privacy-data-handling

Privacy & Data Handling

Reference

• ID: S-RISK-02
• Category: Risk
• Inputs: data inventory, regulatory requirements (GDPR, CCPA), system architecture
• Outputs: data handling policy document → artifacts/risk/
• Used by: Ops & Risk Agent
• Tool scripts: ./tools/artifact/validate.sh

Purpose

Establishes clear policies for data retention, access controls, and deletion procedures that comply with privacy regulations and protect user data throughout its lifecycle from collection to destruction.

Procedure

Step 0: Identify Data Processors

Read company.config.yaml and extract third-party data processors:

• analytics.provider (e.g., PostHog, Mixpanel, Plausible)
• email.provider (e.g., Resend, SendGrid, Postmark)
• observability.error_tracking (e.g., Sentry, Datadog)
• ai.llm_provider (e.g., Anthropic, OpenAI)
• support.platform (e.g., Intercom, Crisp)

For each identified processor, add a specific line to the DPA (Data Processing Agreement) checklist:

• [ ] DPA signed with [provider]
• [ ] Data residency confirmed (EU/US)
• [ ] Sub-processor list reviewed
• [ ] Deletion/export API documented

1. Inventory all personal and sensitive data collected: what data, where stored, why collected.
2. Classify data by sensitivity level: public, internal, confidential, restricted.
3. Define retention periods for each data category based on business need and regulatory requirements.
4. Design access control policies: who can access what data, under what conditions, with what audit trail.
5. Define the data deletion procedure: user-initiated deletion, retention-based expiry, right-to-be-forgotten flows.
6. Document data processing lawful bases (consent, legitimate interest, contractual necessity).
7. Define data breach notification procedures: detection, assessment, notification timeline.
8. Design audit logging for all data access and modifications.
9. Map data flows to third-party processors and define DPA requirements.
10. Save the data handling policy to artifacts/risk/.
11. Validate the artifact using ./tools/artifact/validate.sh.

Quality Checklist

• [ ] All personal data fields are inventoried
• [ ] Data classification levels are assigned
• [ ] Retention periods are defined and justified
• [ ] Access controls follow principle of least privilege
• [ ] Deletion procedures cover all storage locations (DB, backups, caches, logs)
• [ ] Breach notification timeline meets regulatory requirements
• [ ] Third-party data processors are documented with DPAs
• [ ] Artifact passes validation