agency-mcp-builder

Designs, builds, and tests Model Context Protocol servers with typed tools, resources, and prompts.

Updated Jul 27, 2026
One-click install
npx skills add https://github.com/imMamdouhaboammar/Mimera --skill agency-mcp-builder-immamdouhaboammar
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: agency-mcp-builder
Source: https://github.com/imMamdouhaboammar/Mimera/tree/main/.agents/skills/specialized-mcp-builder
Command: npx skills add https://github.com/imMamdouhaboammar/Mimera --skill agency-mcp-builder-immamdouhaboammar

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? AI agents often fail to use custom tools correctly because of vague names, untyped parameters, and poor error handling. This Skill guides the design and implementation of MCP servers so agents can reliably discover, call, and act on external APIs, databases, and services. ## Core Features & Use Cases - Agent-Friendly Tool Design: Enforces verb-noun tool naming, descriptive documentation, and typed parameters with Zod (TypeScript) or Pydantic (Python) so agents pick the right tool on the first try. - Production-Quality Server Implementation: Provides runnable TypeScript and Python MCP server templates with structured error handling, environment-based secrets, and stateless tool design. - Resources, Prompts, and Testing: Covers exposing data as MCP resources, creating prompt templates, and testing the full agent tool-call loop including error paths. - Use Case: You need to give an AI agent access to your support ticket system. Use this Skill to build an MCP server exposing search_tickets and get_ticket_stats tools with validated inputs and actionable error messages. ## Quick Start Ask the agent to design and implement an MCP server that exposes tools for your API, including typed parameters, descriptions, and error handling.

Frequently Asked Questions about agency-mcp-builder

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

FAQPage Schema
How do I build an MCP server in TypeScript?

Use the official @modelcontextprotocol/sdk package with McpServer and StdioServerTransport. Register tools with server.tool(), defining a name, description, and Zod schema for typed parameters, then connect via a stdio transport.

How do I build an MCP server in Python?

Use the FastMCP class from the mcp.server.fastmcp package. Decorate async functions with @mcp.tool() and type parameters with Pydantic Field descriptions so agents receive validated, self-documenting inputs.

What makes a good MCP tool name and description?

Use unambiguous verb-noun names like search_tickets_by_status instead of generic names like query. Descriptions should state when to use the tool, not just what it does, since agents select tools based on name and description alone.

Should I use TypeScript or Python for MCP server development?

Both are officially supported. TypeScript uses the MCP SDK with Zod validation and suits Node.js ecosystems, while Python uses FastMCP with Pydantic and fits data or API-heavy stacks. Choose based on your existing runtime and team expertise.

How do I handle errors in MCP tools?

Wrap external calls in try/catch and return content with isError set to true plus an actionable message. Never crash the server or leak stack traces, so the agent can retry, adjust parameters, or ask the user.

When should I use MCP resources instead of tools?

Use resources to expose read-only context like stats or documents that agents should read before acting, identified by predictable URIs. Use tools for actions that change state or query with parameters.