# AI data loss prevention kit

Use this kit to prevent sensitive data leakage across prompts, retrieval, logs, model providers, tools, and exports. It helps teams translate data classifications into concrete AI controls rather than relying on generic privacy language.

## What it includes

- Data classes for public, internal, confidential, regulated, customer, employee, and secret material.
- A DLP control matrix for masking, retention, provider routing, log redaction, export review, and approval.
- A schema for data-boundary decisions and incident state.
- A review brief for open exposure risks and compensating controls.
- An escalation map for sensitive data categories.

## How to use it

Apply the data class map before prompts, indexes, logs, and tool outputs are connected. Block workflows that cannot enforce provider policy, retention, redaction, or export review for sensitive classes.