Skip to main content

MCP Tools Reference

Complete reference for all Security Scanner MCP tools.

Core Scanning Tools​

scan-security​

Description: Comprehensive security scan - runs all checks at once

Usage: Best for general purpose scanning

Me: Scan this code for security issues

What it does:

  • Runs all 7 scanners simultaneously
  • Combines results
  • Returns comprehensive report

scan-secrets​

Description: Detect hardcoded secrets

Detects:

  • API keys (AWS, Google, GitHub, Stripe, etc.)
  • Passwords and tokens
  • Database connection strings
  • Private keys
  • OAuth secrets

Usage:

Me: Check for hardcoded secrets in this code

scan-injection​

Description: Find SQL/NoSQL/Command injection vulnerabilities

Detects:

  • SQL injection (string concatenation, template literals)
  • NoSQL injection (MongoDB)
  • Command injection (exec, spawn, system)
  • LDAP injection

Usage:

Me: Scan for injection vulnerabilities

scan-xss​

Description: Identify Cross-Site Scripting risks

Detects:

  • dangerouslySetInnerHTML (React)
  • innerHTML / outerHTML
  • jQuery .html()
  • Vue v-html
  • eval() / new Function()

Usage:

Me: Check for XSS vulnerabilities

scan-crypto​

Description: Check cryptographic weaknesses

Detects:

  • Weak hashing (MD5, SHA1)
  • Insecure random (Math.random)
  • Hardcoded keys/IVs
  • SSL validation disabled
  • Vulnerable TLS versions

Usage:

Me: Analyze cryptographic security

scan-auth​

Description: Audit authentication and session security

Detects:

  • JWT misconfigurations
  • Insecure cookies
  • CORS wildcards
  • Weak password policies
  • Session fixation risks

Usage:

Me: Review authentication security

scan-path​

Description: Find file and path vulnerabilities

Detects:

  • Path traversal
  • Dangerous file operations
  • Insecure file uploads
  • Zip Slip (Java)
  • Pickle deserialization (Python)

Usage:

Me: Check for path traversal issues

scan-dependencies​

Description: Check for vulnerable dependencies

Checks:

  • package.json (npm audit)
  • requirements.txt (Python)
  • go.mod (Go)

Usage:

Me: Scan dependencies for vulnerabilities

Infrastructure Tools​

scan-iac​

Description: Scan Infrastructure as Code files

Supports:

  • Dockerfile (CIS Docker Benchmark)
  • Kubernetes YAML (Pod Security Standards)
  • Terraform HCL (Multi-cloud)

Usage:

Me: Scan this Dockerfile

Advanced Tools​

get-fix-suggestion​

Description: Get auto-generated fix code

Returns:

  • Before/After code comparison
  • Explanation
  • Alternative solutions

Usage:

Me: How do I fix this SQL injection?

Parameters:

  • issue: Description of the vulnerability
  • code: Original vulnerable code
  • language: Programming language

generate-security-report​

Description: Create comprehensive security reports

Generates:

  • Mermaid diagrams (pie, bar, flowchart)
  • SARIF format (GitHub Code Scanning compatible)
  • CVE/OWASP information
  • Attack scenario analysis

Usage:

Me: Generate a full security report with diagrams

Output includes:

  • Overall summary
  • Severity distribution chart
  • Vulnerability categories chart
  • Attack scenario flowchart
  • SARIF JSON for CI/CD integration

scan-in-sandbox​

Description: Run scans in Docker isolated environment

Features:

  • Memory/CPU limits
  • Network isolation
  • External tools (Trivy, Checkov, GitLeaks)

Usage:

Me: Scan this code in a sandbox

Security settings:

  • Memory: 128MB - 2GB
  • CPU: 0.1 - 2.0 cores
  • Timeout: 5s - 5min
  • Network: Disabled
  • Privileges: Minimal

Tool Parameters​

Common Parameters​

All scanning tools accept:

  • code: Source code to scan (string)
  • language: Programming language (optional, auto-detected)
  • filename: Original filename (optional)

Language Detection​

Auto-detected languages:

  • JavaScript
  • TypeScript
  • Python
  • Java
  • Go

Tool Response Format​

Standard Issue Format​

interface SecurityIssue {
  type: string;          // Vulnerability type
  severity: string;      // critical | high | medium | low
  message: string;       // Human-readable description
  fix: string;          // Remediation suggestion
  line?: number;        // Line number (1-indexed)
  match?: string;       // Matched text (masked for secrets)
  owaspCategory?: string;  // OWASP Top 10 mapping
  cweId?: string;       // CWE identifier
  metadata?: object;    // Tool-specific data
}

Scan Result Format​

interface ScanResult {
  success: boolean;
  issues: SecurityIssue[];
  summary: {
    critical: number;
    high: number;
    medium: number;
    low: number;
  };
}

Error Handling​

Tools return errors in standard format:

{
  "success": false,
  "error": "Error description",
  "code": "ERROR_CODE"
}

Common error codes:

  • INVALID_INPUT: Invalid code or parameters
  • LANGUAGE_NOT_SUPPORTED: Unsupported language
  • SCAN_TIMEOUT: Scan exceeded time limit
  • DOCKER_NOT_AVAILABLE: Docker not installed (sandbox only)

Best Practices​

  1. Use scan-security first: Get comprehensive overview
  2. Target specific scanners: Use individual tools for deep analysis
  3. Request fixes: Follow up scans with get-fix-suggestion
  4. Generate reports: Use for documentation and CI/CD
  5. Sandbox untrusted code: Use scan-in-sandbox for unknown sources

Next Steps​