Code Injection

Identifies SQL injection, XSS, and command injection

API Field: code_injection_enabled

Overview

Code Injection Detection identifies malicious code patterns that could exploit vulnerabilities in your application. This includes SQL injection, Cross-Site Scripting (XSS), command injection, and other code-based attacks that might be passed through AI interactions.

What It Detects

SQL injection attempts
Cross-Site Scripting (XSS) payloads
Command injection (shell commands)
LDAP injection
XML/XXE injection
Path traversal attacks
Template injection
Code execution attempts

Why It Matters

Code injection attacks can lead to complete system compromise, data theft, service disruption, and ransomware deployment. These attacks exploit the trust boundary between AI-generated content and backend systems.

Technical Details

Risk Score Range

0.0 - 1.0 (High risk: > 0.6)

Confidence Level

Typically 0.88 - 0.98

Processing Time

< 60ms per scan

Common Use Cases

AI-powered search interfacesNatural language to SQL convertersCode generation toolsAutomated testing assistantsDatabase query builders

Detection Examples

Threat Example #1Risk: 97%

'; DROP TABLE users; --

Classic SQL injection attempting to delete database table.

Threat Example #2Risk: 95%

XSS payload attempting to steal session cookies.

Threat Example #3Risk: 99%

| rm -rf / --no-preserve-root

Command injection attempting to delete all files on the system.

API Usage

Enable this scanner in your API request by setting code_injection_enabled to true in your API key settings, or include it in your request:

curl -X POST https://benguard.io/api/v1/scan \
  -H "X-API-Key: ben_your_api_key_here" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "Your user input here"
  }'

The scanner settings are configured per API key in your dashboard under Settings → Scanner Configuration.

Response Format

When this scanner detects a threat, the response will include:

{
  "is_valid": false,
  "status": "threat_detected",
  "risk_score": 0.97,
  "threat_types": ["code_injection"],
  "details": {
    "results": [
      {
        "scanner": "code_injection",
        "threat_detected": true,
        "risk_score": 0.97,
        "confidence": 0.92,
        "details": {
          "reason": "Classic SQL injection attempting to delete database table.",
          "evidence": ["detected pattern in input"]
        }
      }
    ]
  },
  "request_id": "req_abc123"
}

Best Practices

Never execute AI-generated code without sandboxing
Use parameterized queries, not string concatenation
Implement content security policies
Validate all outputs before rendering
Use allowlists for permitted operations

Related Scanners

Consider enabling these related scanners for comprehensive protection:

Prompt Injection

Detects attempts to override system instructions

Data Exfiltration

Detects attempts to extract system prompts and hidden data