Prompts Index

MCP prompt handler providing pre-configured security workflows for AI assistants to guide users through complex security analysis tasks.

Core Purpose

  • Provide structured security workflows
  • Generate multi-step analysis instructions
  • Handle prompt arguments and templating
  • Enable consistent security guidance

Key Components

PromptHandler Class

export class PromptHandler {
    // Available prompts
    SECURITY_ANALYSIS: 'security_analysis'
    THREAT_INVESTIGATION: 'threat_investigation' 
    COMPLIANCE_CHECK: 'compliance_check'
    INCIDENT_RESPONSE: 'incident_response'
    
    // MCP methods
    listPrompts(): Returns all available prompts
    getPrompt(): Generates specific workflow
}

Available Workflows

Security Analysis

  • Comprehensive threat scanning
  • Risk assessment and scoring
  • Remediation recommendations
  • Arguments: content, context, severity_threshold

Threat Investigation

  • Deep-dive into detected threats
  • Attack vector analysis
  • False positive assessment
  • Arguments: scan_id, focus_area

Compliance Check

  • Regulatory compliance verification
  • PII/PHI/PCI detection
  • Risk matrix generation
  • Arguments: content, regulations, profile_name

Incident Response

  • Step-by-step incident handling
  • Time-based action phases
  • Evidence preservation guide
  • Arguments: incident_type, report_id, urgency

Integration in Application

  • Used By: MCP server for prompt requests
  • Integrates With: Tools for security scanning
  • Pattern: Template-based workflow generation
  • Output: Structured conversation messages

Workflow Structure

Each prompt generates:

  1. User message with detailed instructions
  2. Assistant acknowledgment message
  3. Structured steps using MCP tools
  4. Professional report formatting

Example Usage

// Get security analysis workflow
const workflow = handler.getPrompt({
    name: 'security_analysis',
    arguments: {
        content: 'Analyze this text',
        severity_threshold: 'high'
    }
});

// Returns structured messages:
// - Instructions for using airs_scan_content
// - Requirements for analysis
// - Report structure template

Key Features

Dynamic Content

  • Argument-based customization
  • Default value handling
  • Conditional instructions

Professional Output

  • Markdown formatting
  • Clear section headers
  • Step-by-step guidance

Tool Integration

  • References specific MCP tools
  • Provides usage instructions
  • Chains tool operations
Edit this page on GitHub