One prompt. A working team.

dsh-agent-teams turns the current DeepSeek Harness session into a captain that can assemble durable sub-agents, split a goal into dependency-aware tasks, and coordinate work through direct messages.

Ask in natural language. The plugin provides the team protocol, ten coordination tools, persistent state, and a live Web UI—without requiring a separate workflow engine.

Why AgentTeams?

| Capability | What it changes | | --- | --- | | Captain-led delegation | The current session creates the team, assigns roles, and consolidates the final result. | | Durable members | Members are continuable DSH sub-agents that can be woken for focused follow-up turns. | | Dependency-aware tasks | Tasks move through explicit states and cannot be claimed before their dependencies finish. | | Direct messaging | Members send durable mailbox messages directly to teammates or the captain—no relay required. | | Live activity panel | The Web UI shows roles, current work, unread messages, task dependencies, and archived team history. |

Install

[!NOTE] Requires an existing DeepSeek Harness installation.

dsh plugin --profile web add github:StarPivotNet/dsh-plugins-public#path:packages/agent-teams

Then restart dsh web. This package is the StarPivot share copy. Source lineage is recorded in UPSTREAM.md.

Validate the composed profile, restart DSH, and refresh the Web UI:

dsh --profile web --dump-config
dsh web

Then ask for a team directly:

Use AgentTeams to review the commits after v0.5.3 from performance, security, and product perspectives. Return one consolidated report.

How it works

1. The current session creates a team and becomes its captain.
2. The captain adds role-specific members backed by continuable sub-agents. Each add creates and claims that member's first task.
3. Later tasks use only returned ids (t1, t2, …) and assignees that already exist.
4. The spawn brief starts the first task. Later mailbox messages barge in with new instructions; pass mode=queue only to wait out the current turn.
5. Members work, update task state, and report directly to the captain or one another.
6. The captain presents the combined result, then archives the complete team record.

Team state is stored under <workspace>/.agent-teams/; the Web panel reads that disk truth and combines it with live sub-agent activity.

Member creation is zero-interaction by default: the plugin snapshots the LLM provider, model, and reasoning effort actually used by the captain's current step, and restores that snapshot on later continuations. Only an explicit heterogeneous-team request (for example, “backend on provider A/model X, frontend on provider B/model Y”) supplies a member-specific provider + model; there is no per-member model or reasoning prompt.

Configuration

Defaults work without extra setup. A trusted profile can override member behavior:

- id: agent-teams
  config:
    stateDir: .agent-teams
    memberProvider: spawn
    memberModel: deepseek-v4
    memberMaxDepth: 1
    maxMembers: 8

memberProvider is the sub-agent runtime backend (spawn / fork), not an LLM provider. Cross-LLM-provider routing uses the optional provider + model fields of agent_teams_add_member; memberModel is only a model default for all members.

Boundaries

• One captain leads one active team at a time.
• Members act only after they are woken; mail remains durable while a participant is idle.
• State is file-backed and serialized within one DSH process; concurrent processes editing the same team are not coordinated.
• The activity panel reports persisted state as-is. Models may occasionally finish work without performing the expected task-state update.

See docs/usage.md for the full tool reference, state model, Web UI behavior, configuration, and known limits.

Plugin development Skill

The repository also ships the open Agent Skills package dsh-plugin-development:

The skill still lives in this package under skills/dsh-plugin-development.

Documentation

| Guide | Covers | | --- | --- | | Usage | Architecture, UI behavior, tools, configuration, limits, and validation | | Self-iteration | Captain-owned plugin-defect reports filed to this fork's issue tracker | | Verification | Offline, composition, real e2e, and GUI verification | | Plugin development | Human-readable guide built from this plugin | | README writing | Repository documentation conventions |

Development

pnpm install
pnpm build
pnpm verify

License

MIT