Planner's own reporting is thin, and turning plans, buckets and tasks into SQL tables makes questions like "what is overdue across every plan" answerable in one query. The discipline is the same as the rest of this family: discover the schema, then query it, and nothing can write. The CData driver licence is the prerequisite to settle first.
A local MCP server built on the CData JDBC Driver for Microsoft Planner, which exposes Planner as relational SQL models. A model gets table discovery, column discovery and one read-only query tool. It runs over `stdio`, so the client has to be on the same machine.
- The Microsoft Planner tables the driver exposes, returned as CSV with a header row — `msplanner_get_tables`
- The columns of a chosen table, so queries reference real field names — `msplanner_get_columns`
- A SQL `SELECT` against live Planner data — `msplanner_run_query`
- A `Tables` property in the connection file that limits what is exposed; leave it blank and everything is exposed
A Microsoft Planner account and a licensed CData JDBC Driver for Microsoft Planner — a trial licence works, applied with `java -jar cdata.jdbc.msplanner.jar --license`. Java to run the server, Maven to build it (`mvn clean install` produces `CDataMCP-jar-with-dependencies.jar`). The `.prp` connection file names `DriverClass` as `cdata.jdbc.msplanner.MSPlannerDriver`, the driver jar path, a tool `Prefix` such as `msplanner`, and the JDBC connection string generated by the driver's Connection String utility — since Planner uses OAuth, that step sends you through a browser sign-in. MIT licensed.