My app is multi-tenant: users belong to organizations/workspaces and must never see another organization's data. Audit my data access layer and enforce tenant isolation everywhere. 1. Inventory: list every table/collection that holds tenant-owned data and confirm it has a tenant/organization ID column. Add it where missing, including a foreign key to the organizations table. 2. Query audit: find every read, update, and delete query and confirm it filters by the current tenant's ID, derived from the authenticated session, never from a client-supplied organization ID without a membership check. Pay special attention to joins, aggregate queries, search endpoints, exports, and background jobs, which are the places scoping usually gets dropped. 3. Write paths: confirm every insert stamps the tenant ID server-side from the session. 4. Membership checks: anywhere a user can switch or specify an organization, verify the user is actually a member before the query runs. 5. Defense in depth: if I am on Postgres, consider row-level security policies as a backstop. Otherwise, centralize tenant scoping in the data access layer so new queries cannot forget it. Do not break legitimate cross-tenant admin tooling; gate it behind an explicit admin role instead of leaving queries unscoped. Do not trust any tenant identifier sent from the client without verifying membership. Give me a table of every endpoint/query, whether it was scoped, and what you changed. Include a test that logs in as a user from tenant A and tries to read tenant B's records by ID, and show it failing.