cisco-ai-defense/mcp-scanner

Scan MCP servers for potential threats & security findings.

What it solves

MCP Scanner addresses the security risks associated with the Model Context Protocol (MCP), specifically the danger of malicious MCP servers, tools, prompts, and resources. It helps developers and security researchers detect threats like tool poisoning, prompt injection, and malware bundled within MCP server packages before they are deployed or used.

How it works

The tool employs a multi-engine analysis approach to identify security findings:

  • Scanning Engines: It combines the Cisco AI Defense inspect API, YARA rules for pattern matching, and LLM-based semantic analysis.
  • Broad Scope: It scans various MCP components, including tools, prompts, resources, and server instructions. It also performs behavioral code scanning on source code to find mismatches between docstrings and actual implementation.
  • Package & Binary Analysis: It integrates pip-audit for Python dependency vulnerabilities, uses VirusTotal hash lookups for binary malware detection, and utilizes Docker sandboxes for behavioral analysis of PyPI and npm packages.
  • Flexible Deployment: It can be run as a standalone CLI tool, a REST API server, or integrated as a Python SDK.

Who it’s for

  • Security Engineers: To audit MCP servers and tools for vulnerabilities and malicious behavior.
  • AI Developers: To ensure the MCP servers they build or integrate are production-ready and secure.
  • DevOps/CI/CD Pipelines: To automate security scanning of MCP configurations and JSON files in air-gapped or automated environments.

Highlights

  • Multi-Engine Analysis: Uses YARA, LLMs, and Cisco AI Defense for comprehensive detection.
  • Comprehensive Coverage: Scans tools, prompts, resources, instructions, and source code across multiple languages.
  • Sandbox Scanning: Performs behavioral analysis of PyPI and npm packages within Docker sandboxes.
  • Supply Chain Security: Integrates VirusTotal for binary scanning and pip-audit for dependency checks.
  • Static & Live Scanning: Supports both live server connections (SSE/HTTP) and offline static JSON scanning.

Related

  • Project
    korotovsky/slack-mcp-server

    An MCP server that enables AI models to interact with Slack workspaces, allowing them to read, search, and manage messages, users, and channels.

  • Project
    cisco-ai-defense/skill-scanner

    A security scanner for AI Agent Skills that detects prompt injection, data exfiltration, and malicious code using static analysis, behavioral dataflow, and LLM-based semantic analysis.

  • Project
    vmoranv/jshookmcp

    An MCP server that provides AI agents with a comprehensive suite of reverse-engineering tools for JavaScript, WASM, and native processes, featuring an adaptive tool-discovery system to manage token usage.

  • Project
    designcomputer/mysql_mcp_server

    A Model Context Protocol (MCP) server that enables AI applications to securely explore and query MySQL databases through a structured interface.

  • Project
    johnhuang316/code-index-mcp

    An MCP server that provides intelligent code indexing and AST-based analysis to help AI assistants navigate and understand complex codebases.