Runno is a collection of JavaScript toolkits for securely running code in multiple programming languages in environments such as browsers and Node.js. It achieves sandboxed execution through WebAssembly and WASI, supports languages such as Python, Ruby, JavaScript, SQLite, C/C++, and provides integration methods such as web components and MCP servers.

A secure code sandbox environment based on Docker containers, providing an isolated execution space for AI applications.

A secure Docker container code sandbox environment that provides an isolated execution space for AI applications, supporting multiple programming environments and file operations.

This project is a collection of MCP servers designed for Kali Linux, aiming to enhance reverse engineering, security testing, and automation workflows. It integrates tools such as Nmap and Wireshark and provides Docker sandbox support.

An MCP server that provides an isolated Docker environment, supporting code execution, package installation, and environment configuration in multiple programming languages

A Docker-based secure code execution environment that provides isolated sandbox support for the MCP protocol

The CodeSandbox MCP server exposes CodeSandbox SDK operations through tools for AI agents to perform operations such as sandbox creation and file read and write.

The 302AI Sandbox MCP Server is a secure code sandbox service that allows AI assistants to execute arbitrary code, supporting multiple modes and dynamic tool loading.

Sandbox MCP is a model context protocol server that allows LLMs to run code in secure Docker containers, solving the security problem of the inability to test code generated by AI.

This project is an MCP server based on the Tradier Sandbox API, providing options chain data and historical price query functions for Claude Desktop.

An MCP server that provides sandboxed file access, supporting directory exploration, file reading, content search, and secure editing, suitable for AI agents to manage text file collections.

mcp - scaffold is a development sandbox for verifying the Model Context Protocol (MCP) server, providing support for local LLMs (such as LLaMA 7B) and cloud inference, including a chat interface and reference architecture.

gbox is an open - source self - hosted sandbox project that provides a secure environment for MCP integration and other AI agent use cases. It includes tools such as a terminal, browser, and file editor, supports local task execution and file processing, and provides Python and Typescript SDKs.

The TCL MCP Server is a server that supports AI agents to execute TCL scripts and manage the MCP tool ecosystem. It provides a secure sandbox and a developer-friendly experience, and supports multiple runtime environments.

The MCP Data Fetch Server is a secure and sandboxed server that fetches web content and extracts data through the Model Context Protocol (MCP) without executing JavaScript.

VkRunner is a Vulkan shader testing tool based on Piglit's shader_runner, aiming to support test scripts as similar as possible to Piglit's shader_test format. It supports writing shaders in GLSL or SPIR-V and provides a rich set of test commands to verify shader behavior, including drawing, computing, and probe detection functions.

The Modal MCP Toolbox is a collection of tools running on the Modal platform. Based on the Model Context Protocol (MCP), it can provide extended functions for LLMs, such as Python sandbox code execution and FLUX model image generation.

The Local MCP Tools Collection aims to simplify the integration of multi - functional tools for local large language models, supporting multiple MCP servers through a single virtual environment, including web search, Python sandbox, and long - term memory functions.

A secure JavaScript execution service based on the MCP protocol

ipybox is a lightweight Python code execution sandbox based on IPython and Docker, supporting secure code execution in local or remote environments. It is suitable for data analysis and code execution scenarios of AI agents.

The Codex MCP Tool is an open-source MCP server that connects the IDE or AI assistant to the Codex CLI, supporting non-interactive automation, secure sandbox editing, and large-scale code analysis. It provides progress streaming updates and a structured change mode.

An experimental sandbox environment to explore the interaction between MCP hosts, clients, and servers, including attack tests on the MCP server and research on LLM abuse.

Deskmate is a local execution agent that allows users to control personal computers through natural language. It supports multiple AI agent backends and messaging platforms, provides full access to local tools, and does not require sandbox restrictions.

SpotDB is a lightweight data sandbox that provides a securely isolated temporary database environment for AI workflows and data exploration, supporting snapshot restoration and multiple API access methods.

An MCP server based on the Deno sandbox, allowing for the secure execution of TypeScript, JavaScript, and Python code locally, providing a secure execution environment with permission control.

Python MCP Sandbox is an interactive code execution tool that can securely run Python code and manage packages in isolated Docker containers

An MCP server that provides an isolated Docker environment for executing code

The Python MCP Sandbox is an interactive Python code execution tool that allows users to securely execute Python code and install packages in isolated Docker containers.

The Threat.Zone MCP server is a malware analysis service interface built on FastMCP, providing LLMs with a standardized threat analysis toolset that supports file/URL detection, sandbox analysis, network behavior monitoring, and other functions.

This is a prototype of a code execution server based on the MCP protocol, which combines the code generation capabilities of large language models with MCP tool integration, enabling AI agents to run Python code in an isolated sandbox and invoke external tools.

PRIMS is a lightweight open - source MCP server designed for LLM agents, providing a secure one - time Python code sandbox execution environment that supports dependency installation and file mounting.

The Node.js Sandbox MCP Server isolates and runs arbitrary JavaScript code through Docker containers, supports dynamic installation of npm dependencies, and provides two modes: one-time execution and persistent containers.

The Sandbox Fusion MCP Server is an implementation that provides code interpreter functionality for large language models. It supports code execution in multiple programming languages and Jupyter notebook operations, and communicates with clients through stdio interaction.

The Qwen MCP tool is a server based on the model context protocol that integrates Qwen CLI with AI assistants, providing functions such as large context window analysis, file processing, sandbox execution, and multi - model support.

An MCP server for securely running Python code in a sandbox environment, supporting automatic installation of dependency packages and asynchronous execution.

Provide a virtual machine environment based on Magalu Cloud, enabling code execution and virtual resource interaction through the Model Context Protocol (MCP). Support the secure integration of LLMs with sandboxed VMs to enhance agent capabilities.

This project implements a full-stack AI agent system inspired by the model context protocol, including a React front-end and a FastAPI back-end. The back-end serves as an AI coordinator, using Google Gemini to parse natural language instructions, safely call the file system, browser, or GitHub tools to perform tasks, and return the structured results to the front-end. The system focuses on software development scenarios, providing secure sandbox file operations, real-time web browsing, and repository inspection functions, and visualizing the tool call process.

This project is a TypeScript-based experimental platform for MCP servers, used for learning and testing integration with Claude Desktop and Cursor IDE. It extends the basic tutorial functions, providing a modular design and an extensible sandbox environment.

An AWS resource query MCP server based on Python and boto3, deployed through Docker containerization, providing a secure code sandbox execution environment and supporting operations on multiple AWS services

Aggregate all MCP services into a sandbox environment, provide a unified proxy service, and users can access all functions through a single entry point.