agent-team-builder

Generates multi-agent team configurations with roles, prompts, handoff rules, and test scenarios.

1|Updated Aug 8, 2026
One-click install
npx skills add https://github.com/th-efool/SKILLS --skill agent-team-builder-th-efool
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-team-builder
Source: https://github.com/th-efool/SKILLS/tree/main/agent-team-builder
Command: npx skills add https://github.com/th-efool/SKILLS --skill agent-team-builder-th-efool

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Designing a multi-agent system for a business workflow requires defining agent roles, system prompts, tool access, communication patterns, and escalation rules from scratch, which is error-prone and time-consuming without a structured process. ## Core Features & Use Cases - Interactive Discovery: Walks through process name, pain points, volume, success metrics, constraints, and integrations before any design work begins. - Team Architecture Design: Selects 3-7 agents across Coordinator, Specialist, Validator, Interface, and Data role types with hub-and-spoke, pipeline, mesh, or broadcast communication patterns. - Complete Config Generation: Produces team-config.yaml, per-agent prompt files, workflow diagrams, and test-scenarios.yaml conforming to a full schema with A2A messaging, scaling, and shared context support. - Use Case: A sales operations lead describes their lead pipeline, and the skill generates a five-agent sales team config with a lead router, qualifier, nurturer, deal strategist, and proposal writer, each with production-ready system prompts and handoff rules. ## Quick Start Ask the assistant to design an agent team for your business workflow and answer its discovery questions about your process, pain points, and integrations.

Frequently Asked Questions about agent-team-builder

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I design a multi-agent team for a business workflow?▼

Start with a discovery session covering the process name, current state, pain points, volume, success metrics, constraints, and integrations. Then select 3-7 agents across role types like Coordinator, Specialist, and Validator, choose a communication pattern, and generate the team-config.yaml with per-agent prompts.

What communication patterns can agent teams use?▼

Agent teams support four communication patterns: hub-and-spoke for sequential routing through a coordinator, pipeline for linear handoffs, mesh for collaborative work, and broadcast for notifications. The pattern is declared in the team-config.yaml under communication_pattern.

Does this skill deploy or execute the agent team it creates?▼

No, it only generates configuration files such as team-config.yaml, agent prompt files, workflow.md, and test-scenarios.yaml. The generated config is designed to be consumed by a separate agent orchestration framework for actual deployment.

What files are generated for an agent team configuration?▼

Four outputs are generated: team-config.yaml as the master configuration, individual agents/{id}/prompt.md files for each agent, workflow.md with Mermaid diagrams and routing logic, and test-scenarios.yaml with happy path and failure scenarios for validation.

How are security and secrets handled in generated agent configs?▼

Generated configurations never include API keys, passwords, or secrets directly. They use environment variable references instead, and each agent receives least-privilege tool access limited to only what its role requires.

When should I use a single agent instead of an agent team?▼

A single agent fits tasks without handoffs, multiple domains, or parallel workstreams. Teams add value when a workflow has distinct stages, needs validation gates, or requires specialized expertise, but the skill recommends starting with 3-4 agents and expanding based on performance data.