Skip to main content

Guardrails & Security

How Apex AI Brain enforces security and best practices.

Guardrails System​

Guardrails are injected rules that constrain AI outputs to enterprise standards.

AI Response
    ↓
┌─────────────────┐
│ Guardrail Check │
├─────────────────┤
│ ✓ Syntax valid  │
│ ✓ Pattern match │
│ ✓ Security OK   │
│ ✓ Performance OK│
└────────┬────────┘
         ↓
     Approved

Key Guardrails​

Governor Limits Enforcement​

No governor limit violations allowed:

  • Max SOQL queries
  • Max DML statements
  • Batch sizes
  • CPU time limits
  • Heap memory limits

Bulkification Requirements​

All code must be bulkified:

  • No queries in loops
  • Batch DML operations only
  • Efficient collection usage
  • Map-based lookups

FSD Architecture​

All code follows FSD:

  • Proper folder structure
  • Clear layer separation
  • Appropriate dependencies
  • Naming conventions

Security Requirements​

All code validated for:

  • Input sanitization
  • Query parameter binding
  • XSS prevention
  • CSRF protection
  • Data access restrictions

Full Details: See Guardrails & Security Guide