SmartBear MCP Server: Setup for Claude, Cursor and Other Apps

• Lists your BugSnag projects

  Returns the projects in your organization that you can reach, or finds the one that matches an API key.

  Try: “List my BugSnag projects”

  Read · bugsnag_list_projects

• Searches errors in a BugSnag project

  Lists errors in a project with filters and paging, so you can ask for the newest, noisiest or still-open ones.

  Try: “Show the open errors in the web app from the last 24 hours”

  Read · bugsnag_list_project_errors

• Opens one error with its latest occurrence

  Returns an error with totals across all its events and the latest event's stack trace, breadcrumbs and metadata.

  Try: “Explain error 5f2c and show the stack trace”

  Read · bugsnag_get_error

• Opens one error event

  Returns the details of a single event, which is one occurrence of an error.

  Try: “Show event 8a1b in full”

  Read · bugsnag_get_event

• Opens an event from a BugSnag dashboard link

  Takes a pasted dashboard URL and returns the event it points to, so you can drop a link into chat.

  Try: “Look at this BugSnag link and tell me what failed”

  Read · bugsnag_get_event_details_from_dashboard_url

• Lists the occurrences of one error

  Returns the events that were grouped into one error.

  Try: “List the last events for this error”

  Read · bugsnag_get_events_on_an_error

• Lists releases of a project

  Returns the releases recorded for a BugSnag project.

  Try: “What releases shipped this month?”

  Read · bugsnag_list_releases

• Shows all the spans in a performance trace

  Returns every span within one trace, which helps find where a slow request spent time.

  Try: “Show the spans of trace 91ab”

  Read · bugsnag_get_trace

• Changes an error's status

  Updates the status of an error, for example to mark it as fixed or ignored.

  Try: “Mark error 5f2c as fixed”

  Write · bugsnag_update_error

• Sets how network calls are grouped

  Replaces the rules that group network endpoint URLs together in a project's performance data.

  Try: “Group /users/123 and /users/456 as /users/:id”

  Write · bugsnag_set_network_endpoint_groupings

• Lists BearQ environments

  Returns the environments set up in the BearQ workspace and which one is the default, so you can name one in a test run.

  Try: “Which BearQ environments can I run tests in?”

  Read · bearq_list_environments

• Runs every regression-ready BearQ test

  Starts a run of the whole regression suite and returns a task you can follow. It suits release checks.

  Try: “Run the full BearQ regression suite on staging”

  Write · bearq_run_regression_tests

• Runs chosen BearQ test cases

  Starts a run of specific regression-ready test cases by id. Draft cases are rejected.

  Try: “Run BearQ test cases 12 and 15”

  Write · bearq_run_test_cases

• Runs the BearQ tests for a feature area

  Starts a run of every regression test tagged with the functional areas you name, by id or name.

  Try: “Run all BearQ tests for checkout”

  Write · bearq_run_tests_in_functional_areas

• Has BearQ explore the live app

  Explores the running application to find or update its pages and elements in BearQ's model, optionally for one functional area.

  Try: “Update BearQ's model of the settings area”

  Write · bearq_expand_application_model

• Sends a free-form request to BearQ's QA agent

  Passes an instruction to BearQ's QA lead agent, which can list, create and update test cases and manage functional areas. The tool is marked read-only, but the description says it can change things.

  Try: “Ask BearQ to add a test for the login flow”

  Write · bearq_chat_with_qa_lead

• Shows a BearQ task's state and activity

  Returns a task's metadata and activity log right away, without waiting for it to finish.

  Try: “What is the status and log of my last BearQ run?”

  Read · bearq_get_task

• Checks whether a BearQ task is still running

  Returns just the status: running, complete, error or cancelled.

  Try: “Is the BearQ run finished?”

  Read · bearq_get_task_status

• Waits for a BearQ task to finish

  Blocks until the task completes, fails or is cancelled, or the stream times out, then returns the events in order.

  Try: “Wait for the BearQ run to finish and summarize”

  Read · bearq_wait_for_task

• Cancels a running BearQ task

  Stops a task that is still running.

  Try: “Stop the BearQ regression run”

  Write · bearq_stop_task

• Deletes BearQ test cases by id

  Archives the named test cases. The description says they can be restored, and protected tests are refused.

  Try: “Delete BearQ test cases 33 and 34”

  Delete · bearq_delete_test_cases

• Lists your Reflect tests

  Returns the tests in your Reflect account.

  Try: “List my Reflect tests”

  Read · reflect_list_tests

• Shows a Reflect test and its steps

  Returns a test's name, description and every recorded step.

  Try: “Show the steps of the checkout test”

  Read · reflect_get_test_detail

• Runs one Reflect test

  Starts a run of a Reflect test and returns an execution you can check.

  Try: “Run the signup test in Reflect”

  Write · reflect_run_test

• Checks a Reflect test run

  Returns the status of a test execution.

  Try: “Did the signup test pass?”

  Read · reflect_get_test_status

• Lists Reflect test suites

  Returns every suite available in the account.

  Try: “Which Reflect suites do I have?”

  Read · reflect_list_suites

• Runs a Reflect suite

  Starts a suite execution.

  Try: “Run the smoke suite in Reflect”

  Write · reflect_execute_suite

• Cancels a Reflect suite run

  Stops a suite execution that is under way.

  Try: “Cancel the smoke suite run”

  Write · reflect_cancel_suite_execution

• Creates a Reflect test from steps

  Makes a new test holding an ordered list of steps.

  Try: “Create a Reflect test that logs in and opens the dashboard”

  Write · reflect_create_test

• Creates a reusable group of Reflect steps

  Makes a segment, which is an ordered set of steps that tests can reuse. A segment cannot refer to another segment.

  Try: “Create a login segment I can reuse”

  Write · reflect_create_segment

• Connects to a live Reflect recording

  Opens a link to an active recording session so the assistant can drive it step by step.

  Try: “Connect to my Reflect recording session”

  Write · reflect_connect_to_session

• Adds a plain-language step while recording

  Adds a natural language step to the active recording session.

  Try: “Add a step: click the Buy button”

  Write · reflect_add_prompt_step

• Removes the last recorded step

  Deletes the most recent step added to the active recording session.

  Try: “Undo the last recorded step”

  Delete · reflect_delete_previous_step

• Takes a screenshot of the recording

  Captures the current state of the active recording session.

  Try: “Show me what the recorded browser looks like now”

  Read · reflect_get_screenshot

• Lists your Swagger organizations

  Returns the organizations you belong to. On-premise admins see all of them.

  Try: “Which Swagger organizations am I in?”

  Read · swagger_list_organizations

• Searches the Swagger registry

  Finds APIs and domains in the SwaggerHub registry and returns owner, name, description and version.

  Try: “Find our payments APIs in the registry”

  Read · swagger_search_apis_and_domains

• Fetches an API's definition

  Returns the resolved definition for an owner, API name and version, as JSON by default or as stored text.

  Try: “Get the v2 definition of the orders API”

  Read · swagger_get_api_definition

• Validates an OpenAPI or AsyncAPI file

  Checks a definition and returns each finding with its line number, severity and message.

  Try: “Validate this OpenAPI file”

  Read · swagger_validate_api

• Checks an API against your governance rules

  Scans a pasted definition against your organization's standardization rules and lists the violations.

  Try: “Scan this spec against our API standards”

  Read · swagger_scan_api_standardization

• Generates a new API from a description

  Uses SmartBear AI to create and save a new API, or a new version of one, from a prompt, applying your organization's rules.

  Try: “Create an API for a todo service with CRUD endpoints”

  Write · swagger_create_api_from_prompt

• Creates or overwrites an API definition

  Saves a definition to the registry as a new API or an update to an existing one. New APIs start private.

  Try: “Save this spec as the orders API v2”

  Write · swagger_create_or_update_api

• Edits part of an API definition

  Applies exact search-and-replace edits to a stored YAML OpenAPI or AsyncAPI definition.

  Try: “Rename the id field to orderId in the orders API”

  Write · swagger_patch_api

• Fixes an API to meet your standards

  Uses AI to scan a definition for standardization errors and fix them automatically.

  Try: “Fix this API so it meets our governance rules”

  Write · swagger_standardize_api

• Lists your Swagger portals

  Searches the portals where you are at least a designer.

  Try: “Which portals can I edit?”

  Read · swagger_list_portals

• Creates a portal

  Makes a new portal in Swagger.

  Try: “Create a developer portal called Acme APIs”

  Write · swagger_create_portal

• Changes a portal's settings

  Updates one portal's configuration.

  Try: “Change the portal's title”

  Write · swagger_update_portal

• Creates a product in a portal

  Adds a new product to a portal.

  Try: “Add a product called Payments API to the portal”

  Write · swagger_create_portal_product

• Changes a product's settings

  Updates a product's settings inside a portal.

  Try: “Rename the Payments product”

  Write · swagger_update_portal_product

• Publishes a portal product

  Makes the current product content visible to portal visitors, live or as a preview first.

  Try: “Publish the Payments product as a preview”

  Write · swagger_publish_portal_product

• Deletes a portal product

  Removes a product from a portal.

  Try: “Delete the old Beta product”

  Delete · swagger_delete_portal_product

• Adds a documentation page

  Creates a Markdown or HTML page in a portal product in one call and returns a link to edit the draft.

  Try: “Add a Getting started page to the Payments product”

  Write · swagger_create_documentation_page

• Edits a documentation page

  Replaces the content or source of an existing document, in HTML or Markdown.

  Try: “Update the Getting started page with these steps”

  Write · swagger_update_document

• Removes a table of contents entry

  Soft-deletes an entry from a product's table of contents, including nested items.

  Try: “Remove the Changelog entry from the table of contents”

  Delete · swagger_delete_table_of_contents

• Lists the apps and services in the broker

  Returns the pacticipants registered in PactFlow or your Pact Broker.

  Try: “List all services in PactFlow”

  Read · contract-testing_list_pacticipants

• Shows one app or service

  Returns the details of a single pacticipant.

  Try: “Show details for the orders service”

  Read · contract-testing_get_pacticipant

• Checks whether a version is safe to deploy

  Asks the broker whether a version of a service is compatible with what is in an environment.

  Try: “Can I deploy orders-service version 1.4.2 to production?”

  Read · contract-testing_can_i_deploy

• Shows the compatibility matrix

  Returns which consumer and provider versions have been verified against each other.

  Try: “Show the matrix for orders and payments”

  Read · contract-testing_matrix

• Reads the PactFlow audit log

  Returns recorded actions in the workspace.

  Try: “Who changed webhooks this week?”

  Read · contract-testing_get_audit_log

• Writes Pact tests with PactFlow AI

  Generates Pact tests from request and response pairs, code files, or an OpenAPI document and endpoint matcher.

  Try: “Generate Pact tests for the GET /orders endpoint”

  Read · contract-testing_generate_pact_tests

• Publishes consumer Pact contracts

  Uploads one or more consumer contracts to the broker with branch and tag information.

  Try: “Publish these pacts for orders-service on main”

  Write · contract-testing_publish_consumer_contracts

• Publishes a provider API contract

  Uploads an OpenAPI or AsyncAPI contract with self-verification results for bi-directional testing.

  Try: “Publish the payments OpenAPI spec as a provider contract”

  Write · contract-testing_publish_provider_contract

• Records a deployment

  Notes that a version of a service was deployed to an environment.

  Try: “Record that orders-service 1.4.2 is deployed to production”

  Write · contract-testing_record_deployment

• Records a release

  Notes that a version was released to an environment, for mobile and library releases.

  Try: “Record release 2.0 of the mobile app”

  Write · contract-testing_record_release

• Registers a new app or service

  Adds a pacticipant to the workspace.

  Try: “Register a service called notifications”

  Write · contract-testing_create_pacticipant

• Replaces an app's metadata

  Fully replaces the display name, main branch, repository URL and other details of a pacticipant.

  Try: “Set the main branch of orders to main”

  Write · contract-testing_update_pacticipant

• Creates a deployment environment

  Adds an environment such as staging or production.

  Try: “Create a staging environment”

  Write · contract-testing_create_environment

• Creates a webhook on contract events

  Adds a webhook that fires when contract events happen.

  Try: “Create a webhook that pings Slack when a pact changes”

  Write · contract-testing_create_webhook

• Triggers a webhook once

  Runs one webhook as a test.

  Try: “Test the Slack webhook”

  Write · contract-testing_execute_webhook

• Stores a secret for webhooks

  Creates a secret used to authenticate webhooks.

  Try: “Create a secret for the Slack webhook”

  Write · contract-testing_create_secret

• Creates a user (admin)

  Adds a user account to the workspace. It needs admin rights.

  Try: “Create a user for sam@example.com”

  Write · contract-testing_admin_create_user

• Invites users (admin)

  Sends invitations to new users.

  Try: “Invite these three people to PactFlow”

  Write · contract-testing_admin_invite_users

• Replaces a user's roles (admin)

  Replaces every role assigned to a user, which changes what they can do.

  Try: “Give Sam only the viewer role”

  Write · contract-testing_admin_set_user_roles

• Deletes an app and all its data

  Removes a pacticipant together with everything tied to it.

  Try: “Delete the notifications service”

  Delete · contract-testing_delete_pacticipant

• Deletes every integration in the workspace

  Removes all consumer-provider integrations at once.

  Try: “Delete all integrations”

  Delete · contract-testing_delete_all_integrations

• Deletes one integration

  Removes a single consumer-provider integration.

  Try: “Delete the orders to payments integration”

  Delete · contract-testing_delete_integration

• Deletes a branch of a service

  Removes a branch from a pacticipant.

  Try: “Delete the feature-x branch of orders”

  Delete · contract-testing_delete_branch

• Deletes an environment

  Removes an environment by its id.

  Try: “Delete the staging environment”

  Delete · contract-testing_delete_environment

• Deletes a secret

  Removes a secret by its id.

  Try: “Delete the Slack secret”

  Delete · contract-testing_delete_secret

• Deletes a webhook

  Removes a webhook by its id.

  Try: “Delete the Slack webhook”

  Delete · contract-testing_delete_webhook

• Removes a label from a service

  Takes one label off a pacticipant.

  Try: “Remove the legacy label from orders”

  Delete · contract-testing_remove_label_from_pacticipant

• Rotates an API token

  Replaces an API token by id, so the old value stops working.

  Try: “Rotate token 1234”

  Delete · contract-testing_regenerate_api_token

• Deletes a user (admin)

  Removes a user account.

  Try: “Delete the account for sam@example.com”

  Delete · contract-testing_admin_delete_user

• Deletes a team (admin)

  Removes a team.

  Try: “Delete the Platform team”

  Delete · contract-testing_admin_delete_team

• Deletes a role (admin)

  Removes a role.

  Try: “Delete the auditor role”

  Delete · contract-testing_admin_delete_role

• Removes a role from a user (admin)

  Takes one role away from a user.

  Try: “Remove the admin role from Sam”

  Delete · contract-testing_admin_remove_role_from_user

• Removes a user from a team (admin)

  Takes a person out of a team.

  Try: “Remove Sam from the Platform team”

  Delete · contract-testing_admin_remove_user_from_team

• Resets all roles to factory defaults (admin)

  Replaces every role definition with the defaults.

  Try: “Reset all roles to defaults”

  Delete · contract-testing_admin_reset_roles

• Searches QMetry test cases

  Lists test cases in the current QMetry project.

  Try: “List test cases in the payments folder”

  Read · qmetry_fetch_test_cases

• Opens one QMetry test case

  Returns the details of a test case.

  Try: “Show test case PAY-TC-12”

  Read · qmetry_fetch_test_case_details

• Creates a QMetry test case

  Makes a test case with steps, metadata and release or cycle mapping.

  Try: “Create a test case for password reset with three steps”

  Write · qmetry_create_test_case

• Edits or versions a QMetry test case

  Updates a test case, or creates a new version of it.

  Try: “Update the steps of PAY-TC-12”

  Write · qmetry_update_test_case

• Creates a QMetry test suite

  Makes a test suite with metadata and release or cycle mapping.

  Try: “Create a suite for the Q4 release”

  Write · qmetry_create_test_suite

• Updates many test run results at once

  Sets the execution status for one or several test case runs in bulk.

  Try: “Mark runs 1 to 10 as passed”

  Write · qmetry_bulk_update_test_case_execution_status

• Logs a defect in QMetry

  Creates a defect inside QMetry.

  Try: “Log a defect for the failed login test”

  Write · qmetry_create_defect_or_issue

• Imports automation results into QMetry

  Uploads TestNG, JUnit, Cucumber, Robot, HPUFT or QAF results, up to 30 MB, from a file path or encoded content. It is marked read-only but it adds data.

  Try: “Import this JUnit results file into QMetry”

  Write · qmetry_import_automation_test_results

• Sets the current QMetry project

  Chooses the project later QMetry calls apply to.

  Try: “Switch to the Payments project in QMetry”

  Write · qmetry_set_qmetry_project_info

• Lists Zephyr projects

  Returns the projects you can see in Zephyr.

  Try: “List my Zephyr projects”

  Read · zephyr_get_projects

• Lists Zephyr test cases

  Returns test cases in a project.

  Try: “List test cases in project SA”

  Read · zephyr_get_test_cases

• Creates a Zephyr test case

  Adds a new test case to a project.

  Try: “Create a test case for the refund flow”

  Write · zephyr_create_test_case

• Edits a Zephyr test case

  Merges your changes into the current test case so unrelated fields are kept.

  Try: “Set the priority of SA-T12 to High”

  Write · zephyr_update_test_case

• Adds or replaces the steps of a test case

  Appends steps to the end of a test case or, in overwrite mode, deletes all existing steps and replaces them. The tool tells the assistant to ask you which mode to use.

  Try: “Append two steps to test case SA-T12”

  Delete · zephyr_create_test_case_steps

• Lists Zephyr test cycles

  Returns test cycles in a project.

  Try: “List test cycles for the current release”

  Read · zephyr_get_test_cycles

• Creates a Zephyr test cycle

  Adds a new test cycle to a project.

  Try: “Create a cycle called Sprint 12 regression”

  Write · zephyr_create_test_cycle

• Records a test execution

  Creates an execution of a test case inside a test cycle.

  Try: “Add SA-T12 to the Sprint 12 cycle”

  Write · zephyr_create_test_execution

• Updates a test execution

  Changes only the fields you give on an execution, such as its status.

  Try: “Mark the SA-T12 execution as passed”

  Write · zephyr_update_test_execution

• Lists QTM4J projects

  Returns the Jira projects available in QTM4J.

  Try: “Which QTM4J projects do I have?”

  Read · qtm4j_get_projects

• Chooses the active QTM4J project

  Sets the project for the session, which later project calls need, and loads priority and status values.

  Try: “Work in project SCRUM”

  Write · qtm4j_set_project_context

• Searches QTM4J test cases

  Finds test cases in the active project.

  Try: “Find test cases about checkout”

  Read · qtm4j_search_test_cases

• Creates a QTM4J test case

  Makes a test case and resolves names for priority, status, labels and components for you.

  Try: “Create a high priority test case for guest checkout”

  Write · qtm4j_create_test_case

• Edits a QTM4J test case

  Updates a test case, with support for adding or removing labels and components.

  Try: “Add the label smoke to SCRUM-TC-5”

  Write · qtm4j_update_test_case

• Creates a QTM4J test cycle

  Makes a test cycle in the active project.

  Try: “Create a cycle for the 2.1 release”

  Write · qtm4j_create_test_cycle

• Starts a test execution

  Starts a new execution of a test case within a cycle.

  Try: “Start a new execution of SCRUM-TC-5 in cycle SCRUM-TR-101”

  Write · qtm4j_start_new_execution

• Records a test result

  Updates an execution's result, comment, environment, build, assignee or times.

  Try: “Mark that execution as failed with a comment”

  Write · qtm4j_update_test_case_execution

• Uploads automation results

  Uploads JUnit XML, TestNG XML, Cucumber JSON, QAF, HP UFT or SpecFlow results and maps them to a cycle.

  Try: “Upload this JUnit report to cycle SCRUM-TR-101”

  Write · qtm4j_upload_automation_result

• Links Jira bugs to an execution

  Attaches Jira bug keys to a test case execution.

  Try: “Link SCRUM-88 to that failed execution”

  Write · qtm4j_link_bugs_to_test_case_execution

• Removes test cases from a cycle

  Unlinks test cases from a cycle by key or filter, or all at once.

  Try: “Remove all test cases from cycle SCRUM-TR-101”

  Delete · qtm4j_unlink_test_cases_from_test_cycle

• Removes requirements from a test case

  Unlinks requirements by key, or all of them at once.

  Try: “Unlink all requirements from SCRUM-TC-5”

  Delete · qtm4j_unlink_requirements_from_test_case

• Lists Collaborator reviews

  Retrieves reviews, with any filters you give.

  Try: “Show open code reviews assigned to me”

  Read · collaborator_get_reviews

• Opens one Collaborator review

  Finds a review by its id.

  Try: “Show review 4521”

  Read · collaborator_find_review_by_id

• Starts a Collaborator review

  Creates a new review.

  Try: “Create a review for my latest change”

  Write · collaborator_create_review

• Rejects a Collaborator review

  Rejects a review with a reason.

  Try: “Reject review 4521 because tests are missing”

  Write · collaborator_reject_review

• Connects a code host to Collaborator

  Creates a remote system configuration, for example for GitHub or Bitbucket.

  Try: “Add a GitHub configuration to Collaborator”

  Write · collaborator_create_remote_system_configuration

• Deletes a code host connection

  Deletes a remote system configuration by id. It is marked read-only but it removes data.

  Try: “Delete the Bitbucket configuration”

  Delete · collaborator_delete_remote_system_configuration

• Calls any Collaborator review service method

  Invokes a review service method by name with arguments, so its reach depends on the method chosen. It is marked read-only.

  Try: “Finish the current phase of review 4521”

  Write · collaborator_reviewservice_action

Watch out for

• PactFlow tools can delete services, users and all integrations

  contract-testing_delete_pacticipant, delete_all_integrations, admin_delete_user, admin_delete_team and admin_reset_roles are all registered, along with token rotation. Use a token with no admin rights if you do not need these. github.com

• Tools can publish and overwrite live content

  swagger_publish_portal_product makes a portal product visible to visitors, swagger_create_or_update_api overwrites an API definition, and zephyr_create_test_case_steps in overwrite mode deletes all existing steps. Check each call before approving. github.com

• 15 listed write and delete tools are marked read-only

  The server marks tools read-only unless they say otherwise, and 15 tools on this page that change data still carry that mark: bearq_chat_with_qa_lead, bearq_expand_application_model, bearq_run_regression_tests, bearq_run_test_cases, bearq_run_tests_in_functional_areas, bearq_stop_task, collaborator_create_remote_system_configuration, collaborator_create_review, collaborator_delete_remote_system_configuration, collaborator_reject_review, collaborator_reviewservice_action, qmetry_import_automation_test_results, reflect_cancel_suite_execution, reflect_execute_suite and reflect_run_test. A client that auto-approves read-only tools may run them unasked. github.com

• One Collaborator tool calls any review service method

  collaborator_reviewservice_action takes a method name and arguments, so what it can do depends on what the assistant passes. Treat it as write and delete capable. github.com

• API definitions and code can go to SmartBear AI

  swagger_create_api_from_prompt, swagger_standardize_api and the PactFlow test generator send your prompt, spec or code to SmartBear AI services. Do not send anything you cannot share with SmartBear. github.com

• Local tokens sit in your client config

  The local setup keeps product tokens in the env block of your MCP config. Use tokens that reach only the projects you need and keep the file private. github.com

• Usage analytics exist, off unless you set a key

  In HTTP mode the server can report usage events to Amplitude, but only if MCP_SERVER_AMPLITUDE_API_KEY is set. The README says it never sends tool arguments or results. github.com

Security advisories