Integrating mabl MCP with Atlassian MCP

For the full index of agent-friendly pages, see llms.txt.

By using the mabl MCP and Atlassian MCP in tandem, you can consolidate testing and development activities directly into your AI client, such as GitHub Copilot or Claude Desktop. This article outlines three ways to use the mabl MCP and Atlassian MCP to streamline your work:

• Get Jira test case details > create mabl test
• Get Jira task > write code > test changes in mabl
• Create mabl test > update Jira ticket

Get Jira test case details > create mabl test

If your team documents test cases in Jira tickets, you can set up your AI client to create mabl tests based on those test case details. This workflow isn’t just limited to developers. Anyone on the team with access to an AI client that supports MCPs can use these actions to reduce manual effort and ensure tests are accurately aligned with the requirements.

Example

When gathering details for your mabl test, it’s helpful to use back-and-forth dialogue to progressively refine your intent:

• User: “Give me the test case details for the Jira ticket TEST-789.”
• AI: “The field must be editable by editor roles. After entering a new email, the submit button should be enabled. After pressing save, and reloading the page the new email should be persisted.”
• User: “I want to create a new mabl test that validates these test cases. Can you remind me of the applications and environments in my mabl workspace?”
• AI: “Certainly. Your mabl workspace contains the applications Admin Dashboard, Client Portal, and Marketing Site. The available environments are Development, Staging, and Production.”
• User: “And what mabl credentials are available to me?”
• AI: “You have two credentials available: Test user with ‘editor’ role and Admin user with ‘admin’ role.”
• User: “Okay, let’s create a new mabl test with the test case details. Include the following: Test name: “Verify Admin Email Update”; Test environment: Staging; Application: Admin Dashboard; User credentials: Test user with ‘editor’ role.”

At this point, the mabl MCP launches the mabl agent to generate steps according to the prompt. When the user is satisfied, they can save the test and return to their IDE.

Get Jira task > write code > test changes in mabl

When working on new features, developers can use the Atlassian MCP to get details on a task, update their code, and call on the mabl MCP to validate those changes locally. Instead of switching between different UIs, all tasks are handled from their IDE-based client, such as GitHub Copilot.

Example

The following sample dialogue demonstrates this workflow in action:

• User: “Give me the details for TEST-234.”
• AI: “TEST-234: Update the email field so it accepts valid email addresses, including new TLDs like .dev.”

User makes code changes.

• User: “Do I have any mabl tests that validate the user signup flow?”
• AI: “Here are two related mabl tests: User sign-up with new account and New user password reset flow.”
• User: “Can you run those tests locally? Use the URL http://localhost:4000.”

Tip: If the mabl MCP doesn’t return any related mabl tests, create a new mabl test to validate the functionality. If the work item includes a field with test case details, use that as a starting point.

Create mabl test > update Jira ticket

In addition to testing in mabl based on Jira ticket details, you can also update your Jira work items with mabl test details. Both developers and non-developers alike can leverage this workflow to keep their Jira projects up to date with their mabl test activity. For developers working with an IDE-based AI client, this workflow might follow a set of code changes:

Example

The following sample dialogue demonstrates this workflow in action:

• User: “I’d like to create a new test that validates these latest changes to the new user registration flow. How would you test this using an end-to-end tool like mabl? Let me know what I should include in my prompt for best results.”
• AI: “I’d create a new test for the user registration flow. The test would include navigating to the registration page, filling in the required fields (name, email, password), submitting the form, and verifying that the user is successfully registered and logged in. For best results, you should include specific details about the form fields, any validation rules, and the expected confirmation message.”
• User: “Okay, let’s make a new mabl test with those details. Launch the Trainer to the URL http://localhost:4000.”
• AI: “Launching mabl Trainer to create a new test based on your prompt.”

The mabl MCP launches the mabl agent to generate steps according to the prompt. When the user is satisfied, they can save the test and return to their IDE.

• User: “Can you create a comment on this Jira ticket - TEST-456 - with the name and URL for the mabl test I just created?”
• AI: “Adding a comment to Jira ticket TEST-456 with the new mabl test name and URL.”