CISA KEV MCP Server

CISA KEV MCP Server

Provides access to CISA's Known Exploited Vulnerabilities catalog via the Model Context Protocol, enabling querying, searching, and analyzing vulnerability data.

Category
访问服务器

README

CISA KEV MCP Server

This MCP (Model Context Protocol) server provides access to CISA's Catalog of Known Exploited Vulnerabilities (KEV). It allows AI assistants and other clients to query and analyze vulnerability data using the Model Context Protocol.

Features

  • Loads and caches the KEV data with daily refresh
  • Exposes comprehensive tools to query and search the KEV database
  • Provides statistical analysis of vulnerability data
  • Implements filtering by multiple criteria
  • Follows the Model Context Protocol for standardized AI assistant integration
  • Delivers type-safe responses

Installation

npm install
npm run build

Usage

Quick Start with npx (Recommended)

The easiest way to use the KEV MCP server is via npx. The published package includes self-contained bundles that don't require local dependencies:

STDIO Transport (Default)

# Main command - defaults to stdio transport
npx @hrbrmstr/kev-mcp

# Explicit stdio transport
npx @hrbrmstr/kev-mcp-stdio

HTTP Transport

# Using the main command with flag
npx @hrbrmstr/kev-mcp --transport http

# Using the dedicated HTTP command
npx @hrbrmstr/kev-mcp-http

HTTP Request Logging

When using the HTTP transport, the server provides comprehensive structured logging for monitoring, debugging, and security auditing.

Log Format

All logs are output as structured JSON to stderr with the following format:

{
  "level": "info",
  "component": "http-transport",
  "timestamp": "2024-01-15T10:30:45.123Z",
  "method": "POST",
  "path": "/mcp",
  "userAgent": "Mozilla/5.0...",
  "contentType": "application/json",
  "contentLength": 256,
  "remoteAddr": "192.168.1.100",
  "requestId": "abc123def456",
  "duration": 125,
  "status": 200,
  "responseSize": 512,
  "mcpMethod": "tools/call",
  "mcpId": 1,
  "mcpParams": ["query", "limit"]
}

Log Components

  • http-transport: HTTP server and request handling
  • stdio-transport: STDIO transport (when used)
  • server: General server lifecycle events

MCP-Specific Logging

The logging middleware extracts MCP-specific information:

  • mcpMethod: The MCP method being called
  • mcpId: Request ID from the MCP protocol
  • mcpParams: Array of parameter names (not values for security)
  • mcpError: Structured error information when requests fail

Configuration

Environment Variables

  • LOG_HEALTH_CHECKS=true: Enable logging of successful health check requests (default: only errors are logged)
  • PORT: Set the HTTP server port (default: 9191)
  • KEV_URL: Override the default CISA KEV feed URL
  • KEV_MIRROR_URL: Override the KEV feed URL (takes precedence over KEV_URL)
  • KEV_CACHE_PATH: Persist KEV cache JSON to disk (defaults to in-memory only)

Log Analysis Examples

# Start server and pipe logs through jq for formatting
npx @hrbrmstr/kev-mcp-http 2>&1 | jq '.'

# Filter error logs only
npx @hrbrmstr/kev-mcp-http 2>&1 | jq 'select(.level == "error")'

# Monitor request performance
npx @hrbrmstr/kev-mcp-http 2>&1 | jq 'select(.duration > 1000)'

# Track specific MCP methods
npx @hrbrmstr/kev-mcp-http 2>&1 | jq 'select(.mcpMethod == "tools/call")'

# Correlate logs by request ID
npx @hrbrmstr/kev-mcp-http 2>&1 | jq 'select(.requestId == "abc123def456")'

Request Correlation

Each HTTP request receives a unique requestId that appears in all related log entries, making it easy to trace the complete lifecycle of a request from start to finish.

Security Features

  • IP addresses are captured for security monitoring
  • User-Agent strings are logged for client identification
  • Authentication headers are redacted (shown as [REDACTED])
  • Request bodies are not logged to prevent sensitive data exposure
  • Only parameter names (not values) are logged for MCP requests

Local Development

If you've cloned and built the project locally:

npm install
npm run build

# Run locally built version
node build/kev-mcp-bundle.cjs
node build/kev-mcp-bundle.cjs --transport http

Local Tooltest Harness

Run the local tooltest integration against the HTTP transport (no external network calls):

cargo install --git https://github.com/lambdamechanic/tooltest
npm run tooltest:local

MCP Client Configuration

For STDIO Transport (Most Common)

{
  "mcpServers": {
    "cisa-kev": {
      "command": "npx",
      "args": ["@hrbrmstr/kev-mcp"]
    }
  }
}

For HTTP Transport

{
  "mcpServers": {
    "cisa-kev-http": {
      "command": "npx",
      "args": ["@hrbrmstr/kev-mcp-http"]
    }
  }
}

Local Build Configuration

{
  "mcpServers": {
    "cisa-kev": {
      "command": "node",
      "args": ["/path/to/kev-mcp/build/kev-mcp-bundle.cjs"]
    }
  }
}

Self-Contained Bundles

The published package includes bundled executables (~1.2MB each) that contain all dependencies:

  • kev-mcp-bundle.cjs - Main CLI supporting both transports
  • kev-mcp-http-bundle.cjs - HTTP-only version
  • index-bundle.js - Library for programmatic use

These bundles work without node_modules and are perfect for distribution via npx.

Tools Reference

The server provides the following tools for interacting with the KEV catalog:

Basic Data Retrieval

get_kev_count

Returns the total number of vulnerabilities in the KEV catalog.

Parameters: None

Example Response:

"1366"

get_kev_release_date

Returns the release date of the current KEV catalog.

Parameters: None

Example Response:

"2023-11-15"

get_vulnerability_details

Retrieves detailed information about a specific vulnerability by CVE ID.

Parameters:

  • cveId (string): The CVE ID to lookup (e.g., "CVE-2021-34527")

Example Response:

{
  "cveID": "CVE-2021-34527",
  "vendorProject": "Microsoft",
  "product": "Windows Print Spooler",
  "vulnerabilityName": "Microsoft Windows Print Spooler Remote Code Execution Vulnerability",
  "dateAdded": "2021-07-02",
  "shortDescription": "Microsoft Windows Print Spooler contains a remote code execution vulnerability...",
  "requiredAction": "Apply updates per vendor instructions.",
  "dueDate": "2021-07-16",
  "knownRansomwareCampaignUse": "Known",
  "notes": "This vulnerability has been exploited in the wild as part of ransomware attacks.",
  "cwes": ["CWE-269"]
}

Lists and Collections

get_kev_vendors

Returns a list of all unique vendors in the KEV catalog.

Parameters: None

Example Response:

["Adobe", "Apache", "Apple", "Atlassian", ...]

get_kev_products

Returns a list of all unique products in the KEV catalog.

Parameters: None

Example Response:

["Access", "Acrobat Reader", "ActiveMQ", "Adaptive Security Appliance", ...]

get_kev_cves

Returns a list of all CVE IDs in the KEV catalog.

Parameters: None

Example Response:

["CVE-2017-11882", "CVE-2018-13379", "CVE-2018-7600", ...]

get_recent_vulnerabilities

Returns vulnerabilities added to the KEV catalog within a specified time period.

Parameters:

  • days (number, optional): Number of days to look back (default: 30)

Example Response:

{
  "count": 15,
  "vulnerabilities": [
    {
      "cveID": "CVE-2023-12345",
      "vendorProject": "Example Vendor",
      ...
    },
    ...
  ]
}

get_upcoming_due_dates

Identifies vulnerabilities with upcoming remediation due dates.

Parameters:

  • days (number, optional): Number of days to look ahead (default: 30)

Example Response:

{
  "count": 8,
  "vulnerabilities": [
    {
      "cveID": "CVE-2023-67890",
      "vendorProject": "Example Vendor",
      "dueDate": "2023-12-15",
      ...
    },
    ...
  ]
}

Search and Filtering

search_kev

Comprehensive search tool with multiple filtering options.

Parameters:

  • searchText (string, optional): Text to search in vulnerability names, descriptions, and notes
  • ransomwareUse (array of strings, optional): Filter by ransomware usage ("Known" or "Unknown")
  • cwes (array of strings, optional): Filter by CWE identifiers
  • vendors (array of strings, optional): Filter by vendor/project names
  • products (array of strings, optional): Filter by product names
  • dateAddedStart (string, optional): Start date for dateAdded range (YYYY-MM-DD)
  • dateAddedEnd (string, optional): End date for dateAdded range (YYYY-MM-DD)
  • dateAdded (array of strings, optional): Specific dateAdded values to match
  • dueDateStart (string, optional): Start date for dueDate range (YYYY-MM-DD)
  • dueDateEnd (string, optional): End date for dueDate range (YYYY-MM-DD)
  • dueDate (array of strings, optional): Specific dueDate values to match

Example Response:

{
  "count": 5,
  "vulnerabilities": [
    ...
  ]
}

get_related_cves

Finds vulnerabilities related to a specific vendor or product.

Parameters:

  • vendor (string, optional): Vendor name to find related CVEs
  • product (string, optional): Product name to find related CVEs
  • limit (number, optional): Maximum number of results to return (default: 20)

Note: If neither vendor nor product is provided, results are unfiltered (up to limit).

Example Response:

{
  "count": 12,
  "totalMatches": 24,
  "vulnerabilities": [
    ...
  ]
}

search_by_cwe

Locates vulnerabilities associated with a specific Common Weakness Enumeration (CWE).

Parameters:

  • cwe (string): CWE identifier to search for (e.g., "CWE-79" or "79")

Example Response:

{
  "count": 17,
  "cwe": "CWE-79",
  "vulnerabilities": [
    ...
  ]
}

Statistical Analysis

get_kev_statistics

Provides statistical analysis of the KEV catalog.

Parameters: None

Example Response:

{
  "totalVulnerabilities": 436,
  "topVendors": [
    {"vendor": "Microsoft", "count": 89},
    {"vendor": "Adobe", "count": 45},
    ...
  ],
  "ransomwareUsage": {
    "Known": 128,
    "Unknown": 308
  },
  "countByYear": {
    "2021": 168,
    "2022": 175,
    "2023": 93
  }
}

get_cwe_statistics

Provides statistical analysis of CWEs in the KEV catalog.

Parameters: None

Example Response:

{
  "uniqueCweCount": 42,
  "topCwes": [
    {"cwe": "CWE-79", "count": 48},
    {"cwe": "CWE-787", "count": 37},
    ...
  ]
}

Data Management

force_refresh_kev_data

Forces a refresh of the KEV data cache to ensure the latest information is available.

Parameters: None

Example Response:

KEV data successfully refreshed. Current catalog version: 2023.11.15, with 436 vulnerabilities.

Project Structure

├── src/
│   ├── index.ts           # Main server entry point
│   ├── types.ts           # TypeScript interfaces
│   ├── utils.ts           # Utility functions and constants
│   └── tools/             # Individual tools implementation
│       ├── index.ts       # Tool registration
│       ├── getKevCount.ts # Count tool
│       ├── ...
├── build/                 # Compiled JavaScript files
├── package.json
└── README.md

License

MIT

推荐服务器

Baidu Map

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。

官方
精选
JavaScript
Playwright MCP Server

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。

官方
精选
TypeScript
Magic Component Platform (MCP)

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。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

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。

官方
精选
本地
TypeScript
VeyraX

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。

官方
精选
本地
graphlit-mcp-server

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。

官方
精选
TypeScript
Kagi MCP Server

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。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

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。

官方
精选