tools

This MCP server repository provides a modular toolkit of over a dozen tools designed to operate within the Model Context Protocol (MCP). Each core tool lives under a packages/tools-<name> and has a corresponding MCP wrapper under packages/tools-<name>-mcp that exposes the tool via the MCP interface. Tools cover areas like filesystem operations, network requests, data processing (JSON, XML, Base64), PDF text extraction, and RAG utilities, all designed to be composed into larger MCP servers for AI agents. To use, install dependencies, build the monorepo, and run one or more MCP wrappers to expose the tools to MCP clients. Clients can then call the MCP endpoints to execute specific tools, pass structured inputs, and receive typed outputs.

Prerequisites:

• Node.js (LTS) and pnpm installed
• Git to clone the repository

Install and build:

1. Clone the repository git clone https://github.com/SylphxAI/tools.git cd tools

2. Install dependencies using pnpm (workspaces enabled in this repo) pnpm install

3. Build all packages (core utilities, MCP adaptors, and individual tool packages) pnpm build

4. Start an MCP server (example for the Tools MCP set)

   • Each MCP wrapper package can typically be started via its own start script or by invoking the built entry under dist
   • Example (adjust path per actual startup script in your environment): node packages/tools-filesystem-mcp/dist/index.js

Notes:

• If you use the Vercel or Docker deployment options described in the repo, follow those deployment steps after building.
• This repository is marked as deprecated in the README, so production use may require caution and validation.

This MCP toolkit repository is deprecated and archived per the top-level README. Expect discontinued maintenance and potential breaking changes if integrating with newer MCP clients. Common issues to anticipate include mismatched MCP protocol versions, missing environment configuration, and build-time dependencies not aligned with your runtime. Environment variables like MCP_ENV and LOG_LEVEL can help control runtime behavior and logging. Each tool follows a consistent core/adaptor pattern, but individual MCP servers may require explicit build steps before they can be started. If you rely on this for production, consider forking or migrating to an actively maintained alternative.