AK Plan
saltbo/agent-kanban
Breaks a project into dependency-aware Tasks on an Agent Kanban board through Realmroot Toolbox, previews them for approval, then follows each Task through review.
A skill your agent uses when a request requires multi-agent workflow orchestration (task decomposition + dependency/DAG + parallel execution), needs durable task tracking across context compaction…
Install Team Dispatch
$ npx skills add LeoYeAI/openclaw-master-skills --skill team-dispatch -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install LeoYeAI/openclaw-master-skills team-dispatch --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/LeoYeAI/openclaw-master-skills.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/team-dispatch .claude/skills/team-dispatch && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "team-dispatch" agent skill from https://github.com/LeoYeAI/openclaw-master-skills/tree/main/skills/team-dispatch into .claude/skills/team-dispatch/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "team-dispatch", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/LeoYeAI/openclaw-master-skills/tree/main/skills/team-dispatchType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add LeoYeAI/openclaw-master-skills --skill team-dispatch -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install LeoYeAI/openclaw-master-skills team-dispatch --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/LeoYeAI/openclaw-master-skills.git skills-src && mkdir -p .agents/skills && cp -r skills-src/skills/team-dispatch .agents/skills/team-dispatch && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "team-dispatch" agent skill from https://github.com/LeoYeAI/openclaw-master-skills/tree/main/skills/team-dispatch into .agents/skills/team-dispatch/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "team-dispatch", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add LeoYeAI/openclaw-master-skills --skill team-dispatch -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install LeoYeAI/openclaw-master-skills team-dispatch --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/LeoYeAI/openclaw-master-skills.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/skills/team-dispatch .cursor/skills/team-dispatch && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "team-dispatch" agent skill from https://github.com/LeoYeAI/openclaw-master-skills/tree/main/skills/team-dispatch into .cursor/skills/team-dispatch/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "team-dispatch", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/LeoYeAI/openclaw-master-skills.git --path skills/team-dispatch--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add LeoYeAI/openclaw-master-skills --skill team-dispatch -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install LeoYeAI/openclaw-master-skills team-dispatch --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/LeoYeAI/openclaw-master-skills.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/skills/team-dispatch .gemini/skills/team-dispatch && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "team-dispatch" agent skill from https://github.com/LeoYeAI/openclaw-master-skills/tree/main/skills/team-dispatch into .gemini/skills/team-dispatch/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "team-dispatch", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install LeoYeAI/openclaw-master-skills team-dispatchInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add LeoYeAI/openclaw-master-skills --skill team-dispatch -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/LeoYeAI/openclaw-master-skills.git skills-src && mkdir -p .github/skills && cp -r skills-src/skills/team-dispatch .github/skills/team-dispatch && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "team-dispatch" agent skill from https://github.com/LeoYeAI/openclaw-master-skills/tree/main/skills/team-dispatch into .github/skills/team-dispatch/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "team-dispatch", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add LeoYeAI/openclaw-master-skills --skill team-dispatch -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install LeoYeAI/openclaw-master-skills team-dispatch --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/LeoYeAI/openclaw-master-skills.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/skills/team-dispatch .opencode/skills/team-dispatch && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "team-dispatch" agent skill from https://github.com/LeoYeAI/openclaw-master-skills/tree/main/skills/team-dispatch into .opencode/skills/team-dispatch/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "team-dispatch", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Facts
- Skill name
team-dispatch- GitHub stars
- 2.2k
- Token cost
- ~3.2k tokens
- SKILL.md length
- 681 words
- Files
- 98 (incl. scripts, references, assets)
- Skills in repo
- 1,215
- Repo updated
- First seen
- Licence
- MIT
At a glance
A skill your agent uses when a request requires multi-agent workflow orchestration (task decomposition + dependency/DAG + parallel execution), needs durable task tracking across context compaction…
- Works in 4 steps: 首次派发 → 收到 completion event → 立刻执行调度循环 → Prompt 构建 — 结果注入 → …
- A request requires multi-agent workflow orchestration (task decomposition + dependency/DAG + parallel execution)
- SKILL.md covers ⚡ 第〇步：自动安装与环境检测（必须首先执行）, 🌍 国际化（i18n）, 📊 任务分级 and 🔀 拆解模板, plus 2 more sections
- Calls bash
What it does
Team Dispatch is an agent skill from LeoYeAI/openclaw-master-skills. Use when a request requires multi-agent workflow orchestration (task decomposition + dependency/DAG + parallel execution), needs durable task tracking across context compaction, or needs retries/fallbacks and completion notifications.
Its SKILL.md is about 3.2k tokens, which your agent loads only when the skill is triggered. The skill folder holds 103 other files, including scripts, reference files and assets (for example `CHANGELOG.md`, `README.md` and `README.zh.md`).
It sits in Agent Workflows, covering Task breakdown, Context engineering and Task management. The repository describes itself as: 🧠 Curated collection of 1209+ best OpenClaw skills — weekly updated by MyClaw.ai. The licence is MIT.
When your agent uses it
- A request requires multi-agent workflow orchestration (task decomposition + dependency/DAG + parallel execution)
- Needs durable task tracking across context compaction
- Needs retries/fallbacks and completion notifications
Example prompts
- “/team-dispatch”
Workflow steps
4 steps, taken from the step headings in SKILL.md.
What it can do on your machine
Read from SKILL.md and the folder at commit e5199b5. It shows what the files ask for, not the result of running them.
Tool permissions
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Runs code
Ships 1 file in scripts/, which the agent can run.
Shell commands in SKILL.md call:
bash
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Context cost
Team Dispatch loads about 3.2k tokens when it runs, and up to ~4.2k if it reads all its reference files. Until then it costs about 62 tokens; SKILL.md has 681 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
Safety
Auto-check passed
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.
SKILL.md
The full file from LeoYeAI/openclaw-master-skills at commit e5199b5, republished under its MIT licence (© LeoYeAI). 681 words, ~3,210 tokens.
Download SKILL.mdSave it as
.claude/skills/team-dispatch/SKILL.md (or your agent's skills folder). This skill also uses 97 other files; get the full folder from GitHub.- name
- team-dispatch
- description
- Use when a request requires multi-agent workflow orchestration (task decomposition + dependency/DAG + parallel execution), needs durable task tracking across context compaction, or needs retries/fallbacks and completion notifications.
🚀 Team Dispatch v1.0.8 — Multi-Agent Workflow Orchestration
一句需求 → 自动分析 → 智能拆解 → DAG 派发 → 故障重试 → 自动交付。
⚡ 第〇步：自动安装与环境检测（必须首先执行）
真实项目代码空间（默认）
- 默认真实项目/仓库目录：
~/work（可通过<SKILL_DIR>/config.json或~/.openclaw/configs/team-dispatch.json的paths.projectsRoot修改） - 约定：每个项目在
paths.projectsRoot/<project>下创建与产出代码（而不是放在~/.openclaw/workspace/） setup.sh会自动创建该目录（若不存在）
Agent 读到本文件后，必须先执行以下检测流程，确认系统就绪后再处理用户需求。 新用户首次安装、或已有用户技能更新后，均走此流程。
SKILL_DIR 解析
<SKILL_DIR> = 本 SKILL.md 文件所在的目录。
Agent 应从自身读取本文件的路径自动推导，例如：
- 若从
/Users/xxx/skills/team-dispatch/SKILL.md读取，则SKILL_DIR=/Users/xxx/skills/team-dispatch - 若从软连接
~/.openclaw/skills/team-dispatch/SKILL.md读取，用readlink解析真实路径
一键安装
直接运行安装脚本即可完成全部配置：
bash <SKILL_DIR>/scripts/setup.sh如果遇到模型 404/quota 问题，可追加基线模型参数：
bash <SKILL_DIR>/scripts/setup.sh --baseline-models安装脚本自动完成的 8 项配置
setup.sh 会自动检测并配置以下所有项目，已存在的不覆盖：
| # | 配置项 | 说明 |
|---|---|---|
| 1 | 软连接 | ~/.openclaw/skills/team-dispatch → <SKILL_DIR> |
| 2 | 任务目录（使用安装用户的主工作区） | ~/.openclaw/workspace/tasks/{active,done,templates}（不会创建额外的 main 工作区） |
| 3 | 项目模板 | 复制 project.json 到 templates/ |
| 4 | 用户配置 | 生成 ~/.openclaw/configs/team-dispatch.json（语言、通知策略、团队显示名） |
| 5 | 子 Agent agentDir 模板 | 全新安装：cp -R <SKILL_DIR>/assets/agents/<id> ~/.openclaw/agents/<id>（内含 workspace 模板 + 预置文件） |
| 6 | 补齐缺失文件 | 已有 ~/.openclaw/agents/<id>：仅补齐 workspace/ 下缺失文件，不覆盖用户改动；并创建 sessions/ |
| 7 | 写入 openclaw.json | 在 agents.list 中确保存在 main（dispatcher/root）并配置 main.subagents.allowAgents: ["*"]；同时写入 7 个子 Agent 配置（含 workspace、identity、model），已存在的补齐缺失字段 |
| 8 | 重启 Gateway | 配置写入后自动重启 Gateway 使生效 |
子 Agent 完整配置规范
在 openclaw.json 的 agents.list 中：必须存在 main（dispatcher/root），并在 main 上配置 subagents.allowAgents。
main agent 配置示例（安装时应确保存在/补齐）
{
"id": "main",
"default": true,
"name": "main",
"workspace": "/Users/vvusu/.openclaw/workspace",
"agentDir": "/Users/vvusu/.openclaw/agents/main/agent",
"model": "openai-codex/gpt-5.4",
"identity": {
"name": "调度台",
"emoji": "🎯"
},
"subagents": {
"allowAgents": ["*"]
}
}子 Agent 配置示例
{
"id": "<agentId>",
"name": "<中文名>",
"workspace": "~/.openclaw/agents/<agentId>/workspace",
"model": {
"primary": "<见下方模型选择策略>",
"fallbacks": ["bailian/qwen3.5-plus", "bailian/kimi-k2.5", "zai/glm-4.7"]
},
"tools": { "profile": "<coding|full>" },
"skills": [],
"identity": {
"name": "<中文名>",
"emoji": "<角色 emoji>",
"theme": "<英文角色描述>"
}
}模型选择策略
coder→openai-codex/gpt-5.3-codex（OAuth，专为编码优化）- 其他 Agent →
openai-codex/gpt-5.4（OAuth，通用稳定） - fallbacks 从用户已配置的 providers 中选取，确保不会 404
工具集分配
coder,tester→"coding"product,research,trader,writer→"full"
子 Agent Workspace 结构
每个子 Agent 拥有独立的 agentDir + workspace + sessions（符合 OpenClaw 多 Agent 目录规范），互不干扰。
全新安装（您指定的逻辑）
cp -R <SKILL_DIR>/assets/agents/<id> ~/.openclaw/agents/<id>模板目录结构为：
<SKILL_DIR>/assets/agents/<id>/
└── workspace/
├── AGENTS.md
├── SOUL.md
├── IDENTITY.md
├── USER.md
├── TOOLS.md
├── HEARTBEAT.md
├── BOOTSTRAP.md
└── .openclaw/workspace-state.json安装后运行时会在以下位置写入会话与状态：
~/.openclaw/agents/<id>/sessions/已有安装（升级/补齐）
- 若
~/.openclaw/agents/<id>已存在：只补齐workspace/下缺失文件，不覆盖用户已修改的文件 sessions/若缺失会自动创建（本技能脚本也会mkdir -p）
注：子 Agent 通过
sessions_spawn启动时只注入AGENTS.md+TOOLS.md，但完整的 workspace 文件在 Agent 作为独立 agent（而非 subagent）运行时会全部加载。
团队成员一览
| agentId | 中文名 | Emoji | workspace | 工具集 | 角色定位 |
|---|---|---|---|---|---|
coder | 闪电 | ⚡️ | workspace-coder | coding | 编码开发专家 |
product | 诺娃 | 🧭 | workspace-product | full | 产品规划专家 |
tester | 亚特拉斯 | 🔍 | workspace-tester | coding | 测试验证专家 |
research | 露娜 | 🔭 | workspace-research | full | 调研搜索专家 |
trader | 泰坦 | 📈 | workspace-trader | full | 投资分析专家 |
writer | 萊拉 | ✒️ | workspace-writer | full | 内容写作专家 |
shield | 盾卫 | 🛡️ | workspace-shield | full | 安全审计专家 |
检测通过后的输出
安装脚本或 Agent 检测全部通过后，向用户确认：
✅ Team Dispatch 环境就绪！
- 软连接：✅
- 任务目录：✅
- 项目模板：✅
- 用户配置：✅（语言：zh/en）
- 子 Agent Workspace：✅（7 个独立工作目录 + AGENTS.md）
- 子 Agent 配置：✅（workspace + identity + model 完整）
- Gateway：✅（已重启生效）
系统已准备好接收任务，请告诉我您想要完成什么？常见安装问题与自动修复
| 问题 | 原因 | 修复方式 |
|---|---|---|
| 子 Agent 没有独立 workspace | 旧版技能未配置 | 重新运行 setup.sh，自动创建并补齐 |
| 子 Agent 没有角色指令 | 缺少 AGENTS.md | setup.sh 从 assets/agents/ 复制模板 |
| 子 Agent 不知道自己是谁 | 缺少 identity 配置 | setup.sh 自动写入 identity 字段 |
| 用户配置不完整 | 手动创建了简化版 | 用 <SKILL_DIR>/config.json 覆盖 |
| 模型 404 / quota 错误 | 模型不可用 | 运行 setup.sh --baseline-models |
🌍 国际化（i18n）
默认英文，通过 ~/.openclaw/configs/team-dispatch.json 中 "language": "zh" 切换中文。
📊 任务分级
收到需求后，先判断复杂度：
| 级别 | 判断标准 | 处理方式 |
|---|---|---|
| S | 单 Agent 可完成 | 直接 sessions_spawn，不建文件 |
| M | 2-3 个 Agent，线性依赖 | 自动建 DAG + 追踪 |
| L | 4+ Agent，有并行分支 | 自动建 DAG + 追踪 + 进度汇报 |
| XL | 跨多领域，需多轮迭代 | DAG + 分阶段交付 + 用户确认点 |
🔀 拆解模板
根据需求类型自动选择：
开发类
product(PRD) → coder(编码) → tester(测试) → shield(安全审计)
→ writer(文档)安全敏感类（带部署）
product(PRD) → coder(编码) → tester(测试) → shield(安全审计) → writer(文档)研究类
research(调研) → product(分析框架) → writer(成文)全栈类
research(调研) → product(PRD) → coder(编码) → tester(测试) → shield(安全审计) → writer(文档)分析类
research(数据收集) → trader(分析) → writer(报告)
→ product(策略建议)内容类
research(素材收集) → writer(初稿) → product(审核优化)⚠️ 重要限制：sessions_spawn 可能不允许指定 agentId
在某些运行环境/工具策略下，sessions_spawn 可能会拒绝 agentId 参数（错误类似：agentId is not allowed for sessions_spawn (allowed: none)）。
此时 Team Dispatch 必须自动降级为“单 Agent 多角色模拟”模式：
- 仍然按 DAG 拆解任务
- 派发时不传
agentId（让 subagent 继承当前 agent 身份） - 在 task prompt 顶部注入角色指令（例如“你现在扮演 coder/product/tester…”）来模拟团队分工
- 结果收集、依赖解锁、重试逻辑保持不变
如果你有权限配置并希望启用真实多 Agent 派发：
- 在
openclaw.json中设置agents.defaults.subagents.allowAgents: ["*"]（或允许列表） - 重启 Gateway
⚙️ 核心调度流程
⚠️ 关键行为约定：事件驱动的自动调度循环
当你通过
sessions_spawn派发任务后，OpenClaw 会在 subagent 完成时自动向你的会话推送一条 completion event（以 user message 形式到达）。你必须在收到每一条 completion event 后，立刻执行完整的调度循环：
- 从 event 中提取 label，匹配到对应的 task
- 更新该 task 的 status/result/completedAt，写入 JSON
- 扫描所有 pending 任务，找出依赖已全部 done 的任务
- 立刻派发这些就绪的任务（sessions_spawn）
- 检查项目是否全部完成，若完成则归档并通知用户
不要等待用户指令，不要等待其他事件，不要跳过这个循环。每收到一条 completion event 就执行一次。
这是 Team Dispatch 的核心机制——事件驱动的级联派发。如果你不执行这个循环，任务就会"卡住"。
Show full SKILL.md (259 more words)Show less
Step 1: 首次派发
扫描项目 JSON 中所有 status=pending 且依赖全部 done（或无依赖）的任务，批量派发：
for each task where status=="pending" && allDepsAreDone(task):
task.status = "in-progress"
task.startedAt = now()
prompt = buildPrompt(task)
sessions_spawn(agentId=task.agentId, task=prompt,
runTimeoutSeconds=task.timeoutSeconds,
label="{project}:{taskId}")
task.sessionKey = response.childSessionKey
writeJSON()Step 2: 收到 completion event → 立刻执行调度循环
当 subagent completion event 到达时（OpenClaw 自动推送，格式为 "✅ Subagent {agentId} finished\n{result}"），你必须立刻执行以下完整流程，不需要用户指令：
# ① 识别完成的任务
task = findByLabel(event.label) # 或从 event 内容匹配 agentId + 上下文
# ② 更新状态
task.result = summarize(event.result)
task.status = "done"
task.completedAt = now()
writeJSON()
# ③ 扫描并派发就绪的下游任务（关键！）
dispatchReady():
for each t where t.status=="pending" && allDepsAreDone(t):
t.status = "in-progress"
t.startedAt = now()
prompt = buildPrompt(t) # 含上游结果注入
sessions_spawn(agentId=t.agentId, task=prompt, ...)
t.sessionKey = response.childSessionKey
writeJSON()
# ④ 检查项目是否全部完成
checkProjectDone():
if all tasks are done/skipped:
project.status = "completed"
moveToArchive() # tasks/active → tasks/done
finalizeAndNotify() # 向用户发送最终交付清单再次强调：Step 2 是一个自动触发的循环。你不需要也不应该等用户说"继续"或"下一步"。每收到一个 completion event，就执行一轮 ②→③→④。这样任务会像多米诺骨牌一样自动级联推进，直到全部完成。
Step 3: Prompt 构建 — 结果注入
下游 Agent 自动获得上游输出：
buildPrompt(task):
prompt = task.description
if task.dependsOn:
prompt += "\n\n--- 上游任务结果 ---"
for depId in task.dependsOn:
dep = findTask(depId)
prompt += "\n\n[{dep.id}] ({dep.agentId}): {dep.result}"
return promptStep 4: 并行派发注意事项
当多个任务同时就绪时（如 t4-tests 和 t5-docs 都依赖 t3-core），应同时派发所有就绪任务，而非逐个等待：
# 错误做法 ❌：派发 t4，等 t4 完成，再派发 t5
# 正确做法 ✅：同时派发 t4 和 t5，各自独立完成后触发下游🛡️ 故障处理
故障类型与自动恢复
| 类型 | 触发条件 | 自动处理 |
|---|---|---|
| 超时 | 超过 timeoutSeconds | 超时 × 1.5 后重试 |
| 失败 | Agent 返回 failed | 重试（最多 retryLimit 次） |
| 拒绝 | 并发上限 | 进入 queued，自动出队 |
| 模型错误 | 模型不可用(404等) | 检查修正模型配置 |
降级策略（onFailure 字段）
| 策略 | 值 | 行为 |
|---|---|---|
| 阻塞 | "block" | 中止项目，通知用户（默认） |
| 跳过 | "skip" | 标记 skipped，下游继续 |
| 降级 | "fallback" | 换备选 Agent 重试 |
| 人工 | "manual" | 暂停，等用户提供结果 |
📈 状态机
项目状态
active → completed (全部 done/skipped)
active → blocked (关键任务失败，重试耗尽)
active → cancelled (用户取消)
blocked → active (用户干预后恢复)任务状态
pending → in-progress → done
→ queued → in-progress (并发满)
→ in-progress → failed → pending (重试)
→ skipped (跳过)🔒 并发控制
- 并发上限由
agents.defaults.subagents.maxChildrenPerAgent/maxConcurrent决定（默认 5，建议 10） - 超出时任务进入
queued状态，有 Agent 完成后自动出队 - 优先级：关键路径 > 依赖链长 > 声明顺序
🔗 依赖模式
| 模式 | 图示 | 场景 |
|---|---|---|
| 线性 | A → B → C | Bug 修复、简单功能 |
| 扇出 | A → B, C, D | 一个输入多个并行处理 |
| 扇入 | B, C → D | 多个结果汇合 |
| 菱形 | A → B, C → D | 先分后合 |
⏸️ 用户确认点（XL 级）
关键节点暂停等用户审核：
{
"id": "review-prd",
"agentId": null,
"description": "请用户确认 PRD 是否符合预期",
"status": "pending",
"dependsOn": ["prd"],
"type": "checkpoint"
}checkpoint 类型不派发 Agent，暂停并通知用户。
📁 项目文件
存储路径：<workspace>/tasks/
tasks/
├── active/ # 进行中
│ └── <project>.json
├── done/ # 已完成
│ └── <project>.json
└── templates/
└── project.json # 任务模板🔧 运维工具
| 脚本 | 用途 | 命令 |
|---|---|---|
scripts/setup.sh | 一键安装/配置 | bash <SKILL_DIR>/scripts/setup.sh |
scripts/setup.sh --baseline-models | 切换稳定基线模型 | 解决 404/quota 问题 |
scripts/setup-config.sh | 生成用户配置 | 自动跳过已存在 |
scripts/doctor.sh | 环境健康检查 | bash <SKILL_DIR>/scripts/doctor.sh |
scripts/demo-project.sh | 生成闭环自测 Demo 项目（product → coder → tester） | bash <SKILL_DIR>/scripts/demo-project.sh |
scripts/watch.sh | 低频巡检卡死任务（前台运行） | INTERVAL=300 GRACE=20 bash <SKILL_DIR>/scripts/watch.sh |
scripts/watch-install.sh | 安装 watcher（后台常驻，跨平台：macOS/Linux；Windows见ps1） | bash <SKILL_DIR>/scripts/watch-install.sh |
scripts/watch-uninstall.sh | 卸载 watcher（后台常驻） | bash <SKILL_DIR>/scripts/watch-uninstall.sh |
assets/windows/watch-install.ps1.txt | Windows 安装 watcher（Scheduled Task） | copy <SKILL_DIR>\\assets\\windows\\watch-install.ps1.txt watch-install.ps1; powershell -ExecutionPolicy Bypass -File .\\watch-install.ps1 |
scripts/demo-project.sh | 生成协作开发测试 Demo 项目（写入 tasks/active） | bash <SKILL_DIR>/scripts/demo-project.sh |
✅ 闭环自测（最小协作链路）
- 生成一个带 3 个任务的 Demo DAG（写入
~/.openclaw/workspace/tasks/active/）：
bash <SKILL_DIR>/scripts/demo-project.sh- 然后在 main agent 对话中说：
- “请用 team-dispatch 跑一下刚生成的 demo-collab-loop 项目（product → coder → tester），并在完成后归档到 tasks/done。”
预期：main 会按依赖顺序依次
sessions_spawn(product/coder/tester)，并把结果写回项目 JSON，最终移动到tasks/done/并汇总汇报。
✅ 闭环自测（协作开发测试 Demo）
生成一个最小可验证的协作开发 DAG 项目（默认 team-mvp-1），写入：~/.openclaw/workspace/tasks/active/：
bash <SKILL_DIR>/scripts/demo-project.sh生成后，在主 Agent 里发一句话触发调度（示例）：
请用 team-dispatch 跑一下刚生成的 team-mvp-1 demo 项目，按 DAG 依赖派发 product/coder/tester/writer，完成后归档到 tasks/done。
📋 完整调度示例
输入
用户: "帮我调研 AI Agent 市场，写一篇分析报告"自动处理
1. 分析: 研究类，L 级
2. 拆解: research → product(框架) → writer(成文) + coder(可视化)
3. 建 DAG: tasks/active/ai-report.json
4. 派发: research（无依赖）
5. research done → 派发 product（注入 research 结果）
6. product done → 并行派发 writer + coder（注入上游结果）
7. 全部完成 → 归档 tasks/done/ → 汇报用户输出
✅ 项目完成！
- 📝 分析报告: [内容]
- 📊 可视化: ~/work/ai-report/index.html
- 🔎 预览方式: open ~/work/ai-report/index.html© LeoYeAI, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Files
SKILL.md and 97 other files (scripts, references, assets) in skills/team-dispatch of LeoYeAI/openclaw-master-skills.
- SKILL.md
- CHANGELOG.md
- README.md
- README.zh.md
- _meta.json
- assets/agents/coder.md
- assets/agents/coder/workspace/AGENTS.md
- assets/agents/coder/workspace/BOOTSTRAP.md
- assets/agents/coder/workspace/HEARTBEAT.md
- assets/agents/coder/workspace/IDENTITY.md
- assets/agents/coder/workspace/SOUL.md
- assets/agents/coder/workspace/TOOLS.md
- assets/agents/coder/workspace/USER.md
- assets/agents/product.md
- assets/agents/product/workspace/AGENTS.md
- … and 83 more
Open the folder on GitHubat commit e5199b5
Compare with similar skills
Team Dispatch next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Team Dispatch this skillLeoYeAI/openclaw-master-skills | 2.2k | — | ~3.2k | Automated safety check: Pass | MIT | |
| AK Plansaltbo/agent-kanban | 485 | — | ~903 | Automated safety check: Pass | Custom licence | |
| MemPalace Task HandoffMemPalace/mempalace | 59k | — | ~1.9k | Automated safety check: Pass | MIT | |
| OpenRig Software Factorymvschwarz/openrig | 5.5k | — | ~2.5k | Automated safety check: Pass | Apache-2.0 | |
| agtx One-Shot Project Runnerfynnfluegge/agtx | 1.7k | — | ~3.8k | Automated safety check: Pass | Apache-2.0 | |
| Project Developmentguanyang/open-agent-hub | 973 | 2 repos | ~4.7k | Automated safety check: Pass | MIT |
Similar skills
MemPalace Task Handoff
MemPalace/mempalace
Creates, hands off, claims, executes and closes agent tasks through the MemPalace logstream, with approval of the exact task before it is recorded.
OpenRig Software Factory
mvschwarz/openrig
Helps set up a continuing agent software team for a real repository with OpenRig, choosing between manual work, queue handoffs and an explicit Workflow.
agtx One-Shot Project Runner
fynnfluegge/agtx
Runs a whole project unattended on an agtx kanban board, decomposing the goal, starting tasks, unblocking workers and merging each result.
Project Development
guanyang/open-agent-hub
This skill should be used for project-level decisions about LLM-powered systems: whether an LLM is the right primitive for the task at hand, the shape of a multi-stage batch or agent pipeline, token…
Agtx Task Sweep
fynnfluegge/agtx
Breaks a conversation's results into feature-level tasks and pushes them to the agtx kanban board, where each task gets its own worktree and agent session.
More from LeoYeAI/openclaw-master-skills
All 1,215 skills in this repoDevOps Pipeline Management
LeoYeAI/openclaw-master-skills
Manages pipelines on a DevOps quality and efficiency platform through its OpenAPI: list workspaces and templates, create, update, run and cancel pipelines, and read run records.
Feishu Document Collaboration
LeoYeAI/openclaw-master-skills
Patches OpenClaw's Feishu extension so an edited document triggers an isolated agent session that reads the doc and replies inline, turning it into a live chat space.
Files Memory System
LeoYeAI/openclaw-master-skills
Multi-context memory management system for OpenClaw agents with group-isolated storage, global shared memory, workspace organization, and group-specific skills isolation.
GEO-Claw AI Visibility Agent
LeoYeAI/openclaw-master-skills
Runs a brand's AI-search visibility work end to end: diagnosing how AI platforms represent it, repositioning it, producing AI-optimized content and monitoring ongoing mentions.
Google Workspace CLI
LeoYeAI/openclaw-master-skills
Installs and authenticates the gws CLI, then automates Gmail, Drive, Sheets, Calendar, Docs, Chat and Tasks with ready-made recipes, persona bundles and security audits.
HealthFit Health Advisors
LeoYeAI/openclaw-master-skills
Runs four advisor roles, a fitness coach, nutritionist, data analyst and TCM practitioner, to build a health profile and track workouts, diet and wellness over time.
Related
Categories
Questions about Team Dispatch
What does Team Dispatch do?
A skill your agent uses when a request requires multi-agent workflow orchestration (task decomposition + dependency/DAG + parallel execution), needs durable task tracking across context compaction…. Team Dispatch is an agent skill from LeoYeAI/openclaw-master-skills. Use when a request requires multi-agent workflow orchestration (task decomposition + dependency/DAG + parallel execution), needs durable task tracking across context compaction, or needs retries/fallbacks and completion notifications.
When should I use Team Dispatch?
Team Dispatch fits situations like: A request requires multi-agent workflow orchestration (task decomposition + dependency/DAG + parallel execution); needs durable task tracking across context compaction; needs retries/fallbacks and completion notifications.
How do I install Team Dispatch in Claude Code?
Run `npx skills add LeoYeAI/openclaw-master-skills --skill team-dispatch -a claude-code`. Or copy the skill folder (skills/team-dispatch in LeoYeAI/openclaw-master-skills) into .claude/skills/team-dispatch in your project. Claude Code loads it when a task matches its description.
How do I install Team Dispatch in Codex?
Run `npx skills add LeoYeAI/openclaw-master-skills --skill team-dispatch -a codex`. Or copy the skill folder (skills/team-dispatch in LeoYeAI/openclaw-master-skills) into .agents/skills/team-dispatch in your project. Codex loads it when a task matches its description.
Can I use Team Dispatch in Cursor, Gemini CLI or GitHub Copilot?
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add LeoYeAI/openclaw-master-skills --skill team-dispatch -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/team-dispatch, .gemini/skills/team-dispatch, .github/skills/team-dispatch and .opencode/skills/team-dispatch in your project.
What does Team Dispatch need to run?
Going by SKILL.md and its folder, Team Dispatch needs the command-line tools its instructions call (bash).
Does Team Dispatch access the network?
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Is Team Dispatch safe to install?
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.
What licence does Team Dispatch use?
Team Dispatch is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
How many tokens does Team Dispatch use?
About 3.2k tokens (SKILL.md is roughly 13k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 976 tokens, read only when the agent opens those files.
What are the alternatives to Team Dispatch?
Skills that share tags, products or a category with Team Dispatch: AK Plan (saltbo/agent-kanban, 485 stars), MemPalace Task Handoff (MemPalace/mempalace, 59k stars), OpenRig Software Factory (mvschwarz/openrig, 5.5k stars) and agtx One-Shot Project Runner (fynnfluegge/agtx, 1.7k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Team Dispatch?
LeoYeAI (a GitHub user) maintains it in LeoYeAI/openclaw-master-skills, which has 2,158 GitHub stars. The repository holds 1,215 skills in this directory. The repository was last updated on July 20, 2026.
Source: LeoYeAI/openclaw-master-skills on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.