This dictionary describes the generated canonical fixture, not a recovered CloudTeams or CRM schema. Every domain is available as JSONL and CSV; adapters currently map only accounts, contacts and opportunities.
| Field family | Meaning | Contract |
|---|---|---|
id |
Stable synthetic row identity | Namespaced by seed/configuration/profile-shape fingerprint; never a vendor-native ID |
dataset_id |
Fixture/configuration identity | Constant within one generated dataset |
synthetic |
Provenance marker | Always boolean true; CSV is lower-case true |
*_id |
Foreign-key reference | Must resolve through relationships.json; unmapped target relationships remain in canonical output |
*_date, *_at |
Business date or UTC timestamp | ISO date or UTC Z; chronology is validated |
*_cents |
Monetary amount | Nonnegative integer minor units; decimal export is a presentation projection |
status |
Domain lifecycle state | Validated against domain chronology and configured transitions |
currency |
Scenario currency | Target currency mapping remains a prerequisite |
Population and story configuration live in inputs/company.json and inputs/profile-shapes.json. Templates declare chronology, relationships, currencies, identifiers and target prerequisites. A target adapter may omit fields only when its plan records the omission and preserves the canonical source.
The core CRM entities intentionally carry vendor-neutral fields that can be templated across the five foundation targets. Accounts include lifecycle, ownership, firmographic, regional and financial-context fields; contacts include role, department, language and contactability fields; leads include qualification, provenance and next-action fields; opportunities include forecast, probability, type and next-step fields. Adapters translate these into target naming conventions without asserting that a target accepts every field or that a custom field already exists.