fix(abuse): rename PG user/db sub2api->myapi + scrub upstream binary name from cmdline (paused by HF abuse-handler proc_scan)

fix(entrypoint): pick newer of latest.dump/latest.sql.gz by mtime — prevents stale dump from overriding fresh sql.gz on cold boot (root cause of 05-22 data revert)

fix(backup): also write latest.dump each cycle (-Fc -Z9) so on PGDATA wipe restore can pick fresh custom-format snapshot

entrypoint: detect pg_dump custom (.dump) vs plain SQL.gz

Update README + fix short_description length

LF endings

Add supervisord.conf

Add backup.sh

Add entrypoint.sh

Sidecar: PG18+Redis+supervisord+bucket dumps

Update Dockerfile

Update Dockerfile

Pre-seed .installed lock to skip AutoSetup; DB already migrated

Initial README

Initial Dockerfile

initial commit