native-mcp

Connect Spark Agent to external MCP servers and register their tools.

1|Updated Apr 24, 2026
One-click install
npx skills add https://github.com/automatedigital/spark --skill native-mcp-automatedigital
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: native-mcp
Source: https://github.com/automatedigital/spark/tree/main/skills/mcp/native-mcp
Command: npx skills add https://github.com/automatedigital/spark --skill native-mcp-automatedigital

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill eliminates the need for separate bridge CLIs or manual tool configuration to access external MCP server tools directly within Spark Agent, streamlining workflows that rely on third-party integrations.

Core Features & Use Cases

  • Automatic Tool Registration: Discovers MCP server tools on startup and registers them as native Spark tools with a consistent mcp_{server}_{tool} naming convention.
  • Multi-Transport Support: Works with both local stdio-based MCP servers (run via npx, uvx, or custom commands) and remote HTTP/StreamableHTTP MCP servers.
  • Built-in Security: Filters environment variables passed to stdio subprocesses and redacts credentials from error messages to prevent accidental leakage.
  • Use Case: Connect a GitHub MCP server to access issue tracking and pull request tools directly in your Spark coding workflows, or link a filesystem MCP server to manage project files without switching contexts.

Quick Start

Add your desired MCP server configuration to the mcp_servers section of your ~/.spark/config.yaml file, then restart Spark Agent to auto-register all available MCP tools for use in conversations.

Frequently Asked Questions about native-mcp

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

FAQPage Schema
How do I integrate external MCP tools into my Spark agent workflows?

You can integrate external MCP tools into your Spark agent workflows by connecting Spark Agent to Model Context Protocol servers, which automatically exposes their functions as native callable tools using a consistent naming convention.

Can I connect to both local and remote MCP servers from Spark Agent?

Yes, Spark Agent supports connecting to both local stdio-based MCP servers launched via npx or uvx and remote HTTP/StreamableHTTP MCP servers, enabling flexible tool integration across different environments.

How does Spark Agent handle credentials when connecting to MCP servers?

Spark Agent handles credentials securely by filtering environment variables passed to stdio subprocesses and redacting sensitive information from error messages to prevent accidental leakage during MCP server connections.

Do I need a bridge CLI to access filesystem or GitHub MCP tools in Spark?

No, you do not need a bridge CLI to access filesystem or GitHub MCP tools in Spark. The native MCP integration eliminates separate bridge CLIs by auto-discovering and registering tools directly within Spark Agent.

What happens if my connection to a remote HTTP MCP server drops?

If your connection to a remote HTTP MCP server drops, Spark Agent implements persistent connection management with exponential backoff reconnection to automatically restore tool availability across terminal and editor integrations.

How to configure MCP server registration for Spark Agent?

To configure MCP server registration, add your desired server configuration to the mcp_servers section of your ~/.spark/config.yaml file, then restart Spark Agent to auto-register all available MCP tools for use in conversations.