tools

The Tools MCP server bundles a diverse set of utilities across languages (C++, C, .NET, TypeScript, and Python) along with AI-related components for tasks like symbolically executing code, generating samples, or experimenting with AI-powered workflows. This server exposes a consolidated surface for running and testing small tooling suites and samples included in the repository. You can leverage the available tools for quick prototyping, code generation, algorithmic experiments, or data processing tasks, using the same MCP interface you rely on for other servers. To use it, start the server via your MCP orchestration (e.g., through the mcp_config entry above) and query the available tools through the MCP client, selecting the desired tool or sample from the provided catalog. Tools are designed to be language-agnostic entry points, so you can invoke a Python script, a TypeScript utility, or a C/C++ sample as needed, with outputs returned through the MCP channel.

Prerequisites:

• Docker installed and running on the host
• Access to pull the appropriate drazenzadravec/tools image (may require authentication)

Installation steps:

1. Ensure Docker is up:
   • docker version
   • system supports running Linux containers
2. Pull and run the Tools image (as defined in the MCP config):
   • docker pull drazenzadravec/tools
   • docker run -d --name mcp-tools -i drazenzadravec/tools
3. Verify the server is responding via the MCP client:
   • Use your MCP client to request the server 'tools' and list available tools
4. If you manage multiple MCP servers, configure your orchestrator to reference the 'tools' server with the provided mcp_config entry.

Optional: Customize resource limits, environment variables, or mounted volumes as needed for the tooling environment (see additional_notes for guidance).

Environment variables and configuration options (example):

• ENV_MCP_LOG_LEVEL: Set log verbosity (debug, info, warn, error)
• TOOLSET_TAG: Choose a specific image tag or toolset variant if the image supports it
• VOLUMES: Mount a workspace into the container for input/output files
• MEM_LIMIT / CPU_LIMIT: Restrict resource usage for stability in shared environments Common issues:
• Image not found: ensure the image name 'drazenzadravec/tools' is correct and accessible
• Container starts but tool commands fail: verify that the container exposes the expected MCP entry point and that network access to the MCP controller is allowed
• Permissions: if your tools write to mounted volumes, ensure the host path permissions align with the container user For updates, pull the latest image and restart the MCP server to pick up changes.