fix(entrypoint): arm cron worker via self-ping after boot

instrumentationHook is disabled in the standalone build, so node-cron only
arms on POST /api/worker/ping. Without a self-ping, a fresh container or
restart runs zero ingest passes (worker silently never armed) — exactly
what happened on the :8 deploy. Adds the same self-ping block as
technews/finance entrypoints (identical).

entrypoint.sh

@@ -24,4 +24,21 @@ if [ ! -f "$DB" ]; then
   echo "[entrypoint] initialized $DB from seed snapshot"
 fi
 
+# Arm the ingest/synthesis worker once the server answers.
+# The Next standalone build has instrumentationHook disabled, so node-cron
+# only starts when a request reaches /api/worker/ping. Pinging ourselves
+# after boot means a fresh container or restart re-arms the worker (and
+# triggers the boot pipeline pass) without any external health pinger.
+(
+  APP_IP="$(hostname -i 2>/dev/null | awk '{print $1}')"
+  [ -z "$APP_IP" ] && APP_IP=127.0.0.1
+  i=0
+  until curl -sf -o /dev/null --max-time 2 "http://${APP_IP}:${PORT:-3000}/api/worker/ping"; do
+    i=$((i + 1))
+    [ "$i" -ge 90 ] && exit 0
+    sleep 1
+  done
+  echo "[entrypoint] armed cron worker via /api/worker/ping"
+) >/dev/null 2>&1 &
+
 exec node server.js