dsh-agent-teams: A Multi-Agent Team Collaboration Plugin for DeepSeek Harness

dsh-agent-teams is an Agent core plugin (DSH plugin) for DeepSeek Harness, maintained by NanmiCoder, licensed under MIT, written in TypeScript, and last updated in 2026-08. It turns the current DSH session into a "captain" that can assemble durable sub-agents, split a goal into dependency-aware tasks, and coordinate work through direct messages—without requiring a separate workflow engine. Core capabilities include captain-led delegation, durable members, dependency-aware tasks, direct messaging between members, and a live activity panel Web UI. The plugin provides the team protocol, nine coordination tools, persistent state, and a Web interface, allowing users to request a team to collaborate on complex goals in natural language.

What core features does the dsh-agent-teams DSH plugin support?

The plugin supports captain-led delegation: the current session creates the team, assigns roles, and consolidates the final result. Members are continuable DSH sub-agents that can be woken for focused follow-up turns. Tasks have explicit states and cannot be claimed before their dependencies finish. Members send durable mailbox messages directly to teammates or the captain—no relay required. The Web UI shows roles, current work, unread messages, task dependencies, and archived team history in real time. Team state is stored under <workspace>/.agent-teams/; the Web panel reads that disk truth and combines it with live sub-agent activity.

How to install dsh-agent-teams?

An existing DeepSeek Harness installation is required. The recommended way is via npm:

dsh plugin --profile web add github:nanmicoder/dsh-agent-teams

You can also build from source:

git clone https://github.com/NanmiCoder/dsh-agent-teams.git
cd dsh-agent-teams
pnpm install
pnpm build
dsh plugin --profile web add .

Run pnpm build again after changing the source; the local plugin install remains linked to this checkout. Validate the composed profile and restart DSH:

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

Then ask for a team directly, for example: "Use AgentTeams to review the commits after v0.5.3 from performance, security, and product perspectives. Return one consolidated report."

How does dsh-agent-teams work?

Workflow: the current session creates a team and becomes its captain; the captain adds role-specific members backed by continuable sub-agents; the goal becomes tasks with owners and explicit dependencies; claimed tasks are dispatched through durable mailbox messages that wake each member; members work, update task state, and report directly to the captain or one another; the captain presents the combined result, then archives the complete team record. By default, member creation is zero-interaction: 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.

How to configure dsh-agent-teams?

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

- id: agent-teams
  config:
    stateDir: .agent-teams
    memberProvider: spawn
    memberModel: deepseek-v4

The plugin is maintained by NanmiCoder, licensed under MIT, and has 379 stars and 35 forks on GitHub, created on 2026-08-12 and last updated on 2026-08-15. It depends on several @deepseek-ai packages, including dsh-agent, dsh-session, dsh-subagent, and dsh-tools, ensuring deep integration with the DeepSeek Harness ecosystem.

See more DSH plugins at dsh-plugin.org