Security-Use MCP Server
Enables AI assistants to scan project dependencies and Infrastructure as Code files for security vulnerabilities and misconfigurations. It also provides automated fixing capabilities to remediate identified security issues.
README
Security-Use MCP Server
An MCP (Model Context Protocol) server that gives AI assistants like Cursor, Claude, and other MCP-compatible tools the ability to scan for security vulnerabilities and automatically fix them.
What It Does
This MCP server exposes four powerful security tools to your AI assistant:
| Tool | Description |
|---|---|
scan_dependencies |
Scans your project's dependencies (requirements.txt, package.json, pyproject.toml, etc.) for known vulnerabilities using the OSV database |
scan_iac |
Scans Infrastructure as Code files (Terraform, CloudFormation) for security misconfigurations |
fix_vulnerability |
Automatically updates vulnerable packages to secure versions |
fix_iac |
Generates and applies fixes for IaC security issues |
Supported Formats
Dependency Scanning
- Python:
requirements.txt,pyproject.toml,Pipfile,Pipfile.lock,poetry.lock,setup.py - JavaScript/Node.js:
package.json,package-lock.json - Java:
pom.xml
Infrastructure as Code
- Terraform (
.tffiles) - AWS CloudFormation (
.yaml,.yml,.json) - AWS SAM templates
- AWS CDK synthesized output
IaC Security Rules
The scanner checks for common AWS misconfigurations including:
- CKV_AWS_20: S3 buckets with public access (CRITICAL)
- CKV_AWS_19: S3 buckets without encryption (HIGH)
- CKV_AWS_23: Security groups with unrestricted ingress (HIGH)
- CKV_AWS_16: RDS instances without encryption (HIGH)
- CKV_AWS_3: EBS volumes without encryption (HIGH)
- CKV_AWS_35: CloudTrail not enabled (MEDIUM)
- CKV_AWS_14: IAM users without MFA (MEDIUM)
- CKV_AWS_12: VPC without flow logs (MEDIUM)
Installation
From PyPI
pip install security-use-mcp
From Source
git clone https://github.com/security-use/mcp.git
cd mcp
pip install -e .
Quick Setup for Cursor
-
Install the package (see above)
-
Add to Cursor's MCP configuration (
~/.cursor/mcp.json):
{
"mcpServers": {
"security-use": {
"command": "security-use-mcp",
"args": [],
"env": {}
}
}
}
If you installed from source or use a virtual environment:
{
"mcpServers": {
"security-use": {
"command": "python",
"args": ["-m", "security_use_mcp.server"],
"env": {}
}
}
}
-
Restart Cursor
-
Test it - Open Cursor's AI chat and ask:
"Scan this project for security vulnerabilities"
Usage Examples
Once configured, you can ask your AI assistant things like:
Dependency Scanning
- "Scan this project for vulnerable dependencies"
- "Check if my Python packages have any CVEs"
- "Are there any security issues in my requirements.txt?"
IaC Scanning
- "Scan my Terraform files for security issues"
- "Check this CloudFormation template for misconfigurations"
- "Are my S3 buckets configured securely?"
Fixing Vulnerabilities
- "Fix the requests vulnerability"
- "Update django to a secure version"
- "Fix the S3 bucket public access issue in main.tf"
Example Output
Dependency Scan Results
## Dependency Security Scan Results
**Found 2 vulnerabilities**
### CRITICAL (1)
#### requests (2.25.0)
- **ID**: GHSA-xxxx-yyyy-zzzz
- **Title**: CVE-2023-32681 - Unintended leak of Proxy-Authorization header
- **Fixed in**: 2.31.0
### HIGH (1)
#### django (3.1.0)
- **ID**: CVE-2023-xxxxx
- **Title**: SQL Injection in QuerySet.values()
- **Fixed in**: 3.2.19
IaC Scan Results
## IaC Security Scan Results
**Found 2 security issues**
### CRITICAL (1)
#### CKV_AWS_20: S3 bucket with public access
- **File**: `s3.tf:15`
- **Resource**: aws_s3_bucket.my-bucket
- **Remediation**: Set acl to 'private'
### HIGH (1)
#### CKV_AWS_23: Security group allows unrestricted ingress
- **File**: `sg.tf:8`
- **Resource**: aws_security_group.web
- **Remediation**: Restrict CIDR blocks to specific IPs
Configuration Options
Environment Variables
| Variable | Description | Default |
|---|---|---|
SECURITY_USE_LOG_LEVEL |
Logging level (DEBUG, INFO, WARN, ERROR) | INFO |
SECURITY_USE_CACHE_DIR |
Directory for caching vulnerability data | System temp |
Example configuration with environment variables:
{
"mcpServers": {
"security-use": {
"command": "security-use-mcp",
"args": [],
"env": {
"SECURITY_USE_LOG_LEVEL": "DEBUG"
}
}
}
}
Development
Setup
# Clone the repository
git clone https://github.com/security-use/mcp.git
cd mcp
# Install with dev dependencies
pip install -e ".[dev]"
# Also install the core security-use package
pip install -e ../security-use
Running Tests
# Run all tests
pytest
# Run with coverage
pytest --cov=security_use_mcp
# Run specific test file
pytest tests/test_handlers.py -v
Linting
# Check code style
ruff check src/ tests/
# Auto-fix issues
ruff check src/ tests/ --fix
Testing the Server
You can test the MCP server directly:
# Start the server (it communicates via stdin/stdout)
python -m security_use_mcp.server
# Or use the entry point
security-use-mcp
Troubleshooting
Server Not Starting
-
Check Python version (requires 3.10+):
python --version -
Verify installation:
pip show security-use-mcp pip show security-use -
Test the server directly:
python -c "from security_use_mcp.server import server; print('OK')"
Tools Not Appearing in Cursor
- Restart Cursor after changing
mcp.json - Check that the JSON is valid
- Look for errors in Cursor's Developer Tools (Help > Toggle Developer Tools)
Scan Returns No Results
- Make sure you have dependency files (requirements.txt, package.json, etc.) or IaC files (.tf, .yaml) in your project
- Check that the path is correct when scanning specific directories
Architecture
security-use-mcp/
├── src/security_use_mcp/
│ ├── server.py # MCP server implementation
│ ├── models.py # Data models for results
│ └── handlers/ # Tool handlers
│ ├── dependency_handler.py
│ └── iac_handler.py
└── tests/
├── test_server.py # Server tests
├── test_handlers.py # Handler unit tests
└── test_integration.py # Integration tests
The MCP server wraps the security-use Python package, which provides:
- Dependency scanning via the OSV (Open Source Vulnerabilities) database
- IaC scanning with configurable security rules
- Automated fixing capabilities
Related Projects
- security-use - Core security scanning library
- security-use-vscode - VS Code extension
- MCP Specification - Model Context Protocol documentation
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。