Antispam: online checks no longer switch off Bayes; replies teach ham

- The "online reputation checks" rspamd profile used groups_enabled,
  which means ONLY those groups: with the switch on, a message got the
  RBL lookups alone (1 rule instead of 19, score 0.00) and Bayes never
  ran, so every spam lesson was ignored. It now disables nothing.
- A reply or reply-all teaches the answered message as ham (not
  forwards or new mail, not one's own Sent/Drafts, never over an
  explicit spam mark; once per message). message.sent now carries the
  draft's kind and reply_to_message_id.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

backend/app/outbox/jobs.py

@@ -338,6 +338,12 @@ async def send_job(ctx: dict[str, Any], outbox_id: str) -> None:
             # need to read the outbox row. payload to/cc are "Name <email>" strings.
             "to": _addr_dicts(draft.payload.get("to")),
             "cc": _addr_dicts(draft.payload.get("cc")),
+            # What it answers: "reply" / "reply_all" / "forward" / "new", and the
+            # message (a reply says the user wanted that mail — antispam learns it).
+            "kind": draft.kind,
+            "reply_to_message_id": (
+                str(draft.reply_to_message_id) if draft.reply_to_message_id else None
+            ),
         },
     )
     await bus.publish(