Jira Assistant

What it does

The Jira Assistant helps AI agents interact with Atlassian Jira using Model Context Protocol (MCP). It enables users to search for existing issues, create new ones (Tasks, Epics, and Subtasks), update issue details, transition issue statuses (e.g., To Do → In Progress → Done), add comments, manage assignees, and query issues based on specific criteria. This skill streamlines Jira workflows and reduces manual effort.

When to use it

• Searching for existing Jira issues or tasks.
• Creating new Jira issues of different types (Task, Epic, Subtask).
• Updating the status of an issue (e.g., moving from "To Do" to "In Progress").
• Adding comments and managing assignees on Jira issues.
• Querying for issues that meet specific criteria, such as those assigned to you or with a high priority.

Key capabilities

• Issue Search: Uses both natural language searches (via Rovo Search) and JQL queries to find Jira issues.
• Issue Creation: Creates new Jira issues, guiding the user through necessary steps including viewing issue types and required fields.
• Issue Updates: Allows for updating existing issue details and transitioning statuses.
• Configuration Detection: Automatically detects Jira project configuration from workspace rules or by querying available projects if no rules are found. It prompts the user to specify a project key if needed.
• JQL Support: Supports JQL queries for precise filtering, requiring inclusion of project = {PROJECT_KEY} in all queries.

Example prompts

• "Search for issues assigned to me."
• "Create a new Task issue in the KAN project with the summary 'Implement user authentication'."
• "What are the available issue types for the PROJ project?"
• “Find bugs in progress.”

Tips & gotchas

• Configuration Required: To use this skill effectively, configure it by creating a .cursor/rules/jira-config.mdc file with your Jira project details (Project Key, Cloud ID, URL, Board URL).
• JQL Requirement: When using JQL queries, always include project = {PROJECT_KEY} to ensure accurate results.
• Natural Language First: Start with natural language searches for general queries; use JQL only when precise filtering is needed.