Audit my test and staging environments for real user data and replace it with synthetic data. 1. Find the leaks: check test fixtures and seed files for real names, emails, phone numbers, or content copied from production, staging databases that were populated from a production dump, screenshots or exports in the repo that contain user data, and any developer-local scripts that pull production data down. 2. Assess sensitivity: for anything found, note what it contains (emails only, or names, addresses, payment-adjacent data, private user content) so I know how careful deletion needs to be. 3. Replace with synthetic: build a proper fake data generator for fixtures and seeds (Faker or equivalent) producing realistic but clearly synthetic records. Use obviously-fake domains for emails (example.com or a test domain I own) so a misconfigured email job can never message a real person. 4. Staging policy: going forward, staging gets synthetic or anonymized data only. If I genuinely need production-shaped data in staging, write an anonymization script that scrubs emails, names, and free-text fields before import, and document that the anonymized copy is the only allowed path. 5. Clean history: if real data was committed to the repo, flag that it lives in git history and give me the steps to purge it (filter-repo or equivalent) plus credential rotation if any secrets rode along. Do not just delete rows and move on; if real user data sat in a shared staging environment, tell me who had access so I can judge exposure. Give me: what was found and where, the synthetic data generator, the anonymization script for staging, and the git history purge steps if needed.