anmalkov/mcp-crisp repository overview

⁠CRISP MCP Server

. .

⁠What is CRISP MCP Server?

The CRISP MCP Server is a comprehensive security analysis tool that provides:

• Security Recommendations for Azure resources based on best practices
• Threat Modeling with categorized threats and mitigation strategies
• Security Controls organized by Azure service types
• Security Plan Templates for developing comprehensive security strategies

⁠🔧 Usage

⁠VS Code Integration

1. Install and Configure MCP in VS Code

   • Install the VS Code 1.99 or later
   • To enable MCP support in VS Code, enable the chat.mcp.enabled setting.

2. Add an MCP server to your workspace

   • Create a .vscode/mcp.json file in your workspace folder
   • Add to your VS Code settings or MCP configuration:

    {
        "servers": {
            "crisp": {
                "command": "docker",
                "args": [
                    "run",
                    "--rm",
                    "-i",
                    "--name", "crisp-mcp-server",
                    "anmalkov/mcp-crisp:latest"
                ],
                "env": {}
            }
        }
    }
   
   

   Copy

⁠Manual Usage

⁠Pull and Run

docker pull anmalkov/mcp-crisp:latest
docker run -it --rm anmalkov/mcp-crisp:latest

The server will start and listen for MCP protocol messages via stdio.

⁠🛠️ Available Tools

The CRISP MCP Server provides 7 powerful tools for security analysis:

⁠1. get_resources

Description: Retrieve all supported Azure resources.

• Input: None required
• Output: List of supported Azure resource types
• Use Case: Discover what Azure services have security recommendations available

⁠2. get_resources_recommendations

Description: Get security recommendations for specific Azure resources.

• Input: resource_names (optional) - Comma-separated list of Azure resource names
• Output: Detailed security recommendations with categories
• Use Case: Get tailored security advice for your Azure infrastructure

Example:

{
  "resource_names": "Azure App Service,Azure Key Vault,Azure Storage Account"
}

⁠3. get_threat_categories

Description: Get all available threat categories.

• Input: None required
• Output: List of threat categories (e.g., AI, Applications, Containers, Data Stores)
• Use Case: Understand the scope of threats covered in the threat model

⁠4. get_threats_by_categories

Description: Get detailed threats for specific categories.

• Input: category_names (optional) - Comma-separated list of threat categories
• Output: Detailed threat descriptions and mitigation strategies
• Use Case: Focus on specific threat vectors relevant to your architecture

Example:

{
  "category_names": "AI,Containers,Applications"
}

⁠5. get_recommendation_categories

Description: Get all security recommendation categories.

• Input: None required
• Output: List of security recommendation categories
• Use Case: Understand the different types of security controls available

⁠6. get_recommendations_by_categories

Description: Get security recommendations filtered by category.

• Input: category_names (optional) - Comma-separated list of recommendation categories
• Output: Categorized security recommendations
• Use Case: Focus on specific areas of security (e.g., IAM, Network, Data Protection)

Example:

{
  "category_names": "Identity and Access Management (IAM),Network Boundaries,Data Classification and Protection"
}

⁠7. get_template

Description: Get the security plan template in markdown format.

• Input: None required
• Output: Comprehensive security plan template
• Use Case: Create structured security documentation and plans

⁠💬 Prompts

The CRISP MCP Server now supports three intelligent prompts designed to enhance security analysis workflows through natural language interactions and advanced automation:

⁠1. code_analysis

Description:
Performs a comprehensive security analysis of code to identify vulnerabilities, insecure patterns, and offers remediation suggestions.

• Input: Source code
• Output: Security analysis report as a markdown file under docs/security
  Filename format: YYYYMMDD-<short-description>-report.md
• Use Case: Run the code analysis to ensure code security before deployment

⁠2. resource_recommendations

Description:
Provides detailed security recommendations for Azure resources based on the Microsoft Cloud Security Benchmark.

• Input: resource_names – Comma-separated list of Azure resource names
  (e.g., "Azure App Service,Azure SQL Database")
• Output: Recommendation report as a markdown file under docs/security
  Filename format: YYYYMMDD-azure-resource-recommendations.md
• Use Case: Ensure Azure infrastructure adheres to recognized security standards

⁠3. build_security_plan

Description:
Generates a structured and detailed security plan by analyzing architectural diagrams.

• Input: Text-based diagram file (e.g., Draw.io .drawio, Mermaid.js .mmd) placed in the ./diagrams directory
• Output: Comprehensive security plan in markdown format, saved under docs/security
  Filename format: YYYYMMDD-security-plan.md
• Use Case: Automate the creation of security plan from existing architecture diagrams

⁠📋 Security Coverage

The CRISP MCP Server covers comprehensive security domains:

⁠Security Domains

• Identity and Access Management (IAM)
• Data Classification and Protection
• Network Boundaries
• Input Validation
• Keys and Secret Management
• Observability - Monitoring and Logging
• PaaS Services Security
• IaaS Services Security
• Storage Services Security
• Microservices (AKS) Security
• IoT Security
• CI/CD Security

⁠Threat Categories

• AI Threats (Prompt injection, Model theft, etc.)
• Application Threats
• Container Security
• Data Store Threats
• DevOps Security
• General Security Threats

⁠🐳 Docker Image Information

⁠Available Tags

• latest - Latest stable release
• v{version} - Specific version releases (e.g., v0.1.0)

⁠Multi-Platform Support

• linux/amd64
• linux/arm64

⁠Image Size

Optimized Docker image with minimal footprint for efficient deployment.

⁠🛡️ Security Considerations

• The server provides security recommendations but doesn't access your actual Azure resources
• All threat and recommendation data is based on security best practices and frameworks
• No sensitive data is stored or transmitted by the server
• Runs in isolated Docker container for additional security

⁠📝 License

This project is licensed under the MIT License - see the LICENSE⁠ file for details.

⁠Built with ❤️ for the security community