docs/zh-CN/skills/dmux-workflows/SKILL.md
使用dmux(AI代理的tmux窗格管理器)进行多代理编排。跨Claude Code、Codex、OpenCode及其他工具的并行代理工作流模式。适用于并行运行多个代理会话或协调多代理开发工作流时。
npx skillsauth add affaan-m/everything-claude-code dmux-workflowsInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
使用 dmux(一个用于代理套件的 tmux 窗格管理器)来编排并行的 AI 代理会话。
dmux 是一个基于 tmux 的编排工具,用于管理 AI 代理窗格:
n 创建一个带有提示的新窗格m 将窗格输出合并回主会话安装: npm install -g dmux 或参见 github.com/standardagents/dmux
# Start dmux session
dmux
# Create agent panes (press 'n' in dmux, then type prompt)
# Pane 1: "Implement the auth middleware in src/auth/"
# Pane 2: "Write tests for the user service"
# Pane 3: "Update API documentation"
# Each pane runs its own agent session
# Press 'm' to merge results back
将研究和实现拆分为并行轨道:
Pane 1 (Research): "研究 Node.js 中速率限制的最佳实践。
检查当前可用的库,比较不同方法,并将研究结果写入
/tmp/rate-limit-research.md"
Pane 2 (Implement): "为我们的 Express API 实现速率限制中间件。
先从基本的令牌桶算法开始,研究完成后我们将进一步优化。"
# Pane 1 完成后,将研究结果合并到 Pane 2 的上下文中
在独立文件间并行工作:
Pane 1: "创建计费功能的数据库模式和迁移"
Pane 2: "在 src/api/billing/ 中构建计费 API 端点"
Pane 3: "创建计费仪表板 UI 组件"
# 合并所有内容,然后在主面板中进行集成
在一个窗格中运行测试,在另一个窗格中修复:
窗格 1(观察者):“在监视模式下运行测试套件。当测试失败时,
总结失败原因。”
窗格 2(修复者):“根据窗格 1 的错误输出修复失败的测试”
为不同任务使用不同的 AI 工具:
Pane 1 (Claude Code): "Review the security of the auth module"
Pane 2 (Codex): "Refactor the utility functions for performance"
Pane 3 (Claude Code): "Write E2E tests for the checkout flow"
并行审查视角:
Pane 1: "审查 src/api/ 中的安全漏洞"
Pane 2: "审查 src/api/ 中的性能问题"
Pane 3: "审查 src/api/ 中的测试覆盖缺口"
# 将所有审查合并为一份报告
对于涉及重叠文件的任务:
# Create worktrees for isolation
git worktree add -b feat/auth ../feature-auth HEAD
git worktree add -b feat/billing ../feature-billing HEAD
# Run agents in separate worktrees
# Pane 1: cd ../feature-auth && claude
# Pane 2: cd ../feature-billing && claude
# Merge branches when done
git merge feat/auth
git merge feat/billing
| 工具 | 功能 | 使用时机 | |------|-------------|-------------| | dmux | 用于代理的 tmux 窗格管理 | 并行代理会话 | | Superset | 用于 10+ 并行代理的终端 IDE | 大规模编排 | | Claude Code Task 工具 | 进程内子代理生成 | 会话内的程序化并行 | | Codex 多代理 | 内置代理角色 | Codex 特定的并行工作 |
ECC 现在包含一个助手,用于使用独立的 git worktree 进行外部 tmux 窗格编排:
node scripts/orchestrate-worktrees.js plan.json --execute
示例 plan.json:
{
"sessionName": "skill-audit",
"baseRef": "HEAD",
"launcherCommand": "codex exec --cwd {worktree_path} --task-file {task_file}",
"workers": [
{ "name": "docs-a", "task": "Fix skills 1-4 and write handoff notes." },
{ "name": "docs-b", "task": "Fix skills 5-8 and write handoff notes." }
]
}
该助手:
seedPaths 覆盖到每个工作器的工作树中.orchestration/<session>/ 下写入每个工作器的 task.md、handoff.md 和 status.md 文件当工作器需要访问尚未纳入 HEAD 的脏文件或未跟踪的本地文件(例如本地编排脚本、草案计划或文档)时,使用 seedPaths:
{
"sessionName": "workflow-e2e",
"seedPaths": [
"scripts/orchestrate-worktrees.js",
"scripts/lib/tmux-worktree-orchestrator.js",
".claude/plan/workflow-e2e-test.json"
],
"launcherCommand": "bash {repo_root}/scripts/orchestrate-codex-worker.sh {task_file} {handoff_file} {status_file}",
"workers": [
{ "name": "seed-check", "task": "Verify seeded files are present before starting work." }
]
}
tmux capture-pane -pt <session>:0.<pane-index> 检查它。brew install tmux (macOS) 或 apt install tmux (Linux) 安装。development
Share durable, inspectable context and handoffs between Claude, Codex, Hermes, Cursor, OpenCode, and other agents through the local ECC Memory Vault. Use when an agent must save work state, transfer context, resume another agent's task, or search shared project knowledge.
development
Use when multiple consumers and providers must evolve an API or event schema without field drift, integration surprises, or one side silently redefining the interface.
tools
Query live GPU inventory, submit an authenticated Itô fixed-rate RFQ, inspect RFQ or procurement status, and run explicitly gated node qualification through the separately installed canonical CLI. Use when a user asks to find H100/H200 capacity, request a fixed compute rate, check Itô compute status, or validate GPU nodes.
data-ai
Instinct-based learning system that observes sessions via hooks, creates atomic instincts with confidence scoring, and evolves them into skills/commands/agents. v2.1 adds project-scoped instincts to prevent cross-project contamination.