agent-ecosystem-generator

This Skill addresses the cumbersome and error-prone process of manually setting up multi-agent systems. In practice, developers often need to design orchestration patterns, define sub-agent responsibilities, configure MCP tools, and ensure all components work together—a process typically requiring significant repetitive effort and debugging.

Usage is straightforward: simply install this plugin in your Claude Code session and describe your requirements in natural language. For example, input "Design a code review team with security, performance, and testing inspectors." It will automatically launch a requirements interview sub-agent, generate specifications, select the most suitable orchestration pattern (e.g., orchestrator-workers), and ultimately produce a complete, installable plugin or project configuration. The entire process includes requirement clarification, pattern selection, file generation, and semantic validation, supporting up to three rounds of automatic fixes.

It is ideal for teams or individuals needing to build complex multi-agent systems, especially those already using Claude Code for AI-assisted development and looking to orchestrate multiple skills, sub-agents, and MCP tools into a unified workflow. Project managers and AI application architects can also use it to quickly validate multi-agent collaboration scenarios.

For first-time use, start with a simple scenario, such as "Generate a three-person PR review team," to familiarize yourself with the workflow. Note that it is currently in the alpha stage. Project modes and custom MCP paths have limited test coverage; the core process has been end-to-end validated, but complex custom scenarios may require manual adjustments.