kata-agent-orchestrator-design

Design agent orchestrator and reasoning loop with states and specialist delegation rules.

Updated Sep 3, 2025
One-click install
npx skills add https://github.com/guardiatechnology/design-system --skill kata-agent-orchestrator-design
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: kata-agent-orchestrator-design
Source: https://github.com/guardiatechnology/design-system/tree/main/.claude/skills/kata-agent-orchestrator-design
Command: npx skills add https://github.com/guardiatechnology/design-system --skill kata-agent-orchestrator-design

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It helps you design an agent’s operational orchestrator and its internal reasoning loop so the agent follows a bounded, reliable workflow rather than ad-hoc thinking.

Core Features & Use Cases

  • Orchestrator blueprint: Produces orchestrator.md with persona, scope, declared specialists, states, and a tool/workflow table.
  • Reasoning pattern selection: Chooses and justifies a reasoning loop pattern (tool-calling-simple, react, plan-and-execute, or reflexion) based on task structure and constraints.
  • Operational rigor: Ensures constraints like reasoning-pattern justification, HITL requirements, loop-state diagrams, and parameters (iterations, timeouts, sampling).

Quick Start

Ask the AI to generate an orchestrator.md and reasoning-loop.md for the agent "{agent}" using the provided overview.md and system-prompt.md, and include a justified reasoning pattern and the specialist plan.

Frequently Asked Questions about kata-agent-orchestrator-design

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

FAQPage Schema
How do I design an agent orchestrator with a bounded reasoning loop?

Designing an agent orchestrator and reasoning loop involves defining explicit states, workflow stages, specialist delegation rules, and runtime parameters to ensure a bounded, reliable workflow rather than ad-hoc thinking.

What is the best reasoning loop pattern for agent orchestration?

The best reasoning loop pattern depends on your task structure and constraints. You should choose and justify a pattern among tool-calling-simple, react, plan-and-execute, or reflexion to ensure operational rigor and reliable agent behavior.

How do I generate an orchestrator.md file for an AI agent?

To generate an orchestrator.md file, you need to specify the agent persona, scope, declared specialists, states, and a tool/workflow table. This provides a structured blueprint for production handoff and operational execution.

When do I need a plan-and-execute reasoning loop for agent engineering?

You need a plan-and-execute reasoning loop when your task structure requires multi-step planning before execution. Justifying this pattern over tool-calling-simple, react, or reflexion ensures the agent follows a bounded workflow suitable for complex operations.

How to create loop-state diagrams for agent reasoning workflows?

Creating loop-state diagrams for agent reasoning workflows requires defining explicit loop states and runtime parameters like iterations, timeouts, and sampling. These diagrams are suitable for production handoff and operational rigor.

Does agent orchestration require declared specialists and HITL constraints?

Agent orchestration requires declared specialists and HITL constraints to maintain operational rigor. Defining specialist delegation rules and human-in-the-loop requirements ensures the reasoning loop remains bounded and reliable.