goldmembrane/cleaner-code

What it does

Cleaner Code is an MCP (Model Context Protocol) server that functions as an AI code security scanner specifically designed to identify threats and vulnerabilities in AI-generated code. It detects 9 distinct categories of security threats including invisible Unicode characters, Trojan Source attacks, homoglyphs, Glassworm steganography, rules file backdoors, dependency typosquatting, and code obfuscation.

How to set up

As an MCP server, installation involves configuring it within your MCP-compatible environment. Clone or download the goldmembrane/cleanercode repository from GitHub, install required dependencies, and configure the server in your MCP client configuration. The server integrates with supported platforms including Claude, ChatGPT, and AI Studio through the MCP protocol.

Key features

• Detects 9 categories of code security threats
• Static analysis combined with CodeBERT machine learning model
• Identifies invisible Unicode and homoglyph attacks
• Detects Trojan Source and steganographic threats
• Analyzes dependency vulnerabilities and typosquatting
• Recognizes code obfuscation patterns
• Designed specifically for AI-generated code analysis