Does it support human-in-the-loop approvals?

Yes, it includes patterns for 'requires_confirmation' tool configurations, allowing agents to pause for human approval before executing sensitive or dangerous operations.

How does the skill handle agent memory?

It provides guidance on configuring ConversationMemory for short-term turns, VectorMemory for long-term RAG-based storage, and SummaryMemory for condensing long-running dialogues.

Can I use this skill for multi-agent orchestration?

Yes, this skill provides specific implementation patterns for WorkflowAgents (multi-step sequencing) and A2A agents (agent-to-agent communication) designed for complex coordination.

Does this skill support non-OpenAI models?

Yes, it includes configuration patterns for a variety of providers including Azure OpenAI, Anthropic Claude, Google Gemini, and local models via Ollama.

What is the difference between ChatAgent and BaseAgent?

ChatAgent is the standard type for conversational AI, offering built-in memory and automatic tool calling. BaseAgent is a low-level class used when you need full control over message preprocessing, custom protocols, or agent adapters.

Microsoft Agent Type Implementation

Name: Microsoft Agent Type Implementation
Author: pagerguild

bypagerguild

0•

データサイエンスとML

Implements and configures specialized agent types within the Microsoft Agent Framework for diverse conversational and workflow-driven AI applications.

This skill provides Claude Code with expert patterns and implementation logic for the Microsoft Agent Framework. It guides developers through selecting and coding various agent architectures, including ChatAgents for conversational support, BaseAgents for low-level message control, WorkflowAgents for multi-step orchestration, and A2A agents for collaborative multi-agent environments. By leveraging this skill, developers can ensure best practices in model selection, memory management, and tool configuration while building resilient, production-ready AI agent systems.

主な機能

01Configuration guidance for memory management including Vector and Summary memory

02Advanced tool configuration with human-in-the-loop approval workflows

030 GitHub stars

04Agent-to-agent (A2A) communication and discovery protocol implementation

05Standard implementation patterns for ChatAgent, BaseAgent, and WorkflowAgent

06Lifecycle management and error handling strategies for resilient agent execution

ユースケース

01Orchestrating multi-step document processing pipelines via specialized WorkflowAgents

02Building conversational customer support assistants with complex tool-calling requirements

03Designing collaborative agent ecosystems where specialized agents delegate tasks through A2A protocols

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add pagerguild/guilde-lite ms-agent-types

For use in Claude.ai and ChatGPT

Download Skill