Skip to content

Tool catalogue

The DataMaker MCP server exposes the tools below. The in-product agent gets these plus tools from any connected servers: the SAP/OData server, the Tosca server, and the code-execution server (see the bottom of this page). For setup, see MCP server.

Read / discovery

ToolWhat it does
get_templatesAll templates in the active project.
get_template_by_idFull definition of one template.
get_connectionsAll connections in the project.
get_connection_by_idOne connection’s details (URL, auth method, never secrets).
get_endpointsAll REST endpoints in the project.
get_endpoint_by_idOne endpoint’s details.
get_scenariosAll scenarios in the project.
get_scenario_by_idOne scenario’s source and metadata.
get_scenario_environment_variablesA scenario’s environment variables.
list_integrationsConnected integrations (auth-bearing systems, e.g. SAP gateways); returns id, name, kind, origin. Use this to get an integration_id before discovering its services.

Create / generate

ToolWhat it does
create_template_from_jsonInfer fields from a sample JSON object/array and save a template.
create_template_from_csvFetch a CSV (e.g. an S3 URL), AI-detect field types, and save a template.
generate_from_idGenerate rows from a template by ID.
save_scenarioCreate or update a scenario.

Plans

DataMaker’s Plan mode lets the agent design a reviewable test-data plan before running anything. A plan is the spec that compiles into a runnable scenario, not generated data.

ToolWhat it does
create_planSave a reviewable plan (flow, coverage, gaps, entities) the user can open and run.
list_plansList the plans in the project (id, title, status).
get_planFetch one plan by id, including its full spec.

Utility

ToolWhat it does
ask_userAsk a clarifying question rendered as clickable choices; used when a required identifier is ambiguous.
flatten_jsonFlatten a nested JSON object so nested keys join with dots ({user:{name}}{"user.name"}).

Tools from connected servers

The in-product agent also loads tools from separate MCP services registered against each session:

  • SAP / OData (datamaker-sap): discover services from $metadata, read entities with $filter, and write back to entity sets (with CSRF handling).
  • Tricentis Tosca (datamaker-tosca): read Tosca test-case designs and write test data back into Tosca, for both Tosca Cloud and on-prem Tosca Commander (see the table below).
  • Code execution (datamaker-cnt): run the Python a scenario needs.

These tools are provided by their own services, so their exact names and schemas live with those services rather than the core datamaker server. Ask the agent “what tools do you have?” in a chat, or inspect the MCP server’s announced tool list, for the authoritative set in your deployment.

Tosca tools (datamaker-tosca)

Back the agent with a Tricentis Tosca connection. integration_id is optional on the read tools; omit it to auto-select the project’s connected Tosca system.

ToolWhat it does
tosca-list-testcasesList the test cases in a connected Tosca system. Works for Cloud and on-prem; pass space to pick a workspace.
tosca-read-testcase-designRead a test case’s full design (configuration parameters, step folders, module/attribute references, value ranges). The input for data-gap analysis.
tosca-create-testcaseCreate a new Tosca Cloud test case from a (modified) design payload. Used to act on a gap analysis.
tosca-onprem-list-workspacesList the workspaces on an on-prem Tosca Commander system. Needed before querying or writing.
tosca-onprem-queryRun a TQL search and return matching object ids. The on-prem way to locate folders, sheets, and execution lists before reading or writing.
tosca-onprem-create-objectCreate a child object (TestSheet, Attribute, Instance, ExecutionList, …) under a parent, optionally setting attributes. The headline DataMaker → Tosca write.
tosca-onprem-add-execution-entryAdd a test case to an on-prem execution list.
tosca-onprem-get-objectRead an on-prem object’s attributes, e.g. to inspect execution-result verdicts after a run.

Example call

{
"name": "generate_from_id",
"arguments": { "id": "<template-id>", "count": 25 }
}

The exact argument names and options for every tool are part of the schema the MCP server announces on connect; always trust that over a static list.