CodeVitals
Provides offline repository health analysis, including metadata, weighted health score, and summary, for MCP-compatible clients.
README
CodeVitals
AI-powered Repository Health Analysis & Project Intelligence MCP Server.
Overview
CodeVitals is a Model Context Protocol (MCP) server that inspects a local repository and reports its metadata, a weighted health score, and a human-readable summary — entirely offline, with no external API calls. It is designed to be spawned by an MCP-compatible client (such as an AI assistant) and communicates over stdio.
Features
- Repository analysis — dependency counts, framework detection, git status, and file metrics.
- Architecture detection — project type, package manager, language, Docker, CI, and test setup.
- Security analysis — sensitive files,
.env/.gitignorechecks, and hardcoded-secret scanning. - Dependency health — duplicate, unused, and unpinned dependency detection.
- Dead code detection — unused files, orphan files, empty directories, and large files.
- Code quality — file-size distribution, large-function, deep-folder, and naming warnings.
- Health score — a weighted 0–100 score across security, architecture, dependencies, maintainability, and metrics.
- Deterministic summary — strengths, weaknesses, warnings, and quick wins. No LLM involved.
Requirements
- Node.js >= 22
Installation
Install globally:
npm install -g @prakhhxrcodes/codevitals
Or run it without installing:
npx @prakhhxrcodes/codevitals
From source
git clone https://github.com/prakharsharma13/codevitals.git
cd codevitals
npm install
npm run build
CLI Usage
CodeVitals ships a single executable, codevitals, that launches the MCP server
over stdio:
codevitals
The process is not an interactive command-line tool — it does not take subcommands or arguments. It starts a server that waits for an MCP client to connect over stdin/stdout. Running it directly in a terminal is mainly useful for verifying that it starts (it prints a startup message to stderr and then waits). In normal use it is launched automatically by an MCP client (see below).
MCP Usage
Register CodeVitals with any MCP-compatible client. Example client configuration:
{
"mcpServers": {
"codevitals": {
"command": "npx",
"args": ["@prakhhxrcodes/codevitals"]
}
}
}
If installed from source, point the client at the built entry file instead:
{
"mcpServers": {
"codevitals": {
"command": "node",
"args": ["/absolute/path/to/codevitals/dist/index.js"]
}
}
}
Tools
health_check
Checks whether the CodeVitals MCP server is running. No input.
analyze_repository
Analyzes a repository and returns its metadata, health score, and summary.
| Field | Type | Description |
|---|---|---|
repositoryPath |
string |
Absolute path to the repository to scan. |
Configuration
CodeVitals requires no environment variables or configuration files. Analysis
thresholds (for example, the large-file and deep-folder limits used by the
dead-code and code-quality analyzers) ship with sensible defaults defined in
their respective analyzers. The only per-call input is repositoryPath, passed
to the analyze_repository tool.
Example Outputs
analyze_repository returns a JSON document combining three sections:
{
"analysis": {
"dependencyCount": 2,
"devDependencyCount": 2,
"framework": "Unknown",
"isGitRepository": false,
"commitCount": 0,
"totalFiles": 45,
"totalDirectories": 9,
"totalLines": 3290,
"projectType": "Node",
"packageManager": "npm",
"language": "TypeScript",
"hasDocker": false,
"hasCI": false,
"hasTests": false
},
"health": {
"overallScore": 86,
"securityScore": 100,
"dependencyScore": 100,
"maintainabilityScore": 97,
"architectureScore": 35,
"summary": "Repository health: Excellent (86/100)."
},
"summary": {
"strengths": ["Uses TypeScript.", "Security risk is Low."],
"weaknesses": ["Missing Docker support.", "CI/CD pipeline not configured."],
"warnings": [],
"quickWins": ["Add a Dockerfile.", "Add a CI workflow."],
"overallSummary": "Repository scored 86/100. Security risk is Low."
}
}
Architecture
CodeVitals follows a strict layered architecture with constructor dependency injection and a single composition root:
index.ts— bootstraps the process: creates the server, wires the stdio transport, and connects.server/—create-server.tsassembles the application;dependencies.tsconstructs every dependency.tools/— MCP tool registration only; the sole layer aware of the MCP SDK.services/— orchestration and scoring; coordinate analyzers without doing I/O.analyzers/— pure analysis logic; never import the MCP SDK.infrastructure/— filesystem, git, and scanning; no business logic.types/,utils/,constants/— shared, dependency-free helpers.
Dependencies always point inward: tools → services → analyzers → infrastructure. Analyzers and infrastructure never depend on the MCP layer.
Folder Structure
src/
index.ts # entry point (stdio transport)
server/ # server assembly + dependency wiring
tools/ # MCP tool registration
services/ # orchestration + scoring + summary
analyzers/ # analysis logic
infrastructure/ # filesystem, git, scanning
types/ # shared interfaces
utils/, constants/ # helpers
Contributing
Contributions are welcome. See CONTRIBUTING.md for setup and
guidelines. Please make sure npm run build and npm run typecheck pass before
opening a pull request.
FAQ
Does CodeVitals send my code anywhere? No. All analysis runs locally. There are no network or external API calls.
Does it require an API key or an LLM? No. The health score and summary are fully deterministic.
Which languages does it analyze?
It targets JavaScript/TypeScript projects (it reads package.json and scans
common source extensions), but file, git, and structural metrics apply to any
repository.
Can I run it as a standalone CLI report? Not yet — it currently runs as an MCP server. A standalone reporting mode is on the roadmap.
Troubleshooting
The server starts and then exits immediately. That is expected when it is run directly without an MCP client: with no client attached to stdin, the stdio transport reaches end-of-input and shuts down. Run it through an MCP client instead.
analyze_repository returns an error.
Ensure repositoryPath is an absolute path to a directory that contains a
package.json. The tool returns a descriptive error message when the path is
missing or unreadable.
npx @prakhhxrcodes/codevitals cannot be found.
Make sure the package name is spelled correctly (including the @prakhhxrcodes/
scope) and that your Node.js version is 22 or newer.
Roadmap
- Standalone CLI reporting mode
- HTML / PDF report output
- GitHub integration
- npm registry integration for outdated-dependency checks
- AI-generated insights
- Plugin system for custom analyzers
License
MIT © Prakhar Sharma
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.