This article lists available mabl MCP tools and prompts that you can invoke using natural language prompts.
Tools
| Tool | Description | Sample prompt |
|---|---|---|
| #analyze_failure* |
Input: plan run ID or test run ID Return a GenAI analysis of why a test or plan execution failed based on run logs and failure metadata. |
"Using the mabl MCP, please analyze the failure of the ‘checkout flow’ test from last night." |
| #assign_xray_to_mabl_test |
Input: mabl test ID and Xray test case Optional input: Xray project ID (required if you have more than one project) Create or update an association between a mabl test and an Xray test case. Requires prior setup of Xray results reporting. |
"Link Xray test case 'QA-159' to mabl test 'WkY5qe7Is3IOx3iGCK2Zbw-j' and overwrite any existing assignment." |
| #create_mabl_test* |
Input: test name, prompt, and base URL Optional input: application, environment, API spec (API tests only) Create a new browser or API test according to your provided prompt. |
"Create a new mabl browser test based on the details in this Jira ticket: TICKET-123." |
| #create_mabl_test_from_plan* |
Input: Artifact generated from the #plan_new_test tool Create a mabl browser test according to your provided test plan. |
"Please create a mabl test using the plan generated for the 'Invoicing Page Loading' check." |
| #export_to_playwright |
Input: test name or ID Export a mabl test to a Playwright test file for local playback or CI. |
"Can you export the mabl test 'App - User language - Dashboard' to Playwright?" |
| #get_applications | List and search mabl applications. | "What mabl tests do I have for the ‘sandbox’ application?" |
| #get_credentials | List and search your mabl credentials. Use with #run-test to specify which credentials you want to use. | "Can you run that test with the ’staging_env’ credentials?" |
| #get_environments | List and search mabl environments. | "What are the inactive plans in the 'QA' environment?" |
| #get_latest_authored_tests | Get a list of the most recently created tests by the current user. | "What mabl tests have I created recently?" |
| #get_latest_plan_runs |
Input: plan name or ID Get recent mabl plan run results, including a pass/fail summary and run metadata. |
"Summarize the results of the 'production deployment' plan from the last 24 hours. Show me the pass/fail rate and provide details for test failures." |
| #get_latest_test_runs |
Input: test name or ID Get recent results for a specific mabl test, including run IDs, status, timestamps, and error messages. |
"Can you fetch the results for the ‘App - Clear filters’ test?" |
| #get_mabl_deployment |
Input: commit hash or deployment event ID Get the results of mabl test runs associated with your most recent commit. Use to find out if the most recent commit broke tests. |
"What happened with my previous mabl deployment?" |
| #get_mabl_test_details |
Input: test name or ID Get test metadata and step definitions for a specific test |
"Get the test details for the 'App - Clear filters' test." |
| #get_mabl_tests | Search tests in the current workspace by name or query and return matching test metadata. Includes support for semantic search queries based on actions or goals. | "Do I have any mabl tests that cover this dropdown component?" |
| #get_plan_run_result |
Input: plan run ID Get full result details for a specific mabl plan run. |
"Show me a test breakdown for plan run ID 12345-pr." |
| #get_plans | Get all mabl plans. | "Which plans are currently enabled in this workspace?" |
| #get_workspaces | List all of your available workspaces, get the current workspace. | "What's my current mabl workspace?" |
| #get_xray_test_cases |
Optional input: Xray project ID (required if you have multiple Xray projects), limit Get all Xray test cases for a specific project. Requires prior setup of Xray results reporting. |
"Fetch my Xray test cases (limit 10) so I can find the login feature test cases." |
| #plan_new_test* |
Input: Test requirements Plan a new browser test based on requirements. You may refine the plan iteratively. When it aligns with your goals, use the #create_mabl_test_from_plan tool to build out the test. |
"Can you plan a new mabl test based on the following steps: navigate to the product page, click 'Add to Cart', and verify the cart count increases?" |
| #run_mabl_test_cloud |
Input: test name or ID Optional input: to override the default configuration, include the application, environment, and/or URL Triggers a browser or API run in the mabl cloud and provides a URL to review test output in the app. |
"Run the "App - Settings - Address info" test in the Staging environment on Chrome and Firefox." |
| #run_mabl_test_local |
Input: test name or ID Optional input: to enable GenAI Assertions, include Run mabl browser and API tests locally |
"OK, let’s run the first mabl test locally using my local environment." |
| #switch_workspace |
Input: workspace name or ID Switch to a different mabl workspace. |
"Can you switch over to the ‘project_alpha’ workspace?" |
*Indicates a tool that is only available with the Advanced AI add-on
When you invoke MCP tools, it isn’t necessary to use the exact name of the command. Describe what you want to do using natural language and provide inputs when needed. For example:
User: "Do I have any mabl tests that validate this dropdown component?"
AI: Gets matching tests with #get-tests tool.
User: "Can you run that test locally? Use the URL https://localhost:3000."
AI: Runs the test locally with #run-test tool.
Include "mabl" in prompts
If your AI client is connected to other tools and MCPs that perform similar tasks as the mabl MCP, such as creating or running tests, include “mabl” in your prompts to clear up ambiguity. For example: “Please create a mabl test using this prompt…”
Obtaining IDs
In many cases, you can obtain the IDs directly from the AI agent by including it in a prompt. For example: "Get recent plan runs for the plan 'App - a11y tests' and include the plan run IDs, please."
If you can’t get a specific ID directly from the mabl MCP, check out our article on obtaining mabl resource IDs from the mabl app, CLI, and API.
Prompts
The mabl MCP also provides a prompt called create_mabl_test_for_current-work, which you can use to create a new test based on your current work context. To use it, enter the following in the AI client chat: /mcp.mcp-mabl.create_mabl_test_for_current-work.
Managing MCP tool permissions
As you start invoking different MCP tools, your AI agent will prompt you to set permissions for each tool. Depending on which AI client you’re using, permissions might include:
- “allow once”
- “allow in this session”
- “always allow”
- “decline/cancel”
To optimize your experience with the mabl MCP, configure these permissions intentionally based on your specific needs and workflow. For actions that aren’t easily revertible, such as #create_mabl_test and #run_mabl_test_cloud, you may want to grant more limited permissions such as “allow once” or “allow in this session”. For read-only actions, you might want to select “always allow” to reduce bottlenecks.