Tool catalogue
Find the tools available in your DataMaker session.
The connected server announces the authoritative tool names, arguments and response schemas. Catalogue coverage depends on the service, deployment and caller; this page groups common capabilities rather than claiming a fixed tool count.
Core metadata and generation helpers
| Tools | Purpose |
|---|---|
get_templates, get_template_by_id | Discover and inspect templates. |
create_template, create_template_from_json, create_template_from_csv | Propose or save template definitions. |
generate_from_id | Generate using a saved template. |
get_connections, get_connection_by_id | Discover database connections. |
get_endpoints, get_endpoint_by_id, list_integrations | Discover saved requests and systems. |
get_scenarios, get_scenario_by_id, save_scenario | Inspect or save reusable Python workflows. |
get_scenario_environment_variables | Inspect scenario configuration in the authorized context. |
get_sets, get_set_by_id | Inspect reusable sets. |
get_input_sets, get_input_set_by_id | Inspect input sets: columns, test cases and rows. |
list_validation_flows, get_validation_runs, get_validation_status | Inspect reviewed validation flows, results and evidence. |
start_validation_run | Queue a read-only validation of one eligible saved case. |
list_datasets, get_dataset, dataset_preview, dataset_sample, dataset_stats | Inspect chunked data and bounded summaries. |
list_plans, get_plan | Inspect reviewed specifications. |
list_masking_policies, get_masking_policy | Inspect masking policies and their field rules. |
get_uploaded_file | Get a download URL for a file uploaded earlier in the chat. |
flatten_json | Flatten a nested JSON object into dotted keys. |
derive_expectations | Derive reconciliation expectations from an extraction dataset. |
create-skill, list_skills, get_skill | Work with reusable agent instructions. |
ask_user | Request missing information through the chat UI. |
The metadata service includes selected generation helpers; other execution capabilities live on the execution services. Read the specific tool schema rather than inferring parameters from a similar REST route.
Plans and execution
The plan authoring service exposes create_plan and update_plan; update_plan patches an existing plan by block, and any edit to a plan's definition returns it to draft for re-approval. Execution services provide supported scenario execution, data jobs, schema discovery, masking and key-map workflows. See Plans, Data jobs and Masking for their lifecycle and approval requirements.
Plan runs are reported rather than started over MCP. The reporting tools are append_plan_run_log, record_plan_run_file, record_plan_run_delivery and complete_plan_run; the agent uses them to record what a run it is already executing produced. All four are on /mcp-exec/plan-run, and all but record_plan_run_delivery are also on /mcp-exec, so finding one there does not mean the run started from a tool call. Starting a run is a plan action.
Masking has three execution tools, and the difference between them is what they touch:
| Tool | What it does |
|---|---|
dry_run_masking_policy | Rehearses a policy against sample rows you supply. Nothing is persisted, and the report names uncovered fields. |
start_mask | Runs the policy over a ready dataset as a data job, writing a new masked dataset. |
unmask_values | Translates tokens back to originals. Only tokenize rules mint reversible values, and every call needs a human approval granted for that specific key map. |
Writing a policy stays a human step in Masking: its rules decide what happens to real values, and an unlisted field passes through a mask job unchanged.
System tools
Registered system namespaces provide the corresponding capabilities:
- SAP (
/mcp-exec/sap): the service catalogue and entity-set discovery, a schema knowledge graph, filtered reads, single-record export (export_to_sap_endpoint), and authoring helpers -sap-connect,create-service-endpointandcreate-template. The schema and authoring tools carry dialect-neutral names; their originalsap--prefixed spellings (sap-create-endpoint,sap-create-template,sap-schema-searchand the rest) are still registered as deprecated aliases, so prefer the generic name in new work. - Tosca (
/mcp-exec/tosca): test cases, designs and Cloud execution evidence, native authoring withtosca-create-moduleandtosca-create-testcase, and the supported on-premises object operations. - Atlassian (
/mcp-exec/atlassian): Jira and Xray Cloud discovery and test-management operations. - Signavio (
/mcp-exec/signavio): process discovery, reads and summaries. - Other registered systems: inspect the connected catalogue for the supported surface.
Each namespace is also mounted at /mcp/{namespace} for existing external clients.
Read Connections for setup. A missing tool can mean the service is not connected, disabled or unavailable in that release; it does not justify inventing a tool name.
Check a call
Before execution, confirm the active project, target ID, row count and required inputs. Afterward, inspect the result and any pending approval or error. Long-running operations can return a job identifier that needs polling; a tool returning successfully does not necessarily mean the underlying job completed.