Leakguard Mcp

Overview

Leakguard Mcp is a static analyzer that flags lookahead bias and data leakage in machine learning and trading code before backtesting. Running as a Model Context Protocol server, it integrates directly into AI assistants as an invokable tool, catching common pitfalls that undermine model validity.

The server runs locally on your machine via the uvx command uvx leakguard-mcp, speaking MCP over stdio with zero authentication required. Built on the leakguard-mcp Python package, it launches out of the box at version v0.1.1 under the MIT license. Source code is available at doazvjettu/leakguard-mcp on GitHub.

Leakguard Mcp belongs in the Development category and works with MCP clients including Claude Desktop, Claude Code, Cursor, VS Code, Windsurf, Zed, and others. If you maintain this project, claim this listing to verify ownership and keep information up to date.

Installation

// claude_desktop_config.json
{
  "mcpServers": {
    "leakguard_mcp": {
      "command": "uvx",
      "args": [
        "leakguard-mcp"
      ]
    }
  }
}