Netra PII And Input Guardrails

Use this skill to protect prompts and logs from sensitive data and malicious inputs.

When To Use

• You must detect/mask/block PII before sending text to LLMs.
• You need prompt-injection scanning before agent/tool execution.
• You want explicit security actions (MASK, FLAG, BLOCK).

Procedure

1. Choose a PII detector (get_default_detector or RegexPIIDetector).
2. Set action type based on policy (MASK for safe continuation is common).
3. Scan user input with InputScanner before processing.
4. Stop requests on blocked violations.
5. Send masked content downstream when allowed.