@headless-lms/adapter-workflow-hatchet

Hatchet (hatchet.run) implementation of the AutomationEngine port from @headless-lms/core/types — durable execution for dispatched automations.

The adapter declares one workflow, automation-run: a durable parent task that iterates the dispatched actions, spawning one child task per action (automation-run-action, retries: 3) through Hatchet's durable event log. A child that exhausts its retries stops the sequence; the parent then calls finalize exactly once with the results collected so far. Because children are spawned via the durable event log, a crash mid-run resumes after the last completed action instead of restarting the sequence from scratch.

import { HatchetAutomationEngine } from "@headless-lms/adapter-workflow-hatchet";

const container = await createContainer(config, {
  adapters: { automationEngine: new HatchetAutomationEngine() },
});

The Hatchet client reads its own environment (HATCHET_CLIENT_TOKEN and friends — see the Hatchet docs); this adapter reads none of it itself. Pass a pre-built client (or a fake, satisfying HatchetClientLike) as the first constructor argument for tests or non-default setups.

Environment variables (Hatchet client, read by the SDK)

| Variable | Required | Notes | | --------------------- | -------- | ---------------------------------------- | | HATCHET_CLIENT_TOKEN | yes | Hatchet API token for the target tenant. |

See the Hatchet TypeScript SDK docs for the full set of HATCHET_CLIENT_* variables (host/port, TLS, namespace, …).