Block a user
Config-driven local/Databricks/Trino switching
Validate all 12 pipelines produce identical results on DuckDB and Databricks
Deploy Unity Catalog schemas and tables from SQLTable models
Implement stack lake load: populate Iceberg tables with real data from DuckDB
Implement stack lake deploy: create Iceberg schemas and tables from SQLTable models
Add Polaris catalog support alongside Nessie
Validate all 12 pipelines run against Trino-backed Iceberg tables
Add POST /run/{pipeline} async execution with job tracking
Add idempotent pipeline reruns with change detection
Implement stack run: execute any pipeline by name with target selection
Implement stack validate: verify all pipeline output contracts
Perfect error messages: every SchemaError includes lineage context
Build unified staging pipeline: raw files to DuckDB to Iceberg
Implement stack load seed: load seed/reference data into DuckDB
Implement stack load bcda: BCDA FHIR export ingestion into DuckDB
Implement stack load cclf: CCLF file ingestion into DuckDB
Add [api] section to stack.toml with host, port, secret, workers
Add FastAPI app with auth (JWT) and pipeline/bib/status endpoints
Build CLI skeleton with typer: stack run/load/generate/bib/lake/db/validate/docs
Wire stack.toml into all modules replacing hardcoded paths