Sloth-ninja puts a proper firewall between failures and client screens
A production incident exposed a familiar legal-tech risk: useful diagnostics vanished from logs while raw provider errors could still reach users.
Sloth-ninja has tightened how this fork handles database and integration failures across core areas such as login, chat, matters, documents, reviews, workflows and account settings.
The new approach keeps enough redacted diagnostic detail in server logs for the team to investigate real outages, including errors returned in an awkward format by Supabase, the hosted database platform. But browsers now receive a consistent, non-revealing message instead of internal database, provider or configuration details. The work also preserves specific feedback where users genuinely need it, such as validation and selected live chat errors, and adds tests for both sides of the boundary.
Spotted something wrong? Or know the PR text has fresher detail than the writeup above?