dsh-agent-teamsS85

dsh plugin --profile web add @nanmicoder/dsh-agent-teams

The name after --profile is a profile name you choose yourself — swap in whichever one you use.

Once installed, saying something like "use AgentTeams to research X" in any session drives a full multi-agent team. The semantics are ported from Claude Code's AgentTeams as a four-step loop: create the team (the captain is the current session's agent), recruit members (resumable subagents), split tasks and declare dependencies between them, then let members exchange messages directly. That last step is the key design choice — messages go straight to a member's mailbox and wake them, with no relay through the captain, so the captain never becomes the bottleneck. Live team activity lives in a panel pinned to the top-right of the Web GUI, so you can see who is doing what.

• ▸Four-step loop: form the team, recruit members, split tasks with declared dependencies, then direct member-to-member messaging
• ▸Messages use direct mailbox delivery plus wake, never relayed through the captain
• ▸A persistent team activity panel in the Web GUI; the author documents version pinning, npx-without-install and GitHub-latest install paths too

移植 Claude Code AgentTeams 语义的多智能体协作插件，作者是 6 万星 MediaCrawler 维护者

Drop it into your README — it links back to this page.

Badge (inline)Copy Markdown

[![dshfind](https://dshfind.com/api/badge/NanmiCoder/dsh-agent-teams)](https://dshfind.com/en/plugins/NanmiCoder/dsh-agent-teams?ref=badge)

Card (440×200)Copy Markdown

[![dshfind](https://dshfind.com/api/card/NanmiCoder/dsh-agent-teams)](https://dshfind.com/en/plugins/NanmiCoder/dsh-agent-teams?ref=badge)