refactor(api): TOML 配置 SSOT、统一错误契约、Auth/事务加固与可观测性 (#33)

配置 SSOT（TOML + .env）
统一错误契约
Auth 与事务边界
Redis / Celery 可靠性:业务 Redis（DB/0）与 Celery broker/backend（DB/1）显式拆分；连接池、sync client
可观测性（OpenTelemetry + LGTM）

api/app/agents/chat/prompts_conversation.py

@@ -34,6 +34,7 @@ from app.agents.stage_constants import (
 )
 from app.agents.state_schema import KnownFact, PersonaThread
 from app.core.config import settings
+from app.features.conversation.constants import chat
 
 # 风格示例的单一事实源已迁至 `app.agents.style_profiles.ChatStyleProfile.reply_style_examples`；
 # 这里**不再**维护具体字面示例，避免同一模块被当作 few-shot 锚点反复注入，导致风格过拟合。
@@ -292,7 +293,7 @@ def get_opening_prompt(
 
     era_opening_line = ""
     if (
-        settings.chat_era_context_enabled
+        chat.era_context_enabled
         and profile_birth_year is not None
         and _compact_era_hint(
             current_stage,
@@ -450,7 +451,7 @@ def get_guided_conversation_prompt(
     )
 
     era_line = ""
-    if settings.chat_era_context_enabled:
+    if chat.era_context_enabled:
         era_line = _compact_era_hint(
             active_stage,
             birth_year=profile_birth_year,
@@ -696,7 +697,7 @@ _STAGE_TOPIC_CHIP_BANK: Dict[str, List[tuple[str, str]]] = {
         ("support", "家人之间的相互支持"),
         ("responsibility", "肩上的家庭责任"),
     ],
-    "later_life": [
+    "belief": [
         ("value", "现在最看重的事"),
         ("regret", "心里的遗憾"),
         ("pride", "最骄傲的事"),