Aibolit MCP Server

What the cqfn/aibolit-mcp-server MCP server does

The cqfn/aibolit-mcp-server MCP server gives an AI coding agent a focused recommendation for improving Java code. Rather than returning a broad list of small style concerns, it asks Aibolit to help select the most serious design problem in one Java file. The intended result is a clear refactoring target that can improve maintainability, readability, loose coupling, or high cohesion.

This is a good fit when an agent has been asked to make code better but needs help deciding where to start. It can direct the agent toward a structural problem that deserves attention before cosmetic cleanup.

How it works

The server exposes one MCP tool: find_the_most_critical_design_issue. The tool analyzes a single Java file, chooses the design flaw considered most critical, and identifies where that issue appears in the file. Its guidance is specifically intended to support immediate refactoring rather than minor or cosmetic changes.

The cqfn/aibolit-mcp-server MCP server uses Aibolit as the Java code analyzer. Aibolit must be installed separately and available in the environment, as shown by the README's aibolit --version check. The repository documentation describes using the server from AI coding environments, including Claude Code, Cursor, and Windsurf.

Setup and configuration

The documented npm command adds version 0.0.6 to Claude Code:

claude mcp add aibolit npx aibolit-mcp-server@0.0.6

For direct non-interactive execution, the equivalent package runner invocation is npx -y aibolit-mcp-server@0.0.6. Before using the server, install Node, npm, Python, Pip, and Aibolit. Confirm the analyzer installation with:

aibolit --version

The README also mentions Node 18+, npm, and GNU make for contributing and running the project's checks. Those development requirements are separate from the runtime analyzer dependency.

Tools and capabilities

The cqfn/aibolit-mcp-server MCP server provides:

• find_the_most_critical_design_issue: examines one Java file, selects the most serious design flaw, and points to its location.
• Refactoring prioritization based on maintainability, readability, loose coupling, and high cohesion.
• Guidance designed to help an AI agent fix a significant design problem instead of focusing only on low-impact cleanup.

Limitations and notes

Analysis is described for one Java file at a time, so the available material does not establish project-wide analysis or multi-file issue tracking. Only one MCP tool is documented. The server identifies a problem and its location; the material does not promise that it applies the refactoring automatically.

The cqfn/aibolit-mcp-server MCP server requires the Aibolit analyzer in addition to its Node-based package. No environment variables, API credentials, hosted endpoint, or paid service are documented. The repository is marked as MIT licensed.