dtsola

Xiaoyaoclaw Agent Orchestrator — Development skill for Claude Code

Development community

OpenClaw multi-agent daily collaboration orchestrator - split tasks, dispatch via sessions_send, track progress, aggregate results, retry failures.

How to install Xiaoyaoclaw Agent Orchestrator

This entry records only its repository, not the path inside it, so there is no exact command to give. Open dtsola/xiaoyaoclaw-agent-orchestrator and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Xiaoyaoclaw Agent Orchestrator does

OpenClaw multi-agent daily collaboration orchestrator - split tasks, dispatch via sessions_send, track progress, aggregate results, retry failures. 多 Agent 日常协作编排器

Alternatives in Development

  • Codexmcp — Enable seamless collaboration between Claude Code and Codex, transforming from a single agent to multiple agen 2k ★
  • Agent Bridge — A local bridge for bidirectional collaboration between Claude Code and Codex 312 ★
  • Lab Skills — Skills distilled from the Lab's real work and collaboration practices 309 ★

README

OpenClaw Agent Orchestrator 🤝

🇨🇳 中文 | 🌐 English

OpenClaw Agent Orchestrator — multi-agent daily collaboration: split tasks, dispatch via sessions_send, track progress, aggregate results, retry failures

多 Agent 日常协作编排器——把「拆任务 → 分 agent → 管进度 → 聚结果 → 失败重试」封装成标准工作流,让多个常驻 agent 像团队一样协作。 OpenClaw multi-agent daily collaboration orchestrator — split tasks, dispatch to resident agents via sessions_send, track progress, aggregate results with source attribution, retry failures (default max 3).

![license](https://img.shields.io/badge/license-MIT-green) [![ClawHub downloads](https://img.shields.io/badge/dynamic/json?url=https%3A%2F%2Fclawhub.ai%2Fapi%2Fv1%2Fskills%2Fxiaoyaoclaw-agent-orchestrator&query=skill.stats.downloads&label=ClawHub%20downloads&color=blue)](https://clawhub.ai/dtsola/skills/xiaoyaoclaw-agent-orchestrator)

为什么需要它

你有多个 OpenClaw agent(绘图 / 调研 / 编码……),想让它们并行干活时,常见问题:

  • ❌ 手动配通道:跨 agent 要改两层配置(visibility + agentToAgent),allow 还要求双向,一堆坑
  • ❌ 手动一个个发:让 A 画图、B 调研、C 汇总,得自己挨个 sessions_send
  • ❌ 等得焦虑:sessions_send 60s 超时,不知道对方在跑还是死了
  • ❌ 手动拼结果:各 agent 回复散落各处,汇总全靠手
  • ❌ 失败没人管:某 agent 任务失败,整个流程卡住

这个 skill 一次性解决:**一句话派活 → 自动分发 → 进度透明 → 带来源汇总 → 失败自动重试(≤3 次)**。

特性

  • 🤝 强制 sessions_send:任务发到对方常驻会话——完整人格 + 记忆 + 技能,像真实同事一样干活
  • 🚦 三档触发:点名/编排动词直接执行;模糊大任务先问一句;日常问答保持沉默——从不抢活
  • 🔀 并行分发:多个子任务同时发出(fire-and-forget),总时长 ≈ 最慢那个,不串行等
  • 📡 进度透明:sessions_list / sessions_history 实时查状态,60s 超时 ≠ 失败(任务后台继续)
  • 📊 带来源汇总:聚合报告标注「哪条结论来自哪个 agent」,可追溯
  • 🔁 失败重试:默认最多 3 次(可配置),重试带上下文(不重复已完成部分);3 次仍败上报用户
  • 📖 直接读 openclaw.json:agents.list + agentToAgent.allow 双向白名单,真相源只有一个,不搞探测
  • 🐍 零依赖脚本:配置检测 + 状态汇总,纯 Python 标准库,Windows / macOS 双平台

安装

# ClawHub(推荐)
clawhub install xiaoyaoclaw-agent-orchestrator

# 或从 GitHub 手动安装
git clone https://github.com/dtsola/