Safety and Data Protection

Guardrails are content filtering rules the gateway enforces inline: on the way in (prompts) and on the way out (responses). These guides cover the operator side: configuring vendor-provided and custom guardrails, redacting sensitive data, blocking prompt injection, and applying stricter policy bundles to coding assistant traffic.

---

Configure vendor guardrails

Enable provider-backed guardrails and choose block or flag behavior.

Configure custom guardrails

Keyword and regex rules for organization-specific policies.

Detect and redact sensitive data

Keep PII and credentials out of prompts, responses, and logs.

Detect and block prompt injection

Stop jailbreak patterns before they reach a model.

Protect coding assistant traffic

Apply stricter policy bundles to agent and editor traffic without slowing normal use.