logo
Free, unlimited AI code reviews that run on commit
git-lrc git-lrc GitHub Install Now We'd appreciate a star git-lrc - Free, unlimited AI code reviews that run on commit | Product Hunt git-lrc - Free, unlimited AI code reviews that run on commit | Product Hunt

SystemAgent-LocalOS

A localized utility package facilitating direct interaction with the host operating system's interfaces, designed to serve as an office automation intermediary.

Author

SystemAgent-LocalOS logo

gojue

Apache License 2.0

Quick Info

GitHub GitHub Stars 310
NPM Weekly Downloads 0
Tools 1
Last Updated 2026-02-19

Tags

molinggojuebrowsergojue molingmoling molingbrowser automation

SystemAgent-LocalOS MCP Hub

English | 中文 | 日本語

GitHub stars GitHub forks CI Github Version


Overview

SystemAgent-LocalOS functions as an MCP Server focused on local computing and web browsing activities. It achieves system-level manipulation by leveraging native operating system Application Programming Interfaces (APIs). This capability permits sophisticated file system operations, including but not limited to retrieval, modification, combination, metric calculation, and data consolidation, alongside the execution of arbitrary system instructions. It stands as a self-contained, dependency-free solution for automating localized office routines.

Key Benefits

[!IMPORTANT] This utility requires zero external prerequisites or environment setups (like Node.js, Python, or Docker), allowing immediate deployment across diverse OS environments including Windows, Linux, and macOS. This inherent portability avoids common conflicts arising from varied development stacks.

Capabilities

[!CAUTION] Exercise extreme prudence when invoking arbitrary command-line operations due to potential security implications.

  • File System Manipulation: Capabilities cover reading, persistence, merging datasets, statistical analysis, and aggregation of file contents.
  • Shell Execution: Direct invocation of system shell commands is supported.
  • Web Interface Control: Browser automation is powered by the github.com/chromedp/chromedp package, necessitating a functional installation of Google Chrome.
    • For Windows installations, the complete path to the Chrome executable must be established within the system's environment variables.
  • Forward Development Trajectory:
    • Enhanced personal computer data organization strategies.
    • Assistance functionalities for document composition.
    • Integrated scheduling and timeline management.
    • General quality-of-life personal assistance features.

[!WARNING] Current validation has been primarily conducted on macOS; operational stability on other operating systems remains unverified.

Compatible MCP Consumers

Demonstrations

https://github.com/user-attachments/assets/229c4dd5-23b4-4b53-9e25-3eba8734b5b7

SystemAgent-LocalOS operating within the Claude environment.

Configuration Schema

Server Configuration (SystemAgent-LocalOS)

The primary configuration artifact is established at /Users/username/.moling/config/config.json. This file is modifiable as needed. If absent, initialization can be triggered via the command: moling config --init.

Client Integration Setup

To interface with a client like Claude, append the following block to its configuration:

[!TIP]

Typically, only a minimal configuration of 3 to 6 lines is required for connection.

Claude configuration location example: ~/Library/Application\ Support/Claude/claude_desktop_config

{
  "mcpServers": {
    "SystemAgent-LocalOS": {
      "command": "/usr/local/bin/moling",
      "args": []
    }
  }
}

Note that /usr/local/bin/moling specifies the location of the downloaded SystemAgent-LocalOS binary executable.

Automated Client Configuration

Execute moling client --install to automatically inject the necessary settings into supported MCP clients. The utility will autonomously detect compatible clients (e.g., Cline, Claude, RooCode) and apply the configuration.

Operational Modes

  • Standard I/O (Stdio) Mode: An interactive command-line interface designed for straightforward user engagement.
  • Server-Sent Events (SSE) Mode: Optimized for automated, non-graphical execution environments.

Acquisition Methods

Method 1: Installation via Script

Linux/MacOS
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/gojue/moling/HEAD/install/install.sh)"
Windows

[!WARNING] Verification for Windows execution is pending.

powershell -ExecutionPolicy ByPass -c "irm https://raw.githubusercontent.com/gojue/moling/HEAD/install/install.ps1 | iex"

Method 2: Direct Binary Retrieval

  1. Obtain the compressed distribution file from the official releases page.
  2. Unpack the downloaded archive.
  3. Initiate the server process: sh ./moling

Method 3: Compilation from Source

  1. Clone the repository location:
git clone https://github.com/gojue/moling.git
cd moling
  1. Compile the application (requires a functional Golang environment):
make build
  1. Execute the newly built artifact:
./bin/moling

Utilization Guide

Once the server component is active, connect to it using any MCP-enabled client by adjusting that client's configuration to reference the SystemAgent-LocalOS endpoint.

Licensing

Distributed under the Apache License, Version 2.0. See the LICENSE file for comprehensive details.

WIKIPEDIA: A headless browser represents a web browser devoid of a graphical presentation layer. These specialized agents permit the programmatic management of web page content in an environment mirroring standard browsers, yet they are controlled via command lines or network protocols. Their utility is pronounced in web application quality assurance, as they possess the full rendering and interpretation capabilities of a regular browser, including stylesheet rendering (layout, color, typography) and script execution (JavaScript/Ajax)—functions often inaccessible via simpler testing methodologies. Modern browser versions (Chrome 59+, Firefox 56+) natively support remote browser control, superseding earlier solutions like PhantomJS.

== Primary Applications == The fundamental uses for headless browsing capabilities encompass:

  • Automated testing protocols for contemporary web architectures (web validation).
  • Generating high-fidelity static images (screenshots) of web interfaces.
  • Executing unit or integration tests for JavaScript libraries.
  • Programmatic manipulation of user interactions on web documents.

=== Secondary Uses === Headless agents are also instrumental in sophisticated web data harvesting (scraping). Google, for instance, acknowledged in 2009 that using one could aid in indexing content from sites heavily reliant on Ajax. Conversely, these tools have been associated with undesirable activities:

  • Coordinating distributed denial-of-service assaults against web resources.
  • Inflating advertisement view counts.
  • Executing automated manipulation contrary to site terms, such as automated credential testing.

Despite these risks, a 2018 traffic analysis indicated no statistical bias among malicious actors favoring headless browsers over traditional ones for activities like DDoS, SQL injection, or cross-site scripting.

== Control Frameworks == As major browser engines now incorporate native headless capabilities via APIs, several software frameworks centralize this control under unified interfaces. Examples include:

  • Selenium WebDriver – Adheres to W3C WebDriver specifications.
  • Playwright – A library primarily for automating Chromium, Firefox, and WebKit via Node.js.
  • Puppeteer – A Node.js utility specifically for controlling Chrome or Firefox instances.

=== Automated Validation Environments === Certain software platforms designed for quality assurance integrate headless browsing into their operational apparatus.

  • Capybara utilizes headless browsing (via WebKit or Headless Chrome) to simulate user actions within its testing structures.
  • Jasmine typically relies on Selenium but can be configured to employ WebKit or Headless Chrome for browser-based verification.
  • Cypress, a specialized frontend testing framework.
  • QF-Test, a tool for graphical user interface program validation, often employing a headless browser backend.

=== Alternative Paradigms === An alternative strategy involves employing software that exposes browser-like APIs directly. Deno incorporates such browser APIs intrinsically. For the Node.js ecosystem, jsdom offers the most comprehensive implementation of these features. While many alternatives successfully emulate fundamental browser functions (HTML parsing, cookie management, XHR requests, basic JavaScript execution), they frequently lack full DOM rendering or robust event handling, generally resulting in superior execution speed compared to full headless instances.

See Also

`