how-to-data-dictionarylisted

Audit a data dictionary against the warehouse it claims to describe, then generate one that cannot rot - measure every structural fact from the live database, read guarantees from the data contract rather than from today's rows, hand-write only the meaning, and put a drift gate in CI so the schema can never move without the document failing a build. Every check that cannot decide says so instead of passing. Use when asked to "write a data dictionary", "document our tables", "our docs are out of date", "set up a data catalog", "what does this column mean", "which column is revenue", "onboard analysts onto the warehouse", or when someone is about to hand-maintain a column spreadsheet. Ships five standalone modules - dbprofile, claims, audit, gate, render - plus a self-contained browsable HTML dictionary and a mutation suite that proves the detectors fire.
phoebefu6/phoebe-data-skills · ★ 1 · Data & Documents · score 72

Install: claude install-skill phoebefu6/phoebe-data-skills

# how-to-data-dictionary Infrastructure-layer skill (layer 1 of the phoebe-data-skills 4-layer roadmap). A data dictionary is the only document in a data platform that everybody trusts and nobody tests. It is written once, against a schema that then moves, by a person who then leaves. Six months later it reads perfectly and is wrong in ways that each cost a real number. The fix is not a better spreadsheet. It is to split the document in three: the part a database can measure, which should never be typed by a human; the part a CONTRACT guarantees, which should never be inferred from data; and the part only a human can decide, which should never be generated. Then put a gate around all three. Showcase walkthrough (Everrest retail case, real executed charts): `docs/how-to-data-dictionary/` in https://github.com/phoebefu6/phoebe-data-skills ## Where this sits in the lineage `raw dump -> lake -> warehouse` **-> the document that says what the warehouse means**. Reads the stage-2 star schema built by `how-to-schema-and-warehouse`, and reads two documents that already describe it: the enforced `data_contract.yaml` and the signed-off `metric_registry.yaml` from `how-to-agentic-on-your-warehouse`. It does not restate either. It checks the dictionary against both, because a fourth copy of the truth is exactly how this goes wrong. ## The six rules 1. **Description coverage is a vanity metric.** No catalog checks whether a description is TRUE. In the showcase the legacy docum