Open Multi-Agent Files

Breaking changes

This release ships as a minor version, so a caller on a ^1.x range receives it without an explicit upgrade step. A caller that passes malformed task metadata or disables thinking on MiniMax-M3.1-Flash still installs, then fails when the code runs. The individual entries are under Changed and Fixed below.

• Team.addTask() now honors every task option it accepts. Options that were previously discarded — retry settings, memory scope, dependency payload, role, priority, metadata, capability requirements, and consensus verification — now affect queue ordering, retries, capability gating, memory scope, and consensus verification. Malformed metadata that was previously ignored now fails validation before the task is enqueued.
• MiniMax-M3.1-Flash cannot disable thinking. A call with thinking: { enabled: false } now throws before the request, where earlier versions forwarded it with thinking ignored. Use MiniMax-M3 when disabled thinking is required, or pass an explicit extraBody.thinking override.

Added

• MiniMax-M3.1-Flash support, including adaptive thinking, reasoning-effort forwarding, and native reasoning replay on tool-call follow-ups.
• The MiniMax adapter now reports tool-use-only native reasoning echo support for follow-up requests.

Changed

• New MiniMax configurations default to MiniMax-M3.1-Flash; saved configurations, explicit MiniMax-M3, legacy IDs, and custom model IDs continue to be forwarded unchanged.
• For MiniMax-M3, enabled thinking now maps to adaptive and disabled thinking to disabled; reasoning effort remains ignored for M3.
• The oma provider reference and default-model hint for MiniMax now name MiniMax-M3.1-Flash, so newly generated provider templates use that default.
• The core package README, which ships in the core tarball, now summarizes capabilities and quick starts more concisely.

Fixed

• Team.addTask() now forwards every accepted task option instead of only title, description, assignee, and dependencies.
• Malformed task metadata now fails createTask() validation before a task enters the queue instead of being silently dropped.

Compatibility

• Existing explicit MiniMax-M3, legacy, and custom MiniMax model IDs retain pass-through behavior; only newly generated configurations switch to Flash. Generated projects that need M3 should set model explicitly.
• Audit every Team.addTask() call site: options that previously had no effect now affect retries, priority, capability requirements, memory scope, and consensus verification, and consensus or judge settings can add model calls and cost.
• Ensure task metadata satisfies TaskMetadata, including homogeneous arrays; malformed metadata that was previously ignored now causes addTask() to throw.
• No public export or signature was removed or narrowed, and no engines or package-dependency change ships; upgrading needs no code change beyond the two behavior changes above.
• @open-multi-agent/otel is not republished: it did not change, and core 1.22.0 stays inside its ^1.11.0 range.

Packages

• @open-multi-agent/core: 1.22.0
• @open-multi-agent/otel: remains at 0.1.4 and is not republished
• create-oma-app: 0.8.11; generated starters pin core 1.22.0

Thanks

• @March-7: preserve task options in Team.addTask (#633)
• @octo-patch: default new configurations to M3.1 Flash (#631)

Install

:::bash
npm i @open-multi-agent/core@1.22.0
npm create oma-app@latest my-oma