Prompt file imported from Avanade/CoreEx (
.github/prompts/coreex-api-e2e.prompt.md). Copyright stays with the author.
Guide this workspace through adding a complete new entity with CRUD API endpoints end-to-end.
Use .github/skills/coreex-api-e2e/SKILL.md and its referenced workflow as the authoritative workflow.
Operational contract:
- Read the solution-root
AGENTS.mdFeature Configuration first (rop-enabled,outbox-enabled,data-provider—SqlServer/Postgres/Cosmos/None) and check whethersrc/*.Domain/exists — do not ask questions already answered there. - Batch all interview questions into one turn: entity name, CRUD operations, key properties, identifier type, whether a business guard is needed, and (for any ref-data property) whether that reference-data type already exists.
- Invoke L1 skills in sequence —
coreex-refdata(only if a brand-new reference-data type was identified),coreex-aggregate(only when*.Domainproject exists),coreex-contract,coreex-db-migration,coreex-repository,coreex-validator, optionallycoreex-policy,coreex-app-service,coreex-api,coreex-test-api— passing resolved context so no L1 re-asks what was already answered. - Decide the identifier type once, upfront — pass the same value to
coreex-aggregate(TId) andcoreex-contract(IIdentifier<T>). - Skip migration and repository when
data-provider = None. - If any prompt text conflicts with the skill, the skill wins.
Outcome:
- A complete vertical slice from database table to HTTP endpoint is in place, with unit tests for the validator (and policy if applicable) and integration tests for all operations.
