cisco-ai-defense/mcp-scanner
A Python security scanner that detects malicious MCP servers and tools using YARA rules, LLM-as-judge, and Cisco AI Defense APIs.
Not currently ranked — collecting fresh signals.
star history
The MCP Scanner is a developer security tool designed to audit AI agent infrastructure built on the Model Context Protocol. It combines three scanning engines—YARA rules for static pattern matching, LLM-as-judge for behavioral analysis, and Cisco AI Defense for deep inspection—to identify potentially malicious or vulnerable MCP tools. The tool can be run as a standalone CLI or REST API server, and also includes dependency vulnerability scanning for Python packages.
Frequently asked
- What is cisco-ai-defense/mcp-scanner?
- A Python security scanner that detects malicious MCP servers and tools using YARA rules, LLM-as-judge, and Cisco AI Defense APIs.
- Is mcp-scanner open source?
- Yes — cisco-ai-defense/mcp-scanner is open source, released under the Apache-2.0 license.
- What language is mcp-scanner written in?
- cisco-ai-defense/mcp-scanner is primarily written in Python.
- How popular is mcp-scanner?
- cisco-ai-defense/mcp-scanner has 989 stars on GitHub.
- Where can I find mcp-scanner?
- cisco-ai-defense/mcp-scanner is on GitHub at https://github.com/cisco-ai-defense/mcp-scanner.