Gws Modelarmor Sanitize Response

What it does

This skill sanitizes AI agent responses generated by Google Workspace models to ensure they comply with safety guidelines and organizational policies. It acts as a critical filter layer, removing potentially harmful or non-compliant content before the final output reaches the user.

When to use it

• Deploying internal AI agents that generate sensitive corporate reports or communications.
• Integrating Google Workspace models into customer-facing chatbots where brand safety is paramount.
• Automating content creation workflows to prevent accidental leaks of proprietary information.
• Enforcing strict data privacy standards across all model-generated outputs within the organization.

Key capabilities

• Real-time sanitization of generated text responses.
• Enforcement of Google Workspace-specific safety protocols.
• Prevention of policy-violating content from being displayed to end-users.
• Seamless integration within the googleworkspace/cli ecosystem.

Example prompts

• "Configure the agent to sanitize all responses before sending them to HR staff."
• "Apply strict model armor sanitization rules to this customer support chatbot."
• "Ensure that any output from the Google Workspace model passes through the response sanitizer."

Tips & gotchas

This skill is specifically designed for use with Google Workspace models; applying it to other LLMs may yield unexpected results. Always test the sanitization logic against your specific compliance requirements before deploying in a production environment, as strict filtering might occasionally block benign but sensitive queries.