By soein
Orchestrate tmux-based AI swarms for multi-CLI roundtable discussions with @-mentions to AIs like Claude/Gemini/Codex, isolated LLM voting rounds with multi-round rebuttals/file injection for consensus, and supervisor-led execution dispatching tasks to 15 role-based agents (engineers/reviewers). Manage via bash: add/remove roles, promote discuss-to-execute, status, dispatch, stop.
npx claudepluginhub soein/swarmesh --plugin swarmesh在运行中的 discuss 圆桌加入新的 CLI 参与者
列出当前 discuss 会话的所有参与者(可 @点名的名字)
向 discuss 圆桌发一条消息,用 @name 点名触发对方接话
查看 discuss session 最近 N 轮对话(从 jsonl 读取)
启动 discuss 模式:与某个 CLI(Codex/Claude/Gemini)在一个会话内圆桌讨论
动态添加角色到运行中的蜂群(execute 模式)
从蜂群中移除角色
把 discuss 圆桌讨论的结论热升级到 execute 模式,由 supervisor 接手落地
启动蜂群协作环境(execute 模式,supervisor 编排完整蜂群)
查看蜂群当前状态和未读消息
停止蜂群并可选清理数据
向蜂群派发任务(execute 模式),默认发给 supervisor 编排
对 discuss 参与者做隔离投票——每人独立回答同一个问题,最后结构化汇总(对标 pal consensus)
Dynamically add a role to the running swarm (execute mode). Use when user says "add a role", "加个角色", "加 database engineer", "bring backend online", "need another engineer".
Remove a role from the running swarm. Use when user says "remove role", "移除 X", "kick out backend", "少一个角色", "leave role".
Run isolated multi-AI voting with LLM consensus analysis—independent judgment, no groupthink. Dispatcher for ask/list/report/next-round/cancel subcommands. Use when user asks to "vote on X", "get independent opinions", "consensus from 3 AIs", "投票决定 / 独立判断 / 三家意见 / 避免羊群效应", supports multi-round debate and file context injection.
Add a new AI CLI participant to an ongoing discuss-mode roundtable. Use when user asks to "add another AI", "bring in claude/gemini/codex", "add participant to discussion", or mentions "加入讨论" / "加参与者".
List current participants of the discuss-mode roundtable (show who can be @-mentioned). Use when user asks "who's in the discussion", "list participants", "who can I @", "看下有谁", "列出参与者".
Post a message to the discuss roundtable with @-mentions to trigger replies. Use when user wants to "send message to discussion", "@mention a participant", "post to roundtable", "触发 CLI 接话", "发消息", or says "@xxx 讨论 X".
View recent N turns of discuss session history from jsonl (without tmux attach). Use when user asks to "show discussion history", "what did they say", "回看讨论", "看下讨论历史", or needs to catch up on recent conversation.
Start a discuss-mode roundtable with multiple AI CLIs (Codex/Claude/Gemini) in a single tmux session. Use when user asks to "discuss X with multiple AIs", "start a roundtable", "brainstorm with other CLIs", "start multi-AI conversation", or mentions "discuss 模式" / "圆桌讨论" / "多 AI 讨论".
Promote discuss conclusion to execute mode—dump brief, kill discuss, start full swarm, dispatch to supervisor. Use when user says "结案" / "讨论好了开干" / "promote to execute" / "finalize discussion and dispatch" / "讨论结束开始做".
Start AI swarm execute mode with multi-role team (supervisor + engineers + reviewers etc). Use when user asks to "start the swarm", "起蜂群", "多角色协作做项目", "launch multi-role team", "需要多个角色一起做 X".
Show current swarm state (execute or discuss mode), online roles/participants, human inbox messages, task queue progress. Use when user asks "what's going on", "swarm status", "现在咋样了", "进度如何", "谁在线", "看下状态".
Stop the running swarm (execute or discuss mode) with optional data cleanup. Use when user says "stop swarm", "kill all", "停掉", "结束蜂群", "shutdown", "清理蜂群数据".
Dispatch task to swarm (execute mode) — defaults to supervisor for auto-orchestration, or target specific role. Also reads human inbox when called without args. Use when user says "do X" / "实现 Y" / "派给 supervisor" / "让团队做 Z", and when asking for swarm status/replies.
Share bugs, ideas, or general feedback.
Launch agent teams for any kind of work — code, writing, or general tasks
Multi-agent deliberation for AI coding assistants
Multi-agent orchestration for Claude Code. 12 specialized agents working in parallel — planning, building, reviewing, debugging. Plus a Hub for always-alive multi-project sessions controllable from Telegram or Slack.
Collect and synthesize opinions from multiple AI Agents for Claude Code
Multi-agent development orchestration with Claude Code + cmux. Spawn, monitor, and integrate parallel sub-agents visually in terminal panes.
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.