Bitbucket Cloud is now available on the Atlassian Rovo MCP server

[#1]

We're excited to announce that Bitbucket Cloud is now supported on the Atlassian Rovo MCP Server. Meaning AI clients like Claude, ChatGPT, Cursor, and VS Code can now browse repositories, create commits, open pull requests, and check pipeline results  all through the same secure MCP connection that already works with Jira and Confluence.

Available tools and scopes

The table below lists every Bitbucket Cloud tool currently available through the Atlassian Rovo MCP Server, along with the actions each supports and the scopes required.

Tool	Action	Description	Scopes
bitbucketWorkspace	list	List all workspaces	read:workspace:bitbucket
	get	Get a specific workspace	read:workspace:bitbucket
bitbucketRepository	list	List repositories in a workspace	read:repository:bitbucket
	get	Get a specific repository	read:repository:bitbucket
bitbucketRepoContent	branch.get	Get a branch	read:repository:bitbucket
	branch.create	Create a new branch	write:repository:bitbucket read:repository:bitbucket
	commit.create	Create a commit	write:repository:bitbucket
	commit.get	Get a specific commit	read:repository:bitbucket
	files.get	Get file/directory contents at a ref	read:repository:bitbucket
bitbucketPullRequest	create	Create a new pull request	read:pullrequest:bitbucket write:pullrequest:bitbucket
	get	Get a specific pull request	read:pullrequest:bitbucket
	list	List pull requests	read:pullrequest:bitbucket
	comment	Add a comment to a pull request	read:pullrequest:bitbucket
	comments	List comments on a pull request	read:pullrequest:bitbucket
	diff	Get the diff of a pull request	read:pullrequest:bitbucket
	approve	Approve a pull request	read:pullrequest:bitbucket write:pullrequest:bitbucket
	merge	Merge a pull request	read:pullrequest:bitbucket write:pullrequest:bitbucket
bitbucketPipeline	list	List pipelines	read:pipeline:bitbucket
	get	Get a specific pipeline	read:pipeline:bitbucket
	run	Trigger a pipeline	read:pipeline:bitbucket write:pipeline:bitbucket │
	steps	List steps for a pipeline	read:pipeline:bitbucket
	step.get	Get a specific pipeline step	read:pipeline:bitbucket
	step.log	Get the log for a pipeline step	read:pipeline:bitbucket
bitbucketEnvironment	list	List environments	read:pipeline:bitbucket
	get	Get a specific environment	read:pipeline:bitbucket
	create	Create a specific environment	admin:pipeline:bitbucket
	update	Update an environment	admin:pipeline:bitbucket
	delete	Delete an environment	admin:pipeline:bitbucket
bitbucketDeployment	list	List deployments	read:pipeline:bitbucket
	get	Get a specific deployment	read:pipeline:bitbucket

What you can do with these tools

Here are a few examples of how these tools come together in practice:

• Explore a codebase: List workspaces and repositories, browse branches, and read file contents all from your AI client.

• Manage pull requests end to end: Create PRs, review diffs, comments, approve, and merge without switching context.

• Monitor pipelines and deployments: Check pipeline runs, view environments, and track deployments. When a build breaks, your AI client can help you trace the failing commit.

Getting started

Before starting, keep in mind that Bitbucket Cloud tools can be used only with workspaces that are connected to an Atlassian organization. If your workspace has not yet been linked, review this guide and complete this setup process. Additionally, access to Bitbucket tools is presently limited to API token authentication; OAuth is not yet supported, though it is on the roadmap and currently under development

To get set up, an org admin needs to enable API token auth in Admin Hub > Rovo > Rovo MCP. Once that's done, connect through any supported MCP client (Claude, ChatGPT, Cursor, VS Code, and others) and start using Bitbucket tools alongside Jira and Confluence.

For full setup details, visit the Atlassian Rovo MCP Server documentation.

Feedback and support

This is just the beginning, and we’re excited to develop additional tools for Bitbucket Cloud based on this foundation. Your feedback will directly influence the features we prioritize and the tools we create next. If you have ideas for new tools or encounter any problems, please leave a comment below. We are here to assist and we value your input.

[Susan Aryal]

It's great to have an official remote Bitbucket MCP server finally.

If the bitbucket tools were named with their actions it would be great.

For e.g., Instead of having all the pull requests related tool under a single tool called bitbucketPullRequest with different actions, it would be really beneficial if different tools were there for different actions.

We use tools' name to filter what tools to pass to the agent and what not to. It becomes difficult if a single tool does alot of actions.

If I want to pass the tool to add comment to PR to the agent but not add the tool to delete a PR, I cannot do that now. Using the API Token with the restricted scopes is an option but still, the agentic frameworks like strands allow us to allow/deny tools based on their names which would be really great. Just like the confluence, Jira tools names how they have separate tools for separate actions like getConfluencePage, createConfluencePage, etc.