SignalPilot - Governed AI Database Access
MCP Tools
The SignalPilot MCP provides governed database access:
query_database- governed read-only SQL executionvalidate_sql/explain_query/estimate_query_cost- pre-execution checksschema_overview/schema_ddl/schema_link- schema discoverydescribe_table/explore_table/explore_columns/explore_column- deep diveslist_tables/get_relationships/find_join_path- structure and joinscompare_join_types- JOIN impact analysisget_date_boundaries- date ranges across all tablescheck_model_schema/validate_model_output/audit_model_sources- dbt validationanalyze_grain- cardinality and grain analysisdebug_cte_query- step-through CTE debuggingdbt_error_parser- parse dbt error output into structured infolist_projects/get_project- dbt project managementcheck_budget/connection_health/query_history- operational
Local Scripts (via plugin)
For local dbt project work, use these standalone scripts:
scan_project.py- scan a dbt project: models, stubs, deps, hazards, work ordervalidate_project.py- rundbt parseand report structural errors
Use ToolSearch to discover additional tools as needed.
Available Skills
Load these skills as needed for specialized work:
dbt Projects
/signalpilot-dbt:dbt-workflow- Load FIRST for any dbt project. Full 7-step workflow: scan, map, validate, write, verify./signalpilot-dbt:dbt-write- Loaded at Step 2 for writing and verifying SQL models/signalpilot-dbt:dbt-debugging- Load when dbt run/parse fails
SQL (load the one matching your database)
/signalpilot-dbt:duckdb-sql- DuckDB-specific syntax and gotchas/signalpilot-dbt:snowflake-sql- Snowflake-specific patterns/signalpilot-dbt:bigquery-sql- BigQuery-specific patterns/signalpilot-dbt:sqlite-sql- SQLite-specific patterns
General SQL
/signalpilot-dbt:sql-workflow- Structured query building and verification
Quick Start
For dbt projects: Load /signalpilot-dbt:dbt-workflow - it orchestrates the
full lifecycle including scanning, mapping, writing, and verification.
For SQL queries: Load /signalpilot-dbt:sql-workflow + the SQL skill for your
database engine.
For schema exploration: Use the MCP tools directly - schema_overview for a
broad view, describe_table for column details, explore_table for sample data.