# Example Risk Policy

This is a template for risk policies that constrain agent behavior.

## Risk Level Definitions
- **Low:** Reversible actions within normal scope
- **Medium:** Actions with limited blast radius, require logging
- **High:** Actions affecting production, users, or finances — require approval
- **Critical:** Actions that could cause data loss, security breach, or compliance violation — BLOCKED

## Permitted Actions
- List actions the agent may take autonomously

## Restricted Actions (Require Approval)
- List actions that require Team Lead or human approval

## Prohibited Actions
- List actions the agent must NEVER take

## Logging Requirements
- What must be logged for audit trail
- Retention period