Skip to main content
Security policies configure TealGuard to protect your AI agents from data leakage, adversarial inputs, and harmful content generation. All evaluations are deterministic — no LLM in the governance path.

PII Redaction — Basic

Detect and redact emails, phone numbers, SSNs, and credit cards in agent I/O.

Prompt Injection — Strict

High-sensitivity prompt injection detection with automatic denial.

Content Moderation — Standard

Block hate speech, violence, sexual content, and self-harm references.

​
When to Use

Every AI agent that processes user input should have at least one security policy enabled. Start with PII Redaction and Prompt Injection as your baseline, then add Content Moderation for user-facing applications.
Combine security policies with Secrets and Memory policies for defense-in-depth. See the OWASP ASI Starter compliance pack for a comprehensive baseline.