MCP Analyzer Server
Provides security analysis and malicious tool detection for MCP tools with context-aware detection, behavioral pattern analysis, and risk scoring.
README
MCP Analyzer Server
A dedicated MCP (Model Context Protocol) server for security analysis and malicious tool detection.
Features
analyze_mcp_tool: Enhanced context-aware security analysis for individual MCP toolsbatch_analyze_tools: Comprehensive security assessment for multiple tools at once- Context-aware detection: Reduces false positives and negatives
- Behavioral pattern analysis: Identifies tool poisoning and covert operations
- Risk scoring: Provides detailed risk assessment with confidence levels
Installation
npm install
npm run build
Usage
Start the server
npm start
# or
node dist/mcp-analyzer.js
Test the server
echo '{"jsonrpc": "2.0", "id": 1, "method": "tools/list", "params": {}}' | node dist/mcp-analyzer.js
Tools
analyze_mcp_tool
Analyzes a single MCP tool for malicious indicators.
Input:
tool_name: Name of the MCP tooltool_description: Description of the MCP tooltool_input_schema: Input schema (JSON string, optional)
Output:
- Comprehensive security analysis report
- Risk level assessment (LOW/MEDIUM/HIGH)
- Malicious indicators detection
- Security recommendations
batch_analyze_tools
Analyzes multiple MCP tools at once for comprehensive security assessment.
Input:
tools: Array of tools to analyze
Output:
- Batch analysis report for all tools
- Comparative risk assessment
- Security recommendations for the toolset
Security Analysis Features
- Context-aware detection: Uses semantic analysis to reduce false positives
- Behavioral pattern analysis: Identifies tool poisoning and covert operations
- Input schema analysis: Examines tool parameters for dangerous patterns
- Risk scoring: Provides detailed risk assessment with confidence levels
- Security recommendations: Offers specific mitigation strategies
Use Cases
- MCP tool security research: Analyze tools for malicious patterns
- Security assessment: Evaluate tool safety before deployment
- Compliance checking: Ensure tools meet security requirements
- Threat detection: Identify potentially dangerous MCP tools
License
MIT
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。