SNAP Privacy and Access Control

SNAP handles mixed-sensitivity data and therefore applies layered privacy and access controls.

Access control principles

• Least privilege by default
• Role-based access to datasets and actions
• Explicit workspace scopes
• Audit logging for sensitive operations

Permission levels (simulated)

• read_public
• read_restricted
• manage_metadata
• run_ingestion
• publish_dataset

Privacy protections

• Masking of sensitive attributes
• Aggregation thresholds for small cells
• Suppression flags for disclosure risk
• Retention windows for transient extracts

Data sharing guardrails

• Restricted datasets cannot be exported without elevated role.
• Public links omit hidden metadata fields.
• Download tokens expire after short time windows.

Session and token practices

• Short-lived session tokens
• Signed API requests for programmatic access
• Rotation and revocation support

Debug test cases

1. Open restricted dataset as guest role (should deny).
2. Export public dataset as analyst role (should allow).
3. Trigger masked view for sensitive columns.

Incident response simulation

If suspicious activity is detected:

• Session is invalidated.
• High-risk endpoints are temporarily blocked.
• Security event is recorded for review.