native-mcp

Connect MCP servers via stdio or HTTP and register their tools automatically.

Updated May 13, 2026
One-click install
npx skills add https://github.com/superfhp/lumi-agent-body --skill native-mcp-superfhp
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: native-mcp
Source: https://github.com/superfhp/lumi-agent-body/tree/main/skills/mcp/native-mcp
Command: npx skills add https://github.com/superfhp/lumi-agent-body --skill native-mcp-superfhp

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires mcp.

What problem does it solve? Configuring and connecting external MCP servers to an AI agent normally requires custom bridge code and manual tool wiring. This Skill lets you declare MCP servers in a YAML config file and have their tools discovered, registered, and injected into every conversation automatically. ## Core Features & Use Cases - Automatic Tool Discovery: On startup, connects to each configured MCP server, lists its tools, and registers them with a mcp_{server}_{tool} naming prefix. - Dual Transport Support: Runs local stdio servers (npx, uvx, or any command) and connects to remote HTTP/StreamableHTTP servers with custom headers. - Security Controls: Filters environment variables passed to subprocesses and redacts credential patterns from error messages shown to the LLM. - Use Case: Add the GitHub MCP server with a personal access token in ~/.hermes/config.yaml, restart the agent, and immediately call tools like mcp_github_list_issues in any conversation. ## Quick Start Add an MCP server entry under the mcp_servers key in ~/.hermes/config.yaml and restart the agent to make its tools available.

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 my AI agent?▼

Add the server under the mcp_servers key in ~/.hermes/config.yaml with either a command (stdio) or url (HTTP), then restart the agent. It connects at startup, discovers tools, and registers them with the mcp_{server}_{tool} prefix.

How to configure stdio vs HTTP MCP server transport?▼

Stdio servers use command and args fields to launch a subprocess like npx or uvx. HTTP servers use a url field with optional headers for authentication. A server config must have one or the other, not both.

Does the MCP client pass my environment variables to servers?▼

No, only safe baseline variables like PATH, HOME, and XDG_* are inherited by stdio subprocesses. API keys and secrets must be explicitly declared in the server's env config block to be passed through.

Why are my MCP tools not appearing after configuration?▼

Check that servers are listed under mcp_servers (not mcp or servers), YAML indentation is correct, and the mcp Python package is installed. Tools appear with the mcp_{server}_{tool} naming pattern in startup logs.

What happens when an MCP server connection drops?▼

The client automatically reconnects with exponential backoff up to 5 retries, capped at 60 seconds between attempts. If the server remains unreachable after 5 attempts, it gives up until the next discovery call.