native-mcp

Connect Hermes Agent to MCP servers over stdio and HTTP transports.

Updated Mar 13, 2026
One-click install
npx skills add https://github.com/hochoa13/Asistente-Harlest --skill native-mcp-hochoa13
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: native-mcp
Source: https://github.com/hochoa13/Asistente-Harlest/tree/main/skills/mcp/native-mcp
Command: npx skills add https://github.com/hochoa13/Asistente-Harlest --skill native-mcp-hochoa13

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill seamlessly integrates external tools and services exposed via the Model Context Protocol (MCP) directly into the Hermes Agent, eliminating the need for separate bridge applications and making advanced functionalities readily available.

Core Features & Use Cases

  • Unified Tool Access: Discover and register tools from MCP servers as native Hermes Agent tools.
  • Flexible Transport: Supports both stdio (command-line execution) and HTTP/StreamableHTTP transports for connecting to MCP servers.
  • Automatic Reconnection & Security: Ensures persistent connections with automatic retries and includes security filtering for environment variables and credential redaction in error messages.
  • Use Case: Integrate a remote database querying tool or a specialized API service exposed via MCP, allowing the agent to directly execute database queries or API calls as if they were built-in commands.

Quick Start

Configure your MCP servers in ~/.hermes/config.yaml under the mcp_servers key and restart Hermes Agent.

Frequently Asked Questions about native-mcp

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

FAQPage Schema
How do I connect external MCP servers to an AI agent?

You can connect external MCP servers by configuring them in your agent's config.yaml file under the mcp_servers key. This enables automatic discovery and registration of external tools as native agent commands without requiring separate bridge applications.

What is Model Context Protocol integration for service discovery?

Model Context Protocol integration is the process of discovering and registering tools from external MCP servers directly into an AI agent. It eliminates the need for separate bridge applications by making external functionalities readily available as native commands.

Can I use stdio and HTTP transports with MCP servers?

Yes, MCP server connections support both stdio for command-line execution and HTTP/StreamableHTTP transports. This flexible transport allows you to integrate various external tools and services exposed via the protocol.

Does MCP integration support automatic reconnection and credential security?

Yes, MCP integration ensures persistent connections with automatic retries and reconnection. It includes security filtering for environment variables and automatically redacts credentials in error messages to maintain secure operations.

Do I need a bridge application to use MCP tools with my agent?

No, you do not need a separate bridge application. The integration uses zero-config tool injection to seamlessly expose external tools and services from MCP servers directly as built-in agent commands.

What are the limitations of connecting to MCP servers via stdio?

While stdio transport supports command-line execution for MCP servers, it is limited to local or shell-based interactions compared to HTTP. However, both transports include automatic reconnection and credential redaction for secure operations.