agentsphere-mcp-server 1.0.0

mcp server for agentsphere sandbox

agentsphere-mcp-server

Introduction

agent-sphere is a cloud-based secure isolated sandbox infrastructure that provides AI with a stable, fast, and secure runtime environment.

agentsphere-mcp-server is an MCP Server designed for AI to connect and operate agent-sphere sandboxes.

Official Website: https://www.agentsphere.run/home

MCP Tools Description

This MCP Server provides the following 3 tools for AI usage:

1. exec_command

Function: Execute Linux system commands in the sandbox

Parameters:

• cmd (string): The command to execute

Return Value: Execution result containing stdout, stderr, and success fields

2. get_preview_link

Function: Get the access URL for web services in the sandbox

Parameters:

• port (int): Port number

Return Value: Result containing the accessible URL

3. upload_files_to_sandbox

Function: Upload local files or folders to a specified directory in the sandbox

Parameters:

• local_path (string): Absolute path of local file or folder
• target_path (string, optional): Target directory path in the sandbox, defaults to /user_uploaded_files/

Return Value: Result containing the list of successfully uploaded files or error information

Usage

MCP Server Configuration

To configure this server in Efflux, Cursor, or other MCP clients, add the following configuration to your MCP configuration file:

{
  "mcpServers": {
    "agentsphere": {
      "command": "uvx",
      "args": ["agentsphere-mcp-server"],
      "env": {
        "AGENTSPHERE_API_KEY": "your_api_key_here"
      }
    }
  }
}

Note:

• Replace your_api_key_here with your AgentSphere API key
• Please ensure uv is installed (uv is a modern Python dependency and project manager: https://docs.astral.sh/uv/getting-started/installation/)
• Please ensure network connectivity. Users in mainland China are recommended to enable global proxy to ensure proper functionality