Skip to main content
Glama

ToolBox MCP Server

GitHub Release License

An AI-powered automation tool development platform, providing:

🧩 Modular Architecture - Tool hot-reloading via the src/tools directory 🤖 AI Assistance - AI engine for natural language to tool template conversion 🚀 Enterprise-Grade Capabilities - Integration of production environment services such as MongoDB/Redis/SSH 🔄 Real-time Updates - Zero-downtime deployment via buildReload_tool

graph LR
    A[Developer] -->|Create| B(Tool Template)
    B --> C{AI Verification}
    C -->|Pass| D[Automatic Loading]
    C -->|Fail| E[Human Review]
    D --> F[API Exposure]
    F --> G[Client Invocation]
    style C fill:#4CAF50,stroke:#333

中文文档

Contributing

Contributing Guidelines (English) Contributing Guidelines (中文)

Tool Specifications

Related MCP server: Jachy MCP Server

Features

Tools

View the complete tool specifications and detailed documentation: TOOL.md

Resources

Resources are generated dynamically as a result of tool execution. For example, the create_note tool creates a note resource that can be accessed via its URI.

Workflow Tool

The workflow_tool is a powerful tool for orchestrating complex workflows by chaining together multiple tools. It supports:

  • Serial and Parallel Execution: Execute tools in a sequential or parallel manner.

  • Transaction Management: Manage transactions with compensation mechanisms for error handling.

  • Detailed Reporting: Generate detailed reports on workflow execution, including the status, execution time, and results of each step.

  • Customizable Output: Specify an output file to save the workflow execution report.

Prompts

  • summarize_notes: Generates summaries of the notes created using the create_note tool.

Development Guide

Adding New Tools

graph TD
    A[Create Tool Template] --> B{AI Automatic Verification}
    B -->|Schema Validation| C[Generate Test Cases]
    B -->|Risk Exists| D[Human Review]
    C --> E[Unit Testing]
    D -->|Approve| E
    E --> F[Security Scanning]
    F --> G[Build Integration]
    G --> H[Version Release]
    style A fill:#f9f,stroke:#333
    style H fill:#4CAF50,stroke:#333

Detailed Development Process

  1. Template Creation: Create a new tool file in the src/tools/ directory

  2. AI Verification: Automatically check the parameter Schema compliance

  3. Test Generation: Generate test cases based on the function description

  4. Security Review: Static code analysis and dependency checking

  5. Continuous Integration: Automated deployment via GitHub Actions

View the complete development guide: prompt.md Refer to existing implementations: Tool Examples

Development

Install dependencies:

npm install

Build the server:

npm run build

For development with auto-rebuild:

npm run watch

Marketplace Submission

ToolBox Logo A Seamless Integrated Automation Toolkit for Claude Desktop

Installation

To integrate with the Claude Desktop application, add the following server configuration to:

  • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

  • Windows: %APPDATA%/Claude/claude_desktop_config.json

// Cline MCP server configuration file
{
  "command": "node",
  "args": [
    "--inspect=9229",
    "/MCP/ToolBox/build/index.js"
  ],
  "env": {
    "MONGO_URI": "mongodb://user:password@host:port/db",
    "MONGO_INDEX_OPS": "true",
    "REDIS_URI": "redis://:password@host:port",
    "SSH_server1_URI": "username:password@host:port",
    "GEMINI_API_KEY":"GEMINI_API_KEY",
    "SSEPORT": "8080" // Optional: If set, the server will use SSE transport on this port. Otherwise, it defaults to Stdio transport.
  },
  "disabled": false,
  "autoApprove": []
}

Core Values

🚀 Enterprise-Grade Automation Leveraging package.json configuration, providing:

  • Global CLI tool installation (tbx command)

  • Workflow scheduling engine

  • Multi-platform support (Windows/macOS)

  • Hybrid cloud deployment capabilities (MongoDB/Redis/SSH)

🔧 Developer-Friendly

  • Strongly-typed TypeScript implementation

  • Real-time debugging support (--inspect flag)

  • VSCode debugging configuration template

  • Hot-reloading mechanism:

    graph LR
      A[Code Modification] --> B[buildReload_tool]
      B --> C[Automatic Compilation]
      C --> D[Security Verification]
      D --> E[Tool Reloading]
      style B fill:#4CAF50,stroke:#333

    Implemented by calling buildReload_tool:

    • Zero-downtime updates

    • Automatic dependency tree parsing

    • Version compatibility check

    • Sandbox environment testing

Debugging

Debugging MCP servers can be challenging due to their stdio communication. Here are a few approaches:

🚧 Disclaimers

Sensitive Data

DO NOT CONFIGURE CONTAINERS WITH SENSITIVE DATA. This includes API keys, database passwords, etc.

Any sensitive data exchanged with the LLM is inherently compromised, unless the LLM is running on your local machine.

  • ⚠️ The tool is provided "as-is" under MIT License without warranties

  • ⚠️ Developer not liable for direct/indirect damages

  • ⚠️ Users bear all risks from improper container configurations

  • ⚠️ Any illegal or destructive usage is strictly prohibited

  • ⚠️ See LICENSE for full terms

  1. Node.js Inspector: Launch the server with the --inspect=9229 flag:

node --inspect=9229 build/index.js

Then, connect to the server using Chrome DevTools by navigating to chrome://inspect.

  1. MCP Inspector: Utilize the MCP Inspector, a dedicated debugging tool accessible via the inspector npm script:

  2. VSCode Debugging

To debug with VSCode, create a .vscode/launch.json file with the following configuration:

{
    "version": "0.2.0",
    "configurations": [
        {
            "type": "node",
            "request": "attach",
            "name": "ToolBox",
            "address": "localhost",
            "port": 9229,
            "localRoot": "${workspaceFolder}"
        }
    ]
}

Then, launch the server with the --inspect=9229 flag and attach the VSCode debugger.

Available Tools

17 tools
buildReload_toolD

Inputs: no named parameters.

calculator_toolD

Inputs: calculation_type (required), expression, statistics_operation, data_points, +35 more.

cli_toolD

Inputs: command, commands, mode, timeout, +3 more.

compress_toolD

Inputs: action (required), sourcePath (required), destinationPath (required), format (required).

create_noteD

Inputs: title (required), content (required).

excel_toolD

Inputs: action (required), filePath (required), format (required), outputFilePath, +2 more.

fileSystem_toolD

Inputs: operation (required), sourcePath (required), targetPath, content, +6 more.

gemini_image_toolD

Inputs: operation (required), prompt (required), outputDir (required), inputImage, +5 more.

image_toolD

Inputs: sourcePath (required), outputPath, quality, compressionLevel, +4 more.

log_toolD

Inputs: pageSize, page, toolName, status, +4 more.

mongo_toolD

Inputs: dbName (required), collectionName, queryType, operationType, +9 more.

redis_toolD

Inputs: command (required), args.

schedule_toolD

Inputs: action (required), time, delaySeconds, interval, +3 more.

sftp_toolD

Inputs: serverName (required), action (required), localPath (required), remotePath (required).

ssh_toolD

Inputs: serverName (required), command (required).

time_toolD

Inputs: action (required), time_str, timezone, timestamp, +1 more.

workflow_toolD

Inputs: steps (required), version, parallel, outputFile.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    D1.4/5.0

    Scored across 17 tools

    Disambiguation2/5

    The tools cover diverse domains (e.g., file operations, databases, utilities), but there is significant ambiguity due to vague naming and lack of descriptions. For example, 'image_tool' and 'gemini_image_tool' likely overlap in image-related tasks, while 'cli_tool', 'ssh_tool', and 'sftp_tool' could all involve command-line or remote operations without clear boundaries. The absence of descriptions exacerbates confusion, making it difficult for an agent to reliably choose the right tool for a specific need.

    Naming Consistency2/5

    Naming conventions are inconsistent and chaotic. There is a mix of styles: some tools use snake_case (e.g., 'calculator_tool', 'create_note'), others use camelCase (e.g., 'fileSystem_tool'), and some have hybrid or irregular forms (e.g., 'buildReload_tool'). Verb usage varies widely, from generic terms like 'tool' to specific verbs like 'create', with no predictable pattern. This inconsistency makes the tool set harder to navigate and understand at a glance.

    Tool Count3/5

    With 17 tools, the count is borderline high but not extreme, falling into the 16-25 range that feels heavy for a server named 'ToolBox MCP Server', which suggests a broad utility scope. While the number is reasonable for covering multiple domains, it risks being overwhelming without clear organization or descriptions. A more focused set might improve usability, but it doesn't reach the extreme mismatch of 50+ tools.

    Completeness2/5

    The tool set appears to cover various utility domains (e.g., file systems, databases, scheduling, images), but without descriptions, it's impossible to assess true coverage or identify obvious gaps. The lack of CRUD operations for many domains (e.g., no update/delete for notes or files) and the vague tool names suggest significant gaps in functionality. For instance, 'create_note' might exist, but without tools for reading or managing notes, workflows could hit dead ends, leading to potential agent failures.

    Maintenance

    ActivityInactive
    ResponsivenessNo recent community issues

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      A
      quality
      F
      maintenance
      Enables AI models to dynamically create and execute their own custom tools through a meta-function architecture, supporting JavaScript, Python, and Shell runtimes with sandboxed security and human approval flows.
      5
      10
      MIT
    • A
      license
      A
      quality
      D
      maintenance
      A centralized automation hub that provides a unified interface for AI agents to interact with various services, currently featuring Discord forum post creation. It offers an extensible TypeScript architecture designed for easily adding new tool domains and integrating with platforms like Claude Desktop and Cursor.
      1
      ISC