Teamwork Graph CLI GA: connected context at enterprise scale, speed, and security

[#1]

🎉 Teamwork Graph CLI is now GA! Since Open Beta, we've introduced OAuth support, audit logs, granular scope controls, admin self-serve, expanded command coverage, and usability improvements so organizations can adopt Teamwork Graph CLI at scale.

Hi Atlassian Community 👋

We are thrilled to announce that Teamwork Graph CLI is generally available today.

Since Open Beta at Team '26 Anaheim, many of you have put CLI to work and the reception has been incredible:

"It's incredibly powerful. I have Forge apps building on my machine autonomously, and that's all driven from Jira, and sewn into a console app on my machine via the CLI.” - Richard Sworder, Head of Product Ownership, Atlassian Williams F1 Team

"Teamwork Graph CLI is legitimately good at breaking down silos. We used it to look across Jira and team conversations for similar AI QA harness work happening elsewhere, and it did a great job surfacing duplicative efforts that would otherwise have stayed hidden." - Ryan Boyd, VP of IT, SpotOn

You've also told us exactly what you needed to take it further: OAuth support, audit logs, granular scope controls, admin self-serve, and more commands.

We listened. GA delivers all of it, so your organization can adopt Teamwork Graph CLI with the security, control, and confidence enterprise scale requires. For those of you who are already using Teamwork Graph CLI, jump down to see What's new. For those of you who are thinking about getting started, let's dive right in.

What's Teamwork Graph CLI?

Teamwork Graph CLI is a purpose-built command-line interface that gives AI coding agents like Claude Code, Cursor, Codex, and Cowork a connected work context across your entire organization. Agents can reason across Atlassian products (Jira, Confluence, JSM, Bitbucket), Atlassian apps (goals, projects, teams), and 100 connected third-party tools like GitHub, Snyk, and Figma – all from an AI terminal or app.

Standard MCP tools force LLMs to fetch raw data one app at a time. That means they're figuring out how a project connects to a team, which content is relevant, and who owns what on-the-fly at every prompt. Every relationship the agent has to infer consumes tokens, and you end up burning through token budgets to get answers that are still incomplete. Teamwork Graph CLI delivers pre-assembled context directly to the LLM. That way, agents get a structured relationship map they can traverse directly – from "who owns it" to "what depends on it" to "how it connects to business outcomes" – operating more efficiently and fetching only the data it truly needs. The gains show in the results: 44% better answers using 48% fewer tokens, consistent across every test we ran in an internal study.

For more on Teamwork Graph CLI, read here.

What's new?

Setup fast and secure with OAuth 2.1 authentication. Teamwork Graph CLI now supports OAuth 2.1, replacing static API tokens with short-lived, auto-rotating credentials that align with enterprise security standards. If your organization had Teamwork Graph CLI disabled for security reasons, now is a great time to revisit.

Granular controls for users and admins. With granular scope controls, including a dedicated Delete scope, separate from Write, you can give your agents read and write access without handing over delete permissions. We've also expanded the total number of available scopes to keep pace with Teamwork Graph CLI's growing command surface.

Roll it out org-wide with reassurance. All Teamwork Graph CLI requests now flow through Atlassian's audit infrastructure, so admins can see exactly which commands are being invoked, by whom, and when. Audit logs are available at admin.atlassian.com → Insights → Audit Logs and can be exported as JSON for tools like Splunk.

Get more done with new commands. Since Open Beta, we've added 37 new commands, bringing TWG CLI's total coverage to 567. The biggest additions: 32 new Jira commands (boards, dashboards, filters, and bulk operations), 17 new Assets commands, and 14 new Bitbucket commands.

Build automated workflows that require minimal supervision. You shared, we listened. We resolved a number of gaps beta participants flagged on third-party data not surfacing through Rovo search connectors, JSM internal comments, sprint support, and discovery for Jira, and formatting preservation. We also improved update awareness so your CLI and agent skills stay up to date as we ship new improvements.

For the full list of updates, go here.

Ready to get started?

Whether you're new to Teamwork Graph CLI or updating from Open Beta, here's how to get going:

• Install or update Teamwork Graph CLI: Use the latest setup instructions in the product documentation.
• Authenticate with OAuth: Run the OAuth login flow and complete browser authorization.
• Install or refresh agent skills: Make sure your agent has the latest Teamwork Graph CLI guidance.
• Ask your agent for connected work context: Try a prompt (like below). Prompts that ask for the full picture around a Jira issue, Confluence page, PR, incident, or project are most recommended. For more prompts, visit here.

Use Teamwork Graph CLI to get the full picture on PROJ-4821. Who is involved, what Jira work is linked, what Confluence pages or PRs are relevant, and is there any recent activity I should know about?

Use Teamwork Graph CLI to help me prepare for our sprint review. The sprint ends Friday.

How are you using Teamwork Graph CLI?

Share feedback in the comments, through twg feedback, or by creating an issue in the Teamwork Graph CLI GitHub repository.

Don't forget to let us know how you're using Teamwork Graph CLI with your agents and in your workflows in the thread below. We value hearing from folks like you and @Alexander Nilsson on all the creative ways you use our capabilities!

Cheers,
Ashish

---

FAQs

What is the Teamwork Graph CLI?
Teamwork Graph CLI is an agent-native command-line interface that gives AI coding agents, like Claude Code and Codex, access to your connected work context across Atlassian and third-party surfaces. It helps AI agents deliver better results on your agentic workflows (44% better results) while using fewer tokens (48% less), making your AI-powered experiences richer and more cost-efficient.

How do I get started with Teamwork Graph CLI?
To get started, install Teamwork Graph CLI and run the authentication command. You'll see a prompt in your terminal to open a browser, where you'll complete a secure OAuth 2.1 login and select your Atlassian site. Once authenticated, you're ready to run commands. Find the full details in our product documentation.

How does authentication work now that Teamwork Graph CLI is GA?
Teamwork Graph CLI now uses OAuth 2.1, replacing the earlier static token approach. When you authenticate, a browser window opens so you can log in securely and select your Atlassian site. OAuth generates short-lived, auto-rotating tokens, meaning your credentials are more secure and align with enterprise security standards.

Does using Teamwork Graph CLI cost anything? Will it affect my Rovo Credits?
Teamwork Graph CLI is free to install and use. In the future, installing the CLI will remain free, but certain forms of usage will be billed. With 90 days' advance notice, certain enriched, value-added commands will be billed with at least 1 Rovo credit per call. Certain complex queries will cost more than 1 credit per call.

These enriched, value-added commands require TWG CLI to aggregate responses from multiple Atlassian and non-Atlassian apps and run AI inferencing to generate insights. For example, twg collaborator provides insights on collaborators by analyzing common artifacts and interactions between users in an organization. Similarly, twg context aggregates information across multiple Atlassian and non-Atlassian systems to deliver the best context around a work item, Confluence page, or user. These tools offer more value compared to the free tools but cost more due to additional compute requirements.

We will publish the detailed rate card with the 90 days advance notice before we start billing.

Is Teamwork Graph CLI available to all customers?
Yes, Teamwork Graph CLI is available to all Atlassian customers.

I am an admin. Can I re-enable Teamwork Graph CLI if we turned it off during beta?
Yes, if you were waiting for enterprise security and controls to enable Teamwork Graph CLI for org-wide rollout, now is a great time for admins to revisit. With GA, OAuth 2.1 is live, along with granular scope controls and full audit logs. Security teams can review the new controls at admin.atlassian.com → Rovo → Teamwork Graph CLI and re-enable access for your users.

How is Teamwork Graph CLI different from the Rovo MCP?
Both the CLI and Rovo MCP give agents access to Atlassian data via the Teamwork Graph, but they're built for different environments. Most teams will use both. Teamwork Graph CLI is best for terminal, CI/CD, and automation workflows, offering the broadest Atlassian command coverage and strong token efficiency for complex, retrieval-heavy tasks. Rovo MCP is better suited for web-based LLMs, IDEs, and sandboxes where shell access isn't available, with pre-declared tools that work well for simpler, fewer-turn tasks. Use MCP as your everyday web-friendly default and reach for Teamwork Graph CLI when you need deeper coverage, scripted automation, or local tool chaining. For more guidance, go here.

Can I control what permissions my AI agent has?
Yes. Teamwork Graph CLI now supports granular scope controls. Admins and users can grant agents read-only access, read + write access, or read + write + delete access separately, so you're not forced to give broad permissions. During authentication, users can customize their own scopes. Admins can set org-wide scope policies at admin.atlassian.com → Rovo → Teamwork Graph CLI.

I'm an admin. How do I see what Teamwork Graph CLI is doing in my organization?
Teamwork Graph CLI now has full audit log support. All CLI requests flow through Atlassian's audit infrastructure. To view them, go to admin.atlassian.com → Insights → Audit Logs → filter by Activity: "Invoked TWG CLI command." Logs are also available as JSON exports, so you can analyze them in tools like Splunk.

How does Teamwork Graph CLI work with AI agents like Claude Code or Codex?
Teamwork Graph CLI is designed to be agent-first. You can define its usage in Agent Skills files (e.g., AGENTS.md or CLAUDE.md) so your AI coding agent knows when and how to call Teamwork Graph CLI commands. This approach is especially valued for its token efficiency.