Use cases

One broker model across different communication paths.

PigeonMQ does not encode one business domain. Subjects and policy define the application-specific rules; communication contracts and admission semantics stay consistent across the broker.

orders.created

Service-to-service messaging

Use subject policy plus runtime contracts to bind producers and consumers to explicit operations, schema and message context.

jobs.process

Asynchronous workers

Issue bounded publish and receive authority to workers rather than treating every connected worker as indefinitely authorised.

agents.tool.invoke

AI agents and tool runners

Carry delegated intent, tool scope and message context into asynchronous agent-to-tool communication.

customer.profile.export

Regional data processing

Evaluate region and data classification constraints before a message crosses a processing boundary.

deploy.release.request

Operational automation

Apply explicit intent and environment constraints to CI, deployment and infrastructure events.

notifications.send

Domain messaging

Apply schema, sensitive-field and idempotency rules to application-specific communication paths.

Common enforcement model

Every use case follows the same protocol shape:

authenticate principal
        ↓
request subject + operation
        ↓
communication contract
        ↓
submit governed message
        ↓
identity → intent → schema → region → data → idempotency
        ↓
allow | deny | quarantine

What changes by application

The subject definition determines which operations are available and which constraints matter. A deployment event may care about environment and intent; a regional data flow may care about classification and region; an agent tool invocation may care about delegated intent and tool scope.

What stays consistent

Identity is established by the broker, authority is represented by a short-lived contract, each governed operation carries the contract ID, and broker decisions produce typed evidence.