native-mcp

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

Updated May 3, 2026
One-click install
npx skills add https://github.com/JuanMS20/solviora-agent --skill native-mcp-juanms20
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: native-mcp
Source: https://github.com/JuanMS20/solviora-agent/tree/main/skills/mcp/native-mcp
Command: npx skills add https://github.com/JuanMS20/solviora-agent --skill native-mcp-juanms20

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Hermes Agent needs to connect to MCP servers and use their tools without manual wiring across every session.

Core Features & Use Cases

  • Auto-discovery and registration of MCP tools from stdio or HTTP servers.
  • Transparent tool injection into all Hermes toolsets for seamless cross-conversation use.
  • Simple, scalable configuration to connect new MCP servers and reuse existing pipelines.

Quick Start

Configure an MCP server in ~/.hermes/config.yaml under mcp_servers and restart Hermes to auto-discover its tools.

Frequently Asked Questions about native-mcp

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

FAQPage Schema
How do I connect an MCP server to an agent and expose its tools?

Hermes auto-discovers tools from MCP servers by reading the ~/.hermes/config.yaml mcp_servers section, applying transparent tool injection across all conversations without requiring manual wiring for each session.

Does MCP server discovery support both stdio and HTTP transports?

Yes, MCP server discovery supports both stdio and HTTP transports, enabling automated tool registration from local Node.js processes or remote HTTP endpoints into the Hermes agent environment.

What dependencies are required to run stdio and Python-based MCP servers?

Running stdio MCP servers requires Node.js, while Python-based servers require uv and the optional MCP Python package. These dependencies must be installed before starting server discovery.

How does automated tool injection work across multiple agent conversations?

Automated tool injection works by registering tools discovered from configured MCP servers directly into the Hermes agent toolsets, making them transparently available for use across all current and future conversations.

What are the limitations when configuring MCP servers in the config.yaml file?

The main limitation is that configuration requires a valid ~/.hermes/config.yaml mcp_servers section, and you must restart Hermes after adding new MCP servers to trigger the auto-discovery and tool registration process.