openserv-multi-agent-workflows
由 issa-me-sush 提供
Web & Frontend Development简介
Multi-agent workflow examples to work together on the OpenServ Platform.
描述
该技能的功能
Provides patterns and examples for building multi-agent workflows on the OpenServ Platform. Covers agent discovery via the marketplace, task dependency chains, and workflow graph construction using edges. Includes runnable pipeline examples from a simple 2-agent blog workflow to a 6-agent life coaching system.
Workflow edges enforce actual execution order and data flow between agents, which task dependencies alone do not guarantee on the OpenServ Platform.
何时使用
- Automating a research-then-write blog pipeline with two marketplace agents
- Running parallel content creation across research, writing, and image generation agents
- Discovering specialized agents by capability and wiring them into a new workflow
- Setting up conditional branching so rejected tasks route to a handler instead of proceeding
- Combining fan-out agent outputs into a single final task for report generation
示例工作流
User asks: Build a two-agent pipeline that researches a topic and writes a blog post
- 1
Call client.authenticate() to establish a session
- 2
Search for research and writing agents using client.agents.listMarketplace({ search: 'research' }) and listMarketplace({ search: 'content writer' })
- 3
Create the workflow with client.workflows.create(), defining a research task and a writing task with the writing task listing the research task in its dependencies
- 4
Define edges connecting the trigger to the research task and the research task to the writing task
- 5
Call workflows.sync() to deploy and execute the pipeline
A running pipeline where the research agent processes the topic first and passes its output to the writing agent, which produces the final blog post
前置要求
- OpenServ Platform account
安装
clawhub install issa-me-sush/openserv-multi-agent-workflows运行 CLI 命令,或把这段提示词粘贴到 Claude Code、Cursor、Codex 或你在用的任意编程智能体里。