Supported Tools¶
The TestGen MCP server exposes the prompts, tools, and resources listed below.
For setup instructions, see Set up the MCP Server. For example questions to ask an assistant, see MCP Server.
Prompts¶
Prompts are pre-built workflows you can invoke directly through your AI client — typically as a slash command (for example, /testgen:table_health in Claude Code) or from a quick-action menu. They orchestrate several tool calls behind the scenes for common investigations. Exact UX varies by client.
compare_runs— Compare the two most recent test runs to identify regressions and improvements.health_check— Run a data quality health check across all projects and test suites.hygiene_triage— Guided hygiene issue triage workflow — review hygiene issues and decide what to do.investigate_failures— Investigate test failures to identify root causes and patterns.profiling_overview— Explore the profiling results for a table group — understand data shapes, types, null rates, and hygiene issues.table_health— Assess the data quality health of a specific table across all test suites.
Tools¶
Tools are operations the assistant calls during a conversation, picked based on what you ask.
Discover what TestGen knows about¶
get_data_inventory— Get a structural inventory of all projects, connections, table groups, and test suites accessible to the authenticated user.get_test— Get full details of a test definition, including configuration, parameters, and last result.get_test_type— Get detailed information about a specific test type.list_projects— List all projects the authenticated user has access to.list_tables— List tables in a table group.list_test_notes— List notes attached to a test definition, newest first.list_test_suites— List all test suites for a project with their latest run statistics.list_test_types— List available test types with optional filtering.list_tests— List test definitions in a test suite.
Investigate quality issues¶
get_failure_summary— Summarize test failures (Failed and Warning) grouped by test type, table, or column.get_failure_trend— Time-series of test result counts by time bucket — use this to see whether failures are trending up or down.get_recent_test_runs— Get the latest test runs for each test suite in a project, optionally filtered by test suite name.get_source_data— Look up rows from the connected database that match or violate a test definition's criteria.get_source_data_query— Get the SQL query that would be used to look up source data for a test definition, without executing it.get_test_result_history— Get the historical results of a specific test definition across runs, showing how measure and status changed over time.get_test_run_diff— Compare two test runs and report regressions, improvements, persistent failures, and added/removed tests.list_test_results— List individual test results for a test run, with optional filters.search_test_results— Search test results across multiple runs with flexible filters.
Browse profiling results¶
get_hygiene_issue— Get full details of a specific hygiene issue.get_table— Get an overview of a table with profiling highlights: structural metadata, column list, quality scores, and hygiene issue count from the latest profiling run.list_column_profiles— List per-column profile headers (~14 fields each) — the Layer 1 scan of profiling results across columns in a table group.list_hygiene_issues— List hygiene issues for a profiling run.list_profiling_summaries— List aggregated profiling health summaries for a table group or across a project — quality scores, hygiene issue counts, record counts, last profiled date.search_hygiene_issues— Search hygiene issues across profiling runs and table groups.update_hygiene_issue— Update the disposition of a hygiene issue (confirm, dismiss, or mute).
Trigger profiling, tests, and test generation¶
cancel_profiling_run— Request cancellation of a queued or running profiling run.cancel_test_run— Request cancellation of a queued or running test run.generate_tests— Submit a test-generation job for a test suite. Auto-creates test definitions from the latest profiling results for the table group; locked and manually created test definitions are preserved. Returns immediately with a job_execution_id.run_profiling— Submit a profiling run for a table group. Returns immediately with a job_execution_id; uselist_profiling_summariesto track status.run_tests— Submit a test run for a test suite. Returns immediately with a job_execution_id; useget_recent_test_runsto track status.
Resources¶
Resources are static reference documents that AI clients can fetch by URI.
testgen://glossary— Glossary of TestGen concepts, entity hierarchy, result statuses, and quality dimensions.testgen://hygiene-issue-types— Reference table of all hygiene issue types with their data quality dimensions, descriptions, and suggested actions.testgen://test-types— Reference table of all test types with their descriptions and data quality dimensions.