# mcp-server-analyzer [Health: Active]

**Category:** 💻 Developer Tools  
**Repository:** https://github.com/Anselmoo/mcp-server-analyzer  
**GitHub Stars:** 11  
**Views:** 1  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/mcp-server-analyzer

## Description
MCP server for Python, JavaScript, and TypeScript code analysis with Ruff, ty, Vulture, and Biome

## Tools
Capabilities this server exposes over MCP:

- **ruff-check** — Lint Python code using RUFF to identify style violations and potential errors.
- **ruff-format** — Format Python code using RUFF's fast formatter.
- **ruff-check-ci** — Run RUFF linter with CI/CD-specific output formats.
- **vulture-scan** — Detect dead/unused code using VULTURE.
- **analyze-code** — Comprehensive analysis combining RUFF linting and VULTURE dead code detection.

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `uvx` (confidence: high):

```json
"mcpServers": {
  "mcp-server-analyzer": {
    "command": "uvx",
    "args": ["mcp-server-analyzer"]
  }
}
```

## Documentation

## What mcp-server-analyzer MCP server does

The mcp-server-analyzer MCP server gives AI assistants and development environments access to static analysis for Python, JavaScript, and TypeScript code. Python checks use Ruff for linting and formatting, ty for type analysis, and Vulture for identifying likely unused code. JavaScript and TypeScript checks use Biome for linting and formatting.

The server also provides a combined analysis operation that brings together Ruff, ty, and Vulture results. Its quality metrics consider lint findings, type diagnostics, dead-code findings, and a combined score on a 0–100 scale. This makes it suitable for code review assistance, editor-based feedback, and workflows that need a consolidated assessment rather than a single analyzer’s output.

## Tools and capabilities

Available tools include:

- `ruff-check` for Python lint findings, including style and potential-error diagnostics.
- `ruff-format` for applying or reporting Python formatting through Ruff.
- `ruff-check-ci` for Ruff output intended for CI/CD workflows such as GitHub Actions or GitLab CI.
- `ty-check` for Python type checking and diagnostics such as incompatible return values.
- `vulture-scan` for detecting potentially unused imports, functions, and variables.
- `biome-check` for JavaScript and TypeScript linting.
- `biome-format` for JavaScript and TypeScript formatting.
- `analyze-code` for a combined Ruff, ty, and Vulture assessment.

These tools are useful when an agent needs to inspect supplied source code, identify issues, format code, or summarize several Python quality checks together. Biome analysis requires the Node.js dependency described in the project prerequisites.

## Setup and configuration

The mcp-server-analyzer MCP server can be launched with `uvx`, installed with pip, run from a cloned source checkout, or started from the published Docker image. The documented `uvx` command is `uvx mcp-server-analyzer`. The Docker configuration runs the image interactively with `docker run -i --rm ghcr.io/anselmoo/mcp-server-analyzer`.

The README provides stdio configurations for Claude Desktop, Zed, VS Code, and Claude Code. VS Code can use either user settings or a workspace `.vscode/mcp.json`; the latter uses an `mcp` wrapper key. Claude Code uses a project-level `.mcp.json` file. The server requires Python 3.13 or newer. Node.js 22 or newer is listed for Biome JavaScript and TypeScript analysis, while Docker is optional.

## Limitations and notes

Analysis input is written to a temporary file, processed, and deleted after the operation. It also states that analysis makes no outbound network calls and collects no telemetry, usage data, analytics, or crash reports.

The documented analyzer set is limited to Ruff, ty, Vulture, and Biome. Ruff, ty, and Vulture apply to Python analysis, while Biome covers JavaScript and TypeScript linting and formatting. The project documentation does not describe configuration environment variables, remote hosting, or a persistent project index.

The repository is licensed under MIT. Releases and container images are documented as being signed, but those supply-chain controls do not change the server’s analysis scope or the local execution requirements.

_Full upstream README: https://allmcps.com/mcp/mcp-server-analyzer/readme_