skills.rest
breethomas
breethomasCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

agent-workflow

Design AI agent workflows with structured schemas and integration patterns.

16|3|Updated Oct 23, 2025
Search Tags:#architecture#tooling#multi-agent#mcp#orchestration#system-prompt#agent-design
One-click install
npx skills add https://github.com/breethomas/bette-think --skill agent-workflow-breethomas
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-workflow
Source: https://github.com/breethomas/bette-think/tree/main/plugins/bette-think/skills/agent-workflow
Command: npx skills add https://github.com/breethomas/bette-think --skill agent-workflow-breethomas

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Helps teams convert vague product or business needs into production-ready AI agent workflows by providing a structured design process, architecture analysis, and integration patterns that reduce scope creep, tool overload, and reliability issues.

Core Features & Use Cases

  • 9-step agent building process: Define scope, inputs/outputs, system instructions, tooling, orchestration, memory, multimedia, output formatting, and delivery.
  • 8-layer architecture framework: Evaluate infrastructure, agent internet, protocols (MCP), tooling, cognition, memory, application, and ops/governance to prioritize engineering and product tradeoffs.
  • Tooling & integration patterns: MCP, REST/API, database, document processing, and real-time data patterns with error handling and cost controls.
  • Templates & testing: System prompt templates, agent templates (research, decision, QA), orchestration patterns (sequential, parallel, iterative), and testing/validation guidance.
  • Use cases: Customer refund automation, sales qualification multi-agent pipelines, code-review assistants, and content production orchestration.

Quick Start

Design an agent that auto-approves refunds under $200 by pulling order data from Shopify, validating eligibility rules, and returning a structured JSON decision with confidence and reasoning.

Frequently Asked Questions about agent-workflow

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

FAQPage Schema
How do I design a multi-agent orchestration architecture for production?▼

Design multi-agent orchestration by applying an 8-layer architecture framework covering infrastructure, protocols, tooling, cognition, memory, application, and ops governance to prioritize engineering tradeoffs and ensure production reliability.

What is the best way to integrate MCP and REST API tools into an AI agent workflow?▼

The best way to integrate MCP and API tools is using structured integration patterns with built-in error handling and cost controls, evaluating protocol layers and tooling compatibility within your agent architecture.

How do I create a system prompt template for a specific agent task like refund automation?▼

Create system prompt templates by following a 9-step agent building process that defines scope, inputs/outputs, system instructions, tooling, memory, and output formatting specific to your refund automation use case.

Can I use this approach to build both single-agent tasks and parallel multi-agent pipelines?▼

Yes, the architecture framework supports single-agent tasks and multi-agent pipelines by providing sequential, parallel, and iterative orchestration patterns along with agent templates for research, decision, and QA workflows.

What testing and validation guidance is included for deploying reliable agent workflows?▼

Testing and deployment guidance includes validation methods for agent templates, error handling patterns, and cost optimization controls to reduce scope creep and reliability issues during production delivery.

When should I not use a multi-agent architecture for my product?▼

Avoid multi-agent architectures when a single-agent task suffices, as the 8-layer framework helps evaluate infrastructure and ops governance tradeoffs to prevent tool overload and unnecessary complexity.

Related Skills

Explore More Product & Management Skills→
DenissvgnDenissvgn

product-owner

Clarify Product Owner responsibilities and validate product vision against business goals.

Community
Basic
EnigmaWUEnigmaWU
1

implement-agent-observability

Implement trace logging, token counting, cost tracking, and loop-detection for autonomous agents.

Community
Advanced
anliang0306anliang0306
1

polymarket

Integrate Polymarket REST and WebSocket APIs with real-time RTDS subscriptions.

Community
Intermediate
UsernameTronUsernameTron

cc-ref-multi-agent

Document multi-agent orchestration patterns, token economics, and prompt engineering principles.

Community
Advanced