Skip to main content

What You'll Build

An Atlassian workflow assistant that uses Continue with the Atlassian Rovo MCP to:
  • Search and summarize Jira issues across projects
  • Find and digest Confluence documentation
  • Create and update Jira issues with natural language
  • Query Compass components and service dependencies
  • Automate Atlassian workflows with headless CLI runs

​
Prerequisites

Before starting, ensure you have:
The Atlassian Rovo MCP Server is currently in Beta. Core functionality is available, but some features are still under development.
For all options, first:
1

Install Continue CLI

npm i -g @continuedev/cli
2

Understand OAuth Flow

The Atlassian MCP uses OAuth 2.1 for authentication. The first time you connect, you’ll complete a browser-based authorization flow that respects your existing Atlassian permissions.
To use agents in headless mode, you need a Continue API key. All data access respects your existing Jira, Confluence, and Compass user permissions.

​
Choose Your Atlassian Agent

🚀 Two Specialized Agents

We’ve created two ready-to-use agents to help you get started fast - they come pre-configured with optimized prompts, rules, and the Atlassian MCP. You don’t have to use these agents, but they’re designed to give you the best experience right away. Choose the agent that matches your needs, or create your own custom agent.

​
Agent Quick Reference

AgentBest ForUse CasesLink
Jira AgentIssue ManagementSearch issues, create stories, sprint planning, status updates, bulk operationsView Agent
Confluence AgentDocumentationSearch pages, summarize docs, create/update pages, manage spacesView Agent
Cross-product workflows: Both agents can work with Jira, Confluence, and Compass. Choose based on your primary focus area, or create your own agent using the Atlassian MCP. The Atlassian MCP can work with Jira, Compass, or Confluence.
To use Atlassian MCP with Continue CLI, you need either:
  • Continue CLI Pro Plan with the models add-on, OR
  • Your own API keys added to Continue Hub secrets
The agent will automatically detect and use your configuration along with the Atlassian MCP for Jira, Confluence, and Compass operations.

​
Jira Workflows

Using the Jira Agent: All examples below use the Jira Agent which is optimized for issue management.
Use natural language to explore, triage, and manage Jira issues. The agent calls Atlassian MCP tools under the hood.
Running in headless mode: Add -p flag before your prompt and --auto flag at the end:
cn --agent continuedev/atlassian-continuous-ai-jira-agent -p "your prompt" --auto
Important: Complete browser OAuth authentication first before using headless mode.

​
Search and Discovery

Search for bugs across your projects.
cn --agent continuedev/atlassian-continuous-ai-jira-agent "Find all open bugs in Project Alpha"
Get a list of issues assigned to you.
cn --agent continuedev/atlassian-continuous-ai-jira-agent "Show me all issues assigned to me that are in progress"
Analyze sprint workload and priorities.
cn --agent continuedev/atlassian-continuous-ai-jira-agent "List all issues in the current sprint. Group by assignee and priority. Summarize the workload distribution and flag any overloaded team members."

​
Create and Update

Turn natural language into a Jira story.
cn --agent continuedev/atlassian-continuous-ai-jira-agent "Create a story titled 'Redesign onboarding flow' in Project Alpha. Description: We need to simplify the user registration process by reducing steps from 5 to 3. Target completion: Q3."
Create multiple issues from meeting notes or specs.
cn --agent continuedev/atlassian-continuous-ai-jira-agent "Create five Jira issues from these meeting notes: 1) Fix login timeout bug 2) Update API documentation 3) Implement dark mode toggle 4) Review security audit findings 5) Optimize database queries"
Transition issues with natural language.
cn --agent continuedev/atlassian-continuous-ai-jira-agent "Move PROJ-123 to In Progress and add a comment: Starting work on this today. ETA: end of week."

​
Confluence Workflows

Using the Confluence Agent: All examples below use the Confluence Agent which is optimized for documentation management.
Access, search, and manage your team’s documentation directly from Continue.
Running in headless mode: Add -p flag before your prompt and --auto flag at the end:
cn --agent continuedev/atlassian-continuous-ai-confluence-agent -p "your prompt" --auto
Important: Complete browser OAuth authentication first before using headless mode.

​
Search and Summarize

Get a quick summary of a Confluence page.
cn --agent continuedev/atlassian-continuous-ai-confluence-agent "Summarize the Q2 planning page from the Engineering space"
Search for specific information across Confluence.
cn --agent continuedev/atlassian-continuous-ai-confluence-agent "Find all pages about API authentication in our developer docs"
Discover what Confluence spaces you have access to.
cn --agent continuedev/atlassian-continuous-ai-confluence-agent "What Confluence spaces do I have access to?"

​
Create and Update

Generate a new Confluence page with structured content.
cn --agent continuedev/atlassian-continuous-ai-confluence-agent "Create a page titled 'Team Goals Q3' in the Engineering space. Include sections for: Objectives, Key Results, and Timeline."
Modify content on an existing page.
cn --agent continuedev/atlassian-continuous-ai-confluence-agent "Update the 'Onboarding Guide' page to add a new section about our code review process"

​
Compass Workflows

Using Either Agent: Compass workflows work with both agents. Use the Jira agent for component-to-issue workflows, or the Confluence agent for component-to-docs workflows.
Query and manage your service architecture with Compass integration.
Running in headless mode: Add -p flag before your prompt and --auto flag at the end. Complete browser OAuth authentication first before using headless mode.

​
Service Discovery

Understand service relationships.
cn --agent continuedev/atlassian-continuous-ai-jira-agent "What services depend on the api-gateway component?"
Register a new service in Compass.
cn --agent continuedev/atlassian-continuous-ai-jira-agent "Create a service component for the current repository. Use the package.json to infer details."
Import multiple components from structured data.
cn --agent continuedev/atlassian-continuous-ai-jira-agent "Import Compass components and custom fields from this CSV: [paste CSV data]"
Or reference a file:
cn --agent continuedev/atlassian-continuous-ai-jira-agent "Import Compass components from services.csv in the current directory"

​
Combined Workflows

Cross-Product Workflows: These workflows span multiple Atlassian products. Choose the agent based on your primary focus:
  • Use Jira Agent when the workflow is issue-centric
  • Use Confluence Agent when the workflow is documentation-centric
Integrate actions across Jira, Confluence, and Compass for powerful cross-product workflows.
Running in headless mode: Add -p flag before your prompt and --auto flag at the end. Complete browser OAuth authentication first before using headless mode.
Find documentation for a specific component.
# Using Confluence Agent for docs-centric workflow
cn --agent continuedev/atlassian-continuous-ai-confluence-agent "Fetch the Confluence documentation page linked to the user-authentication Compass component"
Create release documentation from completed work.
# Using Jira Agent to pull issue data and create in Confluence
cn --agent continuedev/atlassian-continuous-ai-jira-agent "Generate release notes for all Jira issues completed in Sprint 23. Create a Confluence page in the Release Notes space with: New features (grouped by epic), Bug fixes, Known issues, Deployment instructions"

​
Troubleshooting

  • “Your site admin must authorize this app”: A site admin needs to complete the OAuth flow first before other users can connect
  • Can’t complete browser authorization: Ensure you’re logged into the correct Atlassian Cloud site and have necessary permissions
  • Connection timeout: Increase connectionTimeout in your MCP config or check your network connection
  • “Access denied” or “Forbidden”: The MCP respects your existing Atlassian permissions. Verify you have access to the project/space/component you’re trying to access
  • Can’t create issues: Ensure you have “Create issues” permission in the target Jira project
  • Can’t edit Confluence pages: Verify you have edit permissions in the target Confluence space
  • Standard plan: Moderate usage thresholds
  • Premium/Enterprise plans: Higher quotas (1,000 requests/hour plus per-user limits)
  • If you hit rate limits, reduce query frequency or batch operations
  • Bulk operations may be constrained by rate limits
  • Custom Jira fields may not be recognized without explicit setup
  • Workspace/site switching not available within a single session
  • Check Atlassian Community for known issues

​
Admin Considerations

For Atlassian Site Admins

Installation: The Atlassian Rovo MCP is automatically installed on first OAuth authorization (JIT installation) - no Marketplace installation needed.Managing Access:
  • View connected apps in Admin Hub
  • Users can revoke access from their profile settings
  • Block user-installed apps via “user-installed apps” control in Admin Hub
Security: All traffic is encrypted via HTTPS (TLS 1.2+), and access respects existing user permissions.

​
What You’ve Built

After completing this guide, you have a complete AI-powered Atlassian workflow system that:
  • ✅ Uses natural language — Simple prompts for complex Atlassian operations
  • ✅ Spans multiple products — Seamlessly works across Jira, Confluence, and Compass
  • ✅ Respects permissions — Secure OAuth-based access with existing role enforcement
  • ✅ Runs headlessly — Integrate into CI/CD pipelines and automation scripts

Continuous AI

Your Atlassian workflows now operate at Level 2 Continuous AI - AI handles routine project management tasks with human oversight through review and approval.

​
Next Steps

  1. Explore your projects - Try the Jira search and triage prompts
  2. Organize documentation - Use Confluence workflows to summarize and create pages
  3. Map your services - Query Compass for service dependencies
  4. Customize prompts - Tailor workflows to your team’s specific needs
  5. Monitor usage - Track how AI improves your Atlassian workflows

​
Additional Resources

Atlassian Rovo MCP

Official Atlassian MCP documentation

MCP in Continue

How MCP works with Continue agents

Continue Hub

Create and manage your agents

Atlassian Community

Get help and share feedback

​
Example Use Cases

Release Management

Automate release notes generation, link issues to documentation, and track deployment status across Jira and Confluence.

Sprint Planning

Analyze workload distribution, identify blockers, and generate sprint summaries automatically.

Service Ownership

Map service dependencies in Compass, link to technical documentation in Confluence, and track related issues in Jira.

Onboarding Automation

Create standardized Jira onboarding tickets, link to Confluence guides, and track new hire progress.