Runs agency-orchestrator YAML workflows directly in your Claude Code or Cursor session without API keys. Point it at a YAML file defining multiple agent roles and tasks, and it'll execute them sequentially using the current LLM as each persona. It reads role definitions from markdown files (expects agency-agents-zh installed), handles dependencies between steps, and saves outputs to .ao-output with proper folder structure. If you don't have a YAML ready, describe what you want and it'll generate the workflow for you. Genuinely useful if you're already bought into the multi-agent workflow idea and want to skip the orchestration server setup.

skills.sh installs
952
GitHub stars
8.5K
last updated
Aug 21, 2026

Install

$ npx -y skills add jnmetacode/superpowers-zh --skill workflow-runner --agent claude-code

Installs into .claude/skills of the current project.

Files

工作流执行器:在 AI 工具内运行多角色编排

直接在当前会话中执行 agency-orchestrator 的 YAML 工作流,无需配置 API key。当前 LLM 就是执行引擎——依次扮演每个角色完成任务。

适用场景

  • 用户提供了一个 .yaml 工作流文件(如 运行 workflows/story-creation.yaml)
  • 用户要求多个角色协作完成任务(如"用产品经理和架构师一起评审这个 PRD")
  • 用户安装了 agency-agents-zh 并希望直接在 AI 工具内编排多角色

执行流程(5 步)

按以下顺序执行,不要跳步:

第 1 步:解析工作流

用 Read 工具读取用户指定的 YAML 文件,提取以下字段:

name: "工作流名称"
agents_dir: "agency-agents-zh"    # 角色定义目录
inputs:                            # 输入变量
  - name: xxx
    required: true/false
    default: "默认值"
steps:                             # 执行步骤
  - id: step_id
    role: "category/agent-name"    # 角色路径
    task: "任务描述 {{变量}}"       # 支持模板变量
    output: variable_name          # 输出变量名
    depends_on: [other_step_id]    # 依赖关系

忽略 llm、concurrency、timeout、retry 配置——Skill 模式使用当前会话的 LLM,这些字段仅用于 CLI 模式。

定位角色目录:用 Bash test -d 按以下顺序检查,用第一个存在的:

  1. 当前工作目录下的 {agents_dir}/(如 ./agency-agents-zh/)
  2. ../{agents_dir}/(上级目录)
  3. 相对于 YAML 文件所在目录的 {agents_dir}/
  4. node_modules/agency-agents-zh/

如果全部找不到,停止执行并提示用户:

找不到角色目录。请先安装:
  git clone --depth 1 https://github.com/jnMetaCode/agency-agents-zh.git
  或:npm install agency-agents-zh

第 2 步:收集输入

Sponsors

Details

Repository
github.com/jnmetacode/superpowers-zh
Categories
Backend & APIs, AI & Agent Building, Automation & Workflows
First seen
Jun 3, 2026

More from jnmetacode/superpowers-zh

All 20
1vapiai avatarcreate-toolvapiai/skillsCreate custom tools for Vapi voice assistants including function tools, API request tools, transfer call tools, end call tools, and integrations with Google Cal946skills.sh installs
2mindrally avatarmongodb-developmentmindrally/skillsMongoDB development guidelines with Payload CMS, Mongoose, aggregation pipelines, and TypeScript best practices.916skills.sh installs
3vtex avatarsales-app-extensibilityvtex/skillsApply when building, customizing, or deploying extensions for VTEX Sales App. Covers the complete 7-step workflow from prerequisite checks through code generati909skills.sh installs
4meshy-dev avatarmeshy-3d-agentmeshy-dev/meshy-3d-agentGenerate 3D models, textures, images, rig characters, animate them, and prepare for 3D printing using the Meshy AI API. Handles API key detection, task creation908skills.sh installs
5bobmatnyc avatarzodbobmatnyc/claude-mpm-skillsTypeScript-first schema validation library with static type inference for form validation, API validation, and runtime type checking with compile-time types.899skills.sh installs
6intellectronica avatarnano-banana-2intellectronica/agent-skillsGenerate and edit images using Google's Nano Banana 2 (Gemini 3.1 Flash Image Preview) API. This skill should be used when the user asks to create or modify ima897skills.sh installs