Atlassian Rovo MCP Server

The Atlassian Rovo MCP Server is an official, cloud-hosted Model Context Protocol (MCP) server that acts as a bridge between Atlassian Cloud sites and AI tools. It provides secure, real-time access to data within Jira, Confluence, Jira Service Management, Bitbucket, and Compass.

Key Capabilities:

• Search & Summarize: Query content across Jira, Confluence, JSM, and Bitbucket without switching tools.
• Create & Update: Use natural language to generate or modify issues and pages.
• Automate Workflows: Automate tasks like generating Jira tickets from meeting notes or specifications.

Security Model: Authentication is handled via OAuth 2.1 or API tokens. All actions performed by the AI tool respect the user's existing Atlassian access controls and permissions.

The capture-tasks-from-meeting-notes skill automates the process of turning unstructured meeting notes (from Confluence or pasted text) into actionable Jira tasks. It identifies action items, extracts assignees, looks up Jira account IDs, and creates tasks with appropriate context and descriptions.

Use this skill when:

• You need to create Jira tasks/tickets from meeting notes.
• You need to extract action items from Confluence pages or text.
• You need to parse meeting notes for assigned tasks and generate tickets for team members.

--- 
name: capture-tasks-from-meeting-notes
description: "Analyze meeting notes to find action items and create Jira tasks for assigned work..."
---

The triage-issue skill is designed to intelligently triage bug reports and error messages. It automates the process of searching Jira for duplicates, identifying similar past issues, and helping create well-structured bug tickets or adding context to existing ones. This skill is ideal for reducing manual duplicate checking and ensuring bugs are documented with relevant historical context.

Use this skill when:

• You need to triage an error message or bug report.
• You want to check if an issue is a duplicate.
• You need to find similar past issues.
• You want to create a new bug ticket with proper context.
• You need to add information to an existing Jira ticket.

This skill is a specialized workflow designed exclusively for converting meeting action items into Jira tasks.

Use this skill ONLY when:

• You have meeting notes (either as text or a Confluence URL).
• You need to transform identified action items into formal Jira issues.

DO NOT use this skill for:

• Summarizing meetings (use a general summarization skill).
• Finding or searching for meeting notes (use the search skill).
• Creating calendar events.
• Sending meeting notes via email.
• General note-taking.

The server operates via a client-server model where clients connect to specific endpoints to stream contextual data and receive real-time responses.

1. Connection: A client connects to the server endpoint.
   • Recommended: https://mcp.atlassian.com/v1/mcp/authv2 (for OAuth 2.1 flows).
   • Alternative: https://mcp.atlassian.com/v1/mcp (for API token configurations).
2. Authentication: The system triggers either a secure browser-based OAuth 2.1 flow or uses API token authentication.
3. Data Flow: Once authorized, the client streams contextual data and receives responses from connected Atlassian products.

Note on Endpoints: The legacy Server-Sent Events endpoint (https://mcp.atlassian.com/v1/sse) is supported but deprecated. Custom clients should be updated to use /mcp or /mcp/authv2.

The skill identifies action items using several common text patterns. When parsing, extract the Assignee Name, the Task Description, and optional Context.

Supported Patterns:

1. @mention format: @Sarah to create user stories
2. Name + action verb: Mike will update architecture doc or Lisa should review mockups
3. Structured Action: Action: Sarah - create user stories or Action Item: Mike - update architecture
4. TODO format: TODO: Create user stories (Sarah) or TODO: Update docs - Mike
5. Bullet lists: - Sarah: create user stories or - Mike - update architecture

Effective descriptions provide the context necessary for developers to debug without back-and-forth communication.

✅ Do:

• Use Structured Formats: Use headers to organize information.
• Include Complete Error Messages: Wrap them in code blocks.
• Provide Context: Include environment, version, and timestamp.
• List Concrete Steps: Provide a numbered list of reproduction steps.
• Quantify Impact: State how many users or how frequently the issue occurs.
• Add Relevant Logs: Use code blocks for log excerpts.
• Reference Related Issues: Provide links to similar tickets.

❌ Avoid:

• Pasting raw stack traces without context.
• Vague descriptions like it doesn't work.
• Missing environment information or reproduction steps.
• Formatting errors (e.g., forgetting to use code blocks for logs/errors).

Regardless of the specific template used, every Epic must contain these five core elements to ensure clarity and alignment across teams:

1. Clear objective: A concise statement so anyone understands what is being built or fixed.
2. Source link: A direct link to the authoritative Confluence spec or design document.
3. Success criteria: Measurable outcomes that define when the work is considered 'done'.
4. Scope clarity: Explicitly stating what IS and IS NOT included in the work.
5. Context: Sufficient background information for new team members to understand the importance of the work.

Avoid extracting the following as action items:

• Discussion Notes: Statements describing what was talked about rather than assigned (e.g., John mentioned the documentation needs updating).
• General Statements: Statements lacking a specific assignee (e.g., Someone should create a report).
• Past Actions: Statements describing completed work (e.g., Sarah created the report).

Adjust the focus of your ticket description based on the nature of the work: