# Permission model workbook review brief

Use this brief to translate organizational policy into enforceable AI behavior. The review should separate what a user can know, what an agent can propose, and what the system can actually execute.

## Review sections

- Role definitions and environment boundaries.
- Retrieval sources and data classes by role.
- Tool access and sensitive action limits.
- Approval thresholds and exception paths.
- Audit visibility, denied-request handling, and release decision.