Use the Go-based Jira MCP to Read Issues and Development Links

Who It’s For

If you already use Atlassian Jira and want to view issue status, assignee, description, subtasks, and available transitions in a conversation, or inspect linked branches, pull requests, and commits, this service is a good fit. It is a Jira MCP service with an opinionated usage path.

You need a local machine that can run go install, plus ATLASSIAN_HOST, ATLASSIAN_EMAIL, and ATLASSIAN_TOKEN. Put the token only in environment variables; do not enter it in a conversation. The official documentation does not provide a client list, so do not invent names.

It is not suitable if you do not have a token or want to start creating or editing issues in a real project. The available material describes reading issues and development links, but gives no call format for creating or updating them. It is also not suitable if you need a complete field table or startup parameters: the official documentation does not provide them.

Official documentation:

https://github.com/nguyenvanduocit/jira-mcp

Make Your First Attempt in 3 Minutes

Follow one path: install jira-cli with the official go install command, export the three environment variables, and then read an existing issue.

Prerequisites: Go is already installed locally. Prepare your own site host, email address, and API token. Do not use the example domain, email address, or token value. The official documentation does not specify MCP startup parameters, so do not invent serve or a configuration file name for your first attempt.

Run this first:

go install github.com/nguyenvanduocit/jira-mcp/cmd/jira-cli@latest

Then export:

export ATLASSIAN_HOST=your-instance.atlassian.net
export [email protected]
export ATLASSIAN_TOKEN=your-api-token

Replace the three values with your real values. This is the official installation command to use for the first attempt.

For the first tool call, use jira_get_issue to read an issue whose key you already know. The page says it can return the status, assignee, description, subtasks, and available transitions. This reads Jira; it does not modify Jira. The official documentation does not provide an example issue key, so do not invent one.

Success signal: the issue detail fields are returned rather than an authentication failure. The token will be sent with outbound requests; do not interpret this as a no-network or no-privacy boundary.

Use Cases

Case one: read issue details. Call jira_get_issue. On success, you will see the status, assignee, description, subtasks, and available transitions. The official documentation does not specify parameter names or key format, so do not invent fields when those details are missing.

Case two: read development links. Call jira_get_development_information. On success, you will see branches, pull requests, and commits connected to the issue through integrations with development tools such as GitHub, GitLab, or Bitbucket. This is still a read operation. The official documentation does not specify failure codes or a revocation procedure. If you only need to verify the installation, stop after case one.

Usage Boundaries

The primary installation is go install github.com/nguyenvanduocit/jira-mcp/cmd/jira-cli@latest. The page also includes just install-cli; do not treat it as a second primary installation path.

You must provide ATLASSIAN_HOST, ATLASSIAN_EMAIL, and ATLASSIAN_TOKEN yourself. The official documentation does not provide the URL for obtaining a token, and you should not invent startup parameters or a configuration file format.

The tools explicitly named in the available material are jira_get_issue and jira_get_development_information. Do not call creation, update, or transition interfaces that are not documented. Reading still accesses your Jira site.