How Atlassian and GitHub connect planning, delivery, and code
Atlassian and GitHub offer four ways to connect code, work, and knowledge: GitHub for Atlassian, the Teamwork Graph connector for GitHub, GitHub Copilot for Jira, and the Atlassian MCP Server.
Instead of treating code repositories and project management as separate silos, connecting GitHub through Atlassian’s Teamwork Graph brings pull requests, commit histories, and issue tracking into one unified workflow. Agents don’t just read the state of your code—they actively write results back as structured context, enriching organizational knowledge with every commit.
The end-to-end experience
Here’s how the pieces fit into a single workflow. Each step draws on one or more of the connection points detailed below.
Plan: Define the outcome, decisions, and acceptance criteria in Jira and Confluence.
Connect context: GitHub for Atlassian and the Teamwork Graph connector link work, knowledge, and code.
Build: Develop in GitHub, or delegate work to a coding agent from Jira.
Review and ship: Review a GitHub pull request with the original Jira and Confluence context close at hand.
Learn: GitHub activity flows back into Jira and Rovo, keeping teams and agents aligned.
One connected loop: teams decide what to build in Jira and Confluence → developers build it in GitHub → the code, pull requests, and status flow back into Jira → and that up-to-date picture shapes what the team plans next.
Explore the ways to connect GitHub and Atlassian
GitHub for Atlassian: Track GitHub development from Jira
Best for: developers, product managers, and delivery leads who want implementation status attached to the Jira work item.
Connect GitHub and Jira so branches, commits, pull requests, builds, deployments, and related development activity appear alongside the work they deliver. When teams include Jira work item keys in branch names, commits, and pull requests, stakeholders can follow progress from Jira without asking developers for a separate status update.
Streamline development workflows across GitHub, Jira, and your Atlassian apps having unified data with one, Atlassian-powered app.
How to connect GitHub to Jira
Install GitHub for Atlassian and link your repositories to your Jira site.
Open a Jira work item with clear acceptance criteria.
In GitHub, create a branch and pull request that reference the Jira key.
Return to the Jira work item and open the development panel.
Show the linked branch, commit, pull request, and build or deployment status.
Value moment: product and delivery teammates see real implementation progress in Jira while developers keep working in GitHub.
Teamwork Graph connector for GitHub: Search and reason across GitHub, Jira, and Confluence
Best for: engineering managers, product teams, support, and leaders who need answers that connect code activity with business and delivery context.
Make eligible GitHub repositories, pull requests, pull request comments, issues, issue comments, commits, and branches available as context for Rovo-powered experiences. Rovo can bring GitHub delivery signals together with Jira work and Confluence knowledge, helping teams understand progress, blockers, ownership, and decisions without manually stitching together links.
Query: What's stuck in our sprint with no recent commits? Rovo compares current sprint work in Jira with recent GitHub PR activity, then highlights where the real follow-up is Jira status hygiene rather than missing code.
Value moment: the answer explains the delivery story, not just a list of search results.
Setup: an organization administrator connects GitHub through GitHub for Atlassian and enables GitHub as a Rovo source. Availability depends on the connection type, permissions, and Rovo eligibility.
Atlassian MCP Server with GitHub Copilot: Connect GitHub Copilot to Atlassian through an open protocol
Best for: teams standardizing on MCP-compatible AI clients and administrators who want a governed way to expose Atlassian tools and context.
The Atlassian MCP Server is the connection layer between Atlassian and compatible AI clients that support the Model Context Protocol. Once you add it in Visual Studio Code, GitHub Copilot Chat (agent mode) can use permitted Jira and Confluence context and supported actions from the developer’s existing chat experience — no separate content store, and no leaving the editor. Access follows the user’s Atlassian permissions and administrator controls.
Use Atlassian MCP Server to surface relevant Atlassian context from within VS Code
How to connect GitHub Copilot to Atlassian MCP Server
Configure the Atlassian MCP Server in Visual Studio Code by adding it to your MCP configuration.
Complete the Atlassian OAuth flow so the connection runs under your existing permissions.
In GitHub Copilot Chat (agent mode), ask an example prompt such as “Summarize the acceptance criteria on PROJ-123 and the linked Confluence page.”
Review each proposed action before executing so a human stays in control.
Value moment: organizations can bring governed Atlassian context to the AI client developers already use, without duplicating data or replacing the client.
Requirements: an Atlassian account with access to the relevant Jira and Confluence content, Visual Studio Code with GitHub Copilot Chat (agent mode) enabled, and the Atlassian MCP Server configured. Access respects your Atlassian permissions and any administrator controls.
The same server connects Claude and Claude Code
GitHub Copilot for Jira: Delegate a Jira work item to GitHub Copilot
Best for: software teams that want to start agentic implementation from a well-defined Jira work item.
Assign a Jira work item to GitHub Copilot for Jira. The agent uses work item context, inspects the selected GitHub repository, implements a change, and opens a draft pull request for human review. The handoff begins in Jira and returns to the normal GitHub review workflow.
How to delegate a Jira work item to GitHub Copilot
Open a Jira work item with clear acceptance criteria and assign it to GitHub Copilot for Jira.
Select the target GitHub repository so the agent has the right codebase context.
Let the agent inspect the repository, implement the change, and open a draft pull request in GitHub.
Review the draft pull request in GitHub’s normal review workflow before merging.
Value moment: teams move from planned work to a reviewable pull request without copying requirements into a separate coding-agent prompt.
Requirements: Rovo enabled for the Atlassian site, an eligible GitHub Copilot subscription, and the GitHub Copilot for Jira Marketplace app configured by administrators.
How the experiences work together
Start with GitHub for Atlassian to see code progress on Jira work; add the Teamwork Graph connector for GitHub to ask questions across GitHub, Jira, and Confluence; use the GitHub Copilot for Jira to turn a work item into a draft pull request; and connect through the Atlassian MCP Server when you want GitHub Copilot or another compatible AI client to reach Atlassian over an open protocol.
If you want to… | Start with… |
See code progress on Jira work | GitHub for Atlassian |
Ask questions across GitHub, Jira, and Confluence | Teamwork Graph connector for GitHub |
Turn a Jira work item into a draft pull request with GitHub’s agent | GitHub Copilot for Jira |
Connect GitHub Copilot or another compatible AI client through an open protocol | Atlassian MCP Server |
These experiences are complementary. A team can track implementation through GitHub for Atlassian, search the broader delivery story with the Teamwork Graph, delegate a Jira work item to GitHub Copilot for Jira, and use the Atlassian MCP Server when they want GitHub Copilot or another compatible AI client to reach Atlassian context over a governed, open protocol.
FAQs
Do I have to leave GitHub — or Jira — to see development progress? No. With GitHub for Atlassian, branches, commits, pull requests, builds, and deployments show up on the Jira work item itself. As long as your team includes the Jira work item key in branch names, commits, and pull requests, product and delivery teammates can follow progress from Jira while developers keep working in GitHub — no separate status update needed.
Do these AI agents make changes automatically, without human review? No. Every connection operates under your existing Atlassian permissions and keeps humans in control. Coding agents open draft pull requests for human review rather than merging on their own, and generated code, proposed actions, and pull requests all pass through your normal governance process.
What do we need to get started? Begin by installing and configuring GitHub for Atlassian for your repositories and Jira sites, and reference Jira keys in your GitHub work. To use Rovo-powered experiences, an administrator enables GitHub as a Rovo source, which requires an eligible Rovo experience and the Teamwork Graph connector for GitHub permissions. Individual agent surfaces have their own prerequisites: GitHub Copilot for Jira needs Rovo enabled, an eligible GitHub Copilot subscription, and the GitHub Copilot for Jira Marketplace app.
Getting started
Connect GitHub and Jira. Install and configure GitHub for Atlassian for the repositories and Jira sites your teams use.
Enable GitHub context in Rovo. Confirm the organization has an eligible Rovo experience and the required Teamwork Graph connector for GitHub permissions.
Add the right agent surfaces. Configure GitHub Copilot for Jira and Atlassian MCP Server based on the workflows your team wants.
Keep humans in control. Review generated code, proposed actions, permissions, and pull requests through your existing governance process.