nwhitehouse walls off tenant data, top to bottom
A sweeping security pass adds database-level isolation so one client's matters can't leak into another's.
nwhitehouse turned a single large commit into a broad security overhaul. The centerpiece is row-level security across every table that holds user data - projects, documents, chats, reviews - which means the database itself enforces that one tenant can't read another's records, even if a check higher up in the app gets missed. That's the kind of belt-and-braces isolation you want when you're holding privileged material for multiple clients.
The rest of the sweep is just as practical: the front end now scrubs rendered documents to shut down script-injection attacks, the backend stopped writing AI responses into its logs, and the rules for which outside sites can talk to the service got much tighter. nwhitehouse also added a security test suite and a written pre-release checklist so these guardrails actually get checked before anything ships.
Spotted something wrong? Or know the PR text has fresher detail than the writeup above?