Testkube MCP server

• Lists your test workflows

  Lists Testkube test workflows with filters for resource group, labels and text, and returns names, descriptions and recent status.

  Try: “List my test workflows and their recent execution status.”

  Read · list_workflows

• Shows one workflow in detail

  Returns one workflow with its execution history, health metrics and current status.

  Try: “How is the api-tests workflow doing?”

  Read · get_workflow

• Shows a workflow's YAML

  Returns the full YAML definition of a workflow, including steps and the config options it accepts.

  Try: “Show me the YAML of the api-tests workflow.”

  Read · get_workflow_definition

• Shows workflow reliability numbers

  Returns execution counts, health scores, pass rates and performance data for a workflow.

  Try: “What is the pass rate of api-tests over time?”

  Read · get_workflow_metrics

• Creates a test workflow

  Creates a workflow from YAML. It can be run right away, and its steps run in your environment.

  Try: “Create a workflow that runs my Python API tests.”

  Runs code · create_workflow

• Changes a test workflow

  Replaces an existing workflow with a new YAML definition, and it takes effect immediately.

  Try: “Update api-tests to use the latest Python image.”

  Runs code · update_workflow

• Runs a test workflow

  Starts a workflow, with optional config values and a chosen runner. The workflow's steps run on your Testkube runner.

  Try: “Run api-tests on the staging runner.”

  Runs code · run_workflow

• Lists workflow templates

  Lists reusable workflow templates with names, descriptions and labels.

  Try: “What workflow templates do we have?”

  Read · list_workflowtemplates

• Shows a template's YAML

  Returns the full YAML of one workflow template.

  Try: “Show the YAML of the k6 template.”

  Read · get_workflowtemplate_definition

• Creates a workflow template

  Creates a reusable template from YAML. Workflows can use it right away.

  Try: “Create a template for k6 performance tests.”

  Write · create_workflowtemplate

• Changes a workflow template

  Replaces a template's definition. Workflows that use the template pick up the change.

  Try: “Update the k6 template to a newer image.”

  Write · update_workflowtemplate

• Searches many workflow definitions

  Runs a JSONPath query across all workflow definitions, for example to find which use a certain image. It only registers when the control plane supports it.

  Try: “Which workflows use the node:18 image?”

  Read · query_workflows

• Searches many execution records

  Runs a JSONPath query across execution records from many workflows, such as all failures or durations. It only registers when the control plane supports it.

  Try: “Show all failed executions from the last week.”

  Read · query_executions

• Shows the workflow YAML schema

  Returns the fields and types a TestWorkflow definition allows.

  Try: “What fields can a TestWorkflow have?”

  Read · get_workflow_schema

• Shows the execution data schema

  Returns the fields and types of execution data, to help write queries.

  Try: “What does an execution record contain?”

  Read · get_execution_schema

• Lists labels and values

  Lists workflow labels and their values, useful for filtering other calls.

  Try: “What labels are used on my workflows?”

  Read · list_labels

• Lists resource groups

  Lists the resource groups in your organization.

  Try: “Which resource groups exist?”

  Read · list_resource_groups

• Lists runners

  Lists the agents (runners) that can run workflows, with type, capabilities and status.

  Try: “Which runners can I run tests on?”

  Read · list_agents

• Reads execution logs

  Returns the last 100 lines of an execution's logs by default, and can search the full log by text.

  Try: “Show the logs of the last failed api-tests run.”

  Read · fetch_execution_logs

• Lists test executions

  Lists executions with filters for workflow, status, dates, labels, tags and text.

  Try: “Show failed executions from yesterday.”

  Read · list_executions

• Finds an execution id by name

  Looks up an execution's id from its name.

  Try: “What is the id of execution api-tests-123?”

  Read · lookup_execution_id

• Shows one execution in detail

  Returns status, timing, results, configuration and workers of an execution, and the failing step for one that did not pass.

  Try: “Why did execution api-tests-123 fail?”

  Read · get_execution_info

• Shows resource use of one run

  Returns CPU, memory, disk and network metrics over time for a single execution.

  Try: “Show CPU and memory use for execution api-tests-123.”

  Read · get_workflow_execution_metrics

• Compares resource use across runs

  Computes statistics, trends and outliers for CPU, memory, disk and network across recent runs of a workflow.

  Try: “Is api-tests using more memory over time?”

  Read · get_workflow_resource_history

• Waits for runs to finish

  Polls a list of executions until they finish and returns their final status.

  Try: “Run both workflows and wait until they finish.”

  Read · wait_for_executions

• Stops a running execution

  Aborts a running execution and marks it aborted.

  Try: “Abort the stuck api-tests run.”

  Delete · abort_workflow_execution

• Sets tags on an execution

  Replaces all tags on an execution with the ones you give. An empty set clears them.

  Try: “Tag this run env=prod and bug=found.”

  Write · update_execution_tags

• Lists files from a run

  Lists the files, reports and logs an execution produced.

  Try: “What artifacts did execution api-tests-123 produce?”

  Read · list_artifacts

• Reads an artifact file

  Returns up to 200 lines of an artifact per request, and a summary for binary files.

  Try: “Read the junit report from that run.”

  Read · read_artifact

• Finds available test metrics

  Lists the metric series parsed from k6, JMeter, Artillery, JUnit and Influx reports.

  Try: “Which performance metrics are tracked?”

  Read · list_insight_series

• Lists metric names

  Lists the distinct metric keys available, such as 95th percentile response time.

  Try: “What metric names can I query?”

  Read · list_insight_metric_keys

• Charts a metric over time

  Returns a metric as a time series so you can see values and trends.

  Try: “Show p95 response time for the last month.”

  Read · get_insight_metric_series

• Finds runs behind a metric

  Lists the executions that produced a given metric, newest first.

  Try: “Which runs produced that latency spike?”

  Read · list_insight_executions

• Lists Insights boards

  Lists the dashboards visible to you, shared and private.

  Try: “What Insights boards do I have?”

  Read · list_boards

• Shows a board and its reports

  Returns a board with its reports in order and its layout.

  Try: “Show the contents of the API health board.”

  Read · get_board

• Runs a board's reports

  Runs each report's query and returns the numbers with the query and date range used.

  Try: “Render the API health board for last month.”

  Read · render_board

• Creates a board

  Creates an empty dashboard, shared with the organization unless you make it private.

  Try: “Create a private board called API health.”

  Write · create_board

• Changes a board

  Changes a board's name, description, slug, visibility or layout.

  Try: “Make the API health board shared.”

  Write · update_board

• Adds a report to a board

  Adds a pass/fail, executions, workflows or time-series chart to a board.

  Try: “Add a pass/fail chart for api-tests to the board.”

  Write · add_board_report

• Changes a report

  Changes a report's name, description, kind or settings, merging new settings into the old ones.

  Try: “Rename the report to Weekly pass rate.”

  Write · update_board_report

• Removes a report

  Removes a report from a board. This cannot be undone.

  Try: “Remove the latency report from the board.”

  Delete · remove_board_report

• Deletes a board

  Permanently deletes a board and all its reports. A shared board needs an organization admin.

  Try: “Delete the old API health board.”

  Delete · delete_board

• Builds a dashboard link

  Builds a link to a workflow or an execution in the Testkube dashboard, optionally to a log step.

  Try: “Give me a link to the failed run.”

  Read · build_dashboard_url

Watch out for

• It can create and run workflows that execute code

  create_workflow, update_workflow and run_workflow write and start workflows whose steps run on your Testkube runners. There is no read-only switch, so approve each call. github.com

• It can delete boards and abort runs

  delete_board removes a board and all its reports for good, and abort_workflow_execution stops a running test. Deleting a shared board needs an organization admin. docs.testkube.io

• Docker tokens cannot use the board tools

  Board tools need a signed-in user, so they fail with an API token in Docker. This is a feature limit, but it also means a hosted OAuth login carries your full user rights. docs.testkube.io

• The server sends usage events to Testkube

  Telemetry is on by default and sends the tool name, duration and error status. You can turn it off in the Testkube CLI config. github.com

• Standalone HTTP mode needs its own token

  The optional shttp mode binds to localhost by default and requires a separate inbound bearer token (TK_MCP_SHTTP_AUTH_TOKEN). Bind externally only behind TLS. github.com

Security advisories