skills.rest
hyagoviana
hyagovianaCOMMUNITY
Explore all skills by this author →
GitHub

mcp-builder

Guide creation of MCP servers in Python or TypeScript with validated tool inputs.

1|Updated May 15, 2026
Search Tags:#mcp#model context protocol#tool design#fastmcp#agent evaluation#mcp sdk
One-click install
npx skills add https://github.com/hyagoviana/Sistema_Hyago_Viana_Adabtech --skill mcp-builder-hyagoviana
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: mcp-builder
Source: https://github.com/hyagoviana/Sistema_Hyago_Viana_Adabtech/tree/main/meu-projeto/.claude/skills/mcp-builder
Command: npx skills add https://github.com/hyagoviana/Sistema_Hyago_Viana_Adabtech --skill mcp-builder-hyagoviana

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires anthropic, mcp, and includes scripts (resource) and reference (resource) components.

What problem does it solve?

It helps you design and build high-quality MCP (Model Context Protocol) servers so LLM agents can reliably perform real tasks by using well-structured tools.

Core Features & Use Cases

  • Agent-centric tool design: Build tools around complete workflows (not just raw API endpoints) and optimize for limited agent context.
  • Protocol- and SDK-aligned implementation planning: Study MCP specs and best practices, then translate them into an actionable implementation plan.
  • Production-quality craftsmanship: Add input validation, consistent output formats, pagination/limits, and LLM-friendly, actionable error messages.
  • Testing and evaluations: Create evaluation questions and use an evaluation harness to measure whether tools actually work for realistic tasks.

Quick Start

Use the mcp-builder skill to create an MCP server that exposes tools for an external API, then evaluate the server with an XML QA file to verify the agent can answer end-to-end questions using only your MCP tools.

Frequently Asked Questions about mcp-builder

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

FAQPage Schema
How do I build an MCP server for LLM agents to use external APIs?▼

Build an MCP server by designing agent-centric tools around complete workflows using FastMCP in Python or the MCP SDK in Node/TypeScript, implementing validated inputs, consistent outputs, and actionable error messages.

What is agent-centric tool design in Model Context Protocol?▼

Agent-centric tool design in Model Context Protocol builds tools around complete external service workflows rather than raw API endpoints, optimizing for limited agent context by adding input validation, pagination, and LLM-friendly error messages.

Does the MCP SDK support TypeScript and Python for creating tools?▼

Yes, the MCP SDK supports TypeScript and Python for creating tools. You can build MCP servers using the MCP SDK in Node/TypeScript or FastMCP in Python to integrate external APIs and implement validated tool inputs and outputs.

How do I evaluate MCP tools with read-only tasks?▼

You evaluate MCP tools with read-only tasks by creating evaluation questions and using an evaluation harness to measure whether the tools actually work for realistic end-to-end agent tasks using an XML QA file.

Why do my LLM agent tools fail when handling large API responses?▼

LLM agent tools fail on large API responses when lacking robust pagination and truncation. Designing tool contracts with schemas, annotations, and limits ensures the agent can process data within its limited context window.

Related Skills

Explore More Software Engineering Skills→
arroschavesarroschaves
1

mcp-builder

Guides building MCP servers for LLM-external service integration.

Community
Advanced
ContextJet-aiContextJet-ai
29

measure-agent-task-success

Measure end-to-end success rates and task paths of AI agents using tracing data.

Official
Advanced
pkoka888pkoka888
1

mcp-builder

Build MCP servers exposing validated, agent-friendly tools for external APIs.

Community
Advanced
eduardoabreu81eduardoabreu81
6

mcp-builder

Build Model Context Protocol servers connecting AI agents to external APIs through structured tools.

Community
Advanced