gws-modelarmor-sanitize-prompt

To sanitize user prompts, you can pass the text through a predefined security template using a CLI command with --template and --text inputs. This processes incoming prompts to block or rewrite unsafe content before downstream AI processing.

Prompt sanitization enforces safety rules by applying a predefined template to incoming user inputs. It blocks or rewrites unsafe content, preventing the leakage of sensitive data and ensuring AI responses remain compliant.

To prevent sensitive data leakage in customer-facing AI applications, you should sanitize prompts using a Model Armor template before processing. This enforces safety rules and handles confidential information appropriately.

Yes, you can use a Model Armor template to enforce safety rules across user inputs. Template-based sanitization applies predefined security configurations to incoming prompts, validating and sanitizing the text before it reaches downstream processing.

Yes, prompt sanitization supports CLI commands for automated workflows. You can execute the sanitizer by passing your prompt text and security template via command-line arguments, returning a validated, sanitized prompt for immediate downstream use.