Language

• Block names use business language.
• Commands are imperative verb phrases.
• Events are past-tense facts.
• Ambiguous words are clarified or marked with hotspots.
• The same word does not silently mean different things inside one bounded context.

Commands

• Every command has an initiator.
• Every command targets one primary aggregate or external system.
• Every command has required input fields.
• Every command has expected success event or events.
• Every command has known rejection reasons.
• Authorization is clear for sensitive commands.

Events

• Every event has an owning context.
• Every event has a source aggregate, policy, or external system.
• Event payloads include the facts consumers need.
• Events are not named after technical implementation details.
• Events that cross boundaries are modelled as integration events.

Aggregates And Rules

• Every aggregate has an identity.
• Every aggregate lists the commands it handles.
• Every aggregate lists the events it emits.
• Invariants are visible and testable.
• Aggregates are not oversized collections of unrelated behaviour.
• Child entities and value objects are inside the correct aggregate boundary.

Policies And Processes

• Every policy has a trigger event.
• Every policy has a condition or says it always applies.
• Every policy has a resulting command, notification, or integration event.
• Long-running policies are promoted to sagas or process managers.
• Retries, deadlines, and failure paths are clear for automated workflows.

Read Side

• Every important user decision has a read model.
• Every read model has a consumer.
• Every read model has update sources.
• Queries do not change business state.
• Read model freshness expectations are clear.

Boundaries And Integrations

• Bounded contexts are named after business capabilities.
• External systems have owners and integration types.
• Integration events have publishers, subscribers, and contract payloads.
• Failure handling is defined for unreliable external systems.
• Privacy-sensitive data crossing boundaries is reviewed.

Generation Readiness

• Generation scope is clear.
• Persistence and hosting options are selected.
• Open hotspots are resolved, downgraded, or accepted as risk.
• Test scenarios exist for key commands, events, policies, and read models.
• Business owner and technical owner agree the model is ready.