "db3.atlassian-mcp": {
      "command": "uvx",
      "args": ["db3-atlassian-mcp@latest"],
      "env": {
        "JIRA_BASE_URL": "https://yourcompany.atlassian.net",
        "JIRA_USER": "you@yourcompany.com",
        "JIRA_API_KEY": "your-api-token-here"
      }
    }
  }
}

Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "db3.atlassian-mcp": {
      "command": "uvx",
      "args": ["db3-atlassian-mcp@latest"],
      "env": {
        "JIRA_BASE_URL": "https://yourcompany.atlassian.net",
        "JIRA_USER": "you@yourcompany.com",
        "JIRA_API_KEY": "your-api-token-here"
      }
    }
  }
}

Step 4: Restart your IDE

The MCP server will appear in your sidebar as db3.atlassian-mcp.

Once connected, ask Kiro something like "Read Jira ticket PROJECT-123" to verify it's working.

Related MCP server: Jira & Confluence MCP Server

What You Can Do

Jira

  • Fetch any ticket by key

  • Search tickets with JQL

  • View sub-tasks for a parent ticket

  • Create new tickets with optional parent for sub-task linking

  • Update tickets — change fields, transition status, add comments

  • Assign tickets by name or email (no account IDs needed)

  • Attach files to tickets

Confluence

  • Read any page by pasting its URL or page ID

  • Search pages with CQL

  • Browse all pages in a space

  • Create new pages (with optional parent page)

  • Update existing pages with rich formatting

Bitbucket Cloud (core — always on)

  • List and search repositories in a workspace

  • List branches for a repository

  • List and view pull requests

Bitbucket Cloud (opt-in via ENABLE_BITBUCKET_PR_TOOLS)

  • List and view commits

  • View PR diffs, file diffs, and diffstats

  • Create pull requests

  • Add, reply to, update, and delete PR comments

  • Add inline comments on specific diff lines

Bitbucket Cloud (opt-in via ENABLE_BITBUCKET_ADMIN_TOOLS)

  • View build statuses on commits and PRs

  • List pipeline runs

  • Resolve and reopen PR comment threads

  • Create, update, and delete PR tasks

  • Fetch repository metadata

  • List branches and commits

  • Fetch commit details and build statuses

  • List recent Bitbucket Pipelines runs

  • List and inspect pull requests

  • Fetch pull request diffs, diffstats, comments, and activity

  • Fetch pull request build statuses

  • Summarize pull request status with reviewers, approvals, build status results, and changed files

  • Create pull requests

  • Add pull request comments

Configuration

Credentials

The server reads credentials from a .env file. It searches in this order:

  1. ~/.db3-atlassian-mcp/.env (recommended)

  2. .env in the current working directory

You can also pass credentials via your MCP config's env block (see Step 3 above).

Bitbucket Tool Tiers

By default, only core Bitbucket tools are enabled (repos, branches, PRs) to keep the tool count low. Enable additional tiers via environment variables:

Variable

Default

Tools Added

(core — always on)

—

list_bitbucket_repos, get_bitbucket_repo, search_bitbucket_repos, list_bitbucket_branches, list_bitbucket_pull_requests, get_bitbucket_pull_request

ENABLE_BITBUCKET_PR_TOOLS

false

Commits, diffs, diffstats, PR comments, inline comments, replies, activity, status, create PR (+15 tools)

ENABLE_BITBUCKET_ADMIN_TOOLS

false

Build statuses, pipelines, resolve/reopen comments, PR tasks (+10 tools)

Example with all tiers enabled:

"db3.atlassian-mcp": {
  "command": "uvx",
  "args": ["db3-atlassian-mcp@latest"],
  "env": {
    "JIRA_BASE_URL": "https://yourcompany.atlassian.net",
    "JIRA_USER": "you@yourcompany.com",
    "JIRA_API_KEY": "your-api-token-here",
    "ENABLE_BITBUCKET_PR_TOOLS": "true",
    "ENABLE_BITBUCKET_ADMIN_TOOLS": "true"
  }
}

Bitbucket Cloud

Bitbucket support is optional. Set these values only if you want Bitbucket tools:

BITBUCKET_WORKSPACE=your-workspace
BITBUCKET_USER=you@yourcompany.com
BITBUCKET_API_TOKEN=your-bitbucket-api-token

BITBUCKET_WORKSPACE lets tool calls omit the workspace argument. For bitbucket.org/your-workspace/..., use:

BITBUCKET_WORKSPACE=your-workspace

Use a Bitbucket Cloud API token scoped to the operations you need. Read-only tools need repository and pull request read scopes. Pipeline status tools need pipeline read scope. Creating pull requests or comments needs write access for pull requests.

Alternative: Install from Source

git clone https://github.com/db3net/db3-atlassian-mcp.git
cd db3-atlassian-mcp
bash install.sh

The installer prompts for your credentials, sets up a Python venv, validates the connection, and configures Kiro automatically.

Updating

If using uvx with @latest, it automatically pulls the newest version each time your IDE starts. If installed from source, run the installer again.

Version Pinning

Using @latest is convenient but means updates are applied automatically. For production or security-sensitive environments, pin to a specific version:

"args": ["db3-atlassian-mcp==1.1.0"]

This way you only get updates when you explicitly change the version number. Check PyPI for available versions.

Uninstall

rm -rf ~/.db3-atlassian-mcp

Then remove the "db3.atlassian-mcp" entry from your IDE's MCP config.

Troubleshooting

Server shows "connection failed" or ENOENT

  • Make sure uv is installed: pip install uv or brew install uv

  • Verify uvx is on your PATH: which uvx

  • Check that Python 3.10+ is available: python3 --version

401 Unauthorized

.env file not found

  • Confirm the file exists: cat ~/.db3-atlassian-mcp/.env

  • Make sure all three variables are set: JIRA_BASE_URL, JIRA_USER, JIRA_API_KEY

  • The URL should not have a trailing slash

Tools not showing up

  • Restart your IDE after adding the MCP config

  • Check your IDE's MCP server logs for errors

  • In Kiro: look at the MCP Servers panel in the sidebar and click reconnect

License

MIT

Available Tools

20 tools
attach_fileA

Attach a file to a Jira ticket. Provide the issue key and absolute file path.

Inputs: file_path (required), issue_key (required). Returns: result

create_confluence_pageC

Create a new Confluence page in a space. Optionally specify a parent page ID.

Inputs: body (required), title (required), space_key (required), parent_id. Returns: result

create_ticketB

Create a new Jira ticket. Optionally pass custom_fields for Jira customfield_* or other fields.

Inputs: summary (required), issue_type (required), project_key (required), assignee, +4 more. Returns: result

get_bitbucket_pull_requestC

Fetch Bitbucket Cloud pull request details.

Inputs: repo_slug (required), pull_request_id (required), workspace. Returns: result

get_bitbucket_repoC

Fetch Bitbucket Cloud repository metadata.

Inputs: repo_slug (required), workspace. Returns: result

get_child_issuesA

Get all child/sub-task issues for a given parent ticket.

Inputs: issue_key (required). Returns: result

get_confluence_pageA

Fetch a Confluence page by numeric ID or full URL. Returns title, space, version, and body as plain text.

Inputs: page_id (required). Returns: result

get_create_fieldsA

List Jira fields available when creating a ticket, including IDs, names, required flags, schemas, and allowed values.

Inputs: issue_type (required), project_key (required). Returns: result

get_space_pagesC

List pages in a Confluence space. Example: space_key='SECENG'

Inputs: space_key (required), max_results. Returns: result

get_ticketA

Fetch a Jira ticket by key (e.g. INFOSEC-2239). Returns standard fields, description, and populated custom fields.

Inputs: issue_key (required), include_noisy_custom_fields. Returns: result

list_bitbucket_branchesA

List Bitbucket Cloud branches for a repository, optionally filtered by name.

Inputs: repo_slug (required), query, workspace, max_results. Returns: result

list_bitbucket_pull_requestsC

List Bitbucket Cloud pull requests for a repository.

Inputs: repo_slug (required), state, workspace, max_results. Returns: result

list_bitbucket_reposC

List Bitbucket Cloud repositories in a workspace.

Inputs: role, workspace, max_results. Returns: result

move_confluence_pageA

Move a Confluence page under a new parent page. Provide the page ID and the target parent page ID.

Inputs: page_id (required), target_parent_id (required). Returns: result

search_bitbucket_reposC

Search Bitbucket Cloud repositories by slug, name, or description.

Inputs: query (required), workspace, max_results. Returns: result

search_confluenceB

Search Confluence pages using CQL. Example: 'space = SECENG AND type = page'

Inputs: cql (required), max_results. Returns: result

search_ticketsA

Search Jira tickets using JQL. Example: 'project = INFOSEC AND status = Open'

Inputs: jql (required), max_results. Returns: result

update_confluence_pageC

Update an existing Confluence page's content.

Inputs: body (required), title (required), page_id (required). Returns: result

update_ticketA

Update a Jira ticket. Can change standard fields, custom_fields, transition status, and/or add a comment.

Inputs: issue_key (required), comment, summary, assignee, +4 more. Returns: result

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    B3.4/5.0

    Scored across 20 tools

    Disambiguation5/5

    Each tool targets a distinct resource and action within Confluence, Bitbucket, or Jira. No two tools have overlapping functionality, and the descriptions clearly differentiate their purposes.

    Naming Consistency5/5

    All tool names follow a consistent verb_noun pattern in snake_case (e.g., get_space_pages, create_ticket, list_bitbucket_repos). No mixing of conventions or unclear verbs.

    Tool Count4/5

    20 tools cover three distinct Atlassian products, which justifies a higher count. While slightly above the ideal 3-15 range, the scope is well-defined and each tool serves a clear purpose.

    Completeness4/5

    The tool set covers core CRUD operations for Jira, Confluence, and Bitbucket, along with search and linking. Minor gaps exist, such as missing delete operations and no Bitbucket repo/PR creation, but the most common workflows are supported.

    Maintenance

    ActivityInactive
    ResponsivenessNo recent community issues

    Related MCP Connectors

    Related MCP Servers