zeekay
a60626ae9e
The social-orchestrator migration off upstream Temporal onto the embedded
Hanzo Tasks engine needs five workflow primitives its digest/sendEmail/post
workflows depend on. Implemented server-side (event-sourced, replay-safe) plus
the @hanzoai/tasks TS SDK (the worker social runs):
- Signal → running workflow: engine.signalWorkflow now re-dispatches a workflow
task after appending WORKFLOW_EXECUTION_SIGNALED, so the stateless worker
replays the extended history and fires the handler (digest/sendEmail are
entirely signal-driven). TS decider read the wrong history attr key
("signal" vs "signalName") — fixed.
- continueAsNew (kind=4): closes the run CONTINUED_AS_NEW + starts a successor
under the same workflowId carrying input/search-attrs/memo. TS continueAsNew
emits the command instead of throwing ContinueAsNewNotSupported.
- startChild (kind=5): detached (ABANDON) child, seq-keyed + idempotent; retires
the 0x006D 501 for the command path. TS startChild/executeChild added.
- Typed search attributes: carried on StartWorkflow/SignalWithStart wire, stored
on the execution, visibility-queryable (already read by fieldValue). TS adds
TypedSearchAttributes/defineSearchAttributeKey/SearchAttributeType.
- workflowIdConflictPolicy USE_EXISTING/TERMINATE_EXISTING/FAIL.
- ActivityFailure wraps ApplicationFailure (with .type) so workflow error
matching (err.cause.type === 'refresh_token') works.
Also fixes a pre-existing wire defect: the server registered
SignalWithStartWorkflow at 0x0069 while every client sends 0x0066 (schema +
Go/TS SDK) — signalWithStart 404'd against the real engine. Swapped with the
HTTP-only GetWorkflowHistory (no ZAP client).
DescribeWorkflow(workflowId, "") now resolves the OPEN run (Temporal's current
execution) so a continueAsNew successor is the one signalled/described.
Tests: engine unit (search attrs, conflict policy, continueAsNew, startChild)
+ TS e2e vs real tasksd (signals-across-continueAsNew, startChild, search attrs
run to COMPLETED). go test -race ./pkg/... green; TS vitest green.
Hanzo Tasks
Durable workflow execution engine for AI agent orchestration.
Introduction
Hanzo Tasks is a durable execution engine that powers AI agent orchestration in the Hanzo ecosystem. It enables developers to build scalable, fault-tolerant workflows that automatically handle intermittent failures and retry failed operations.
Tasks provides the backbone for:
- Playground spaces -- each space maps to a Tasks namespace
- Agent execution -- each agent runs as a Tasks worker
- Durable cron and batch jobs -- reliable scheduled and bulk operations
MIT licensed. See LICENSE.
Getting Started
Build
make tasksd
Run
./tasksd start
Or with a config file:
./tasksd --config-file config/development-sqlite.yaml --allow-no-auth start
Build from source
go build ./cmd/tasksd/
./tasksd start
Module
github.com/hanzoai/tasks
Integration
Hanzo Tasks integrates with the broader Hanzo ecosystem:
- Playground connects to the Tasks server via the durable-execution SDK
- Base embeds Tasks for durable cron/batch execution
- Each playground space = a Tasks namespace
- Each agent = a Tasks worker
Repository
This repository contains the source code of the Hanzo Tasks server.
The wire is luxfi/zap (binary, native, on _tasks._tcp:9999).
There is no gRPC and no go.temporal.io anywhere in the build.
To drive Tasks from Go, depend on pkg/sdk — the native
ZAP client. To embed the server in a Go process (e.g. a Hanzo Base
app), depend on pkg/tasks and call tasks.Embed().
Browser clients hit the JSON shim at /v1/tasks/* and the realtime
stream at /v1/tasks/events (Server-Sent Events). AI agents hit the
MCP surface at /v1/tasks/mcp (JSON-RPC 2.0).
Contributing
See CONTRIBUTING.md for development setup and guidelines.