优化: aichat效率批(save无变化捷径/断路器尾窗/system缓存指纹/会话列表刷新收敛) + 新增 LLM prompt caching(cache_control开关默认关/system稳定段易变段分离治缓存命中率) + 销账

src/composables/ai/useAiLifecycleEvents.ts

@@ -13,7 +13,6 @@ import { nextMsgId, convStates, clearApprovalTimer, flushCurrentText, clearAllTo
 import { clearTextIdleTimer, pendingMaxRounds, pendingHelp, pendingDirAuths } from './useAiPendingState'
 import { clearStreamWatchdog } from './useAiStream'
 import { setStreaming } from './streamingGuard'
-import { loadConversations } from './useAiConversations'
 import type { AiChatEvent, AiMessage, MessageId } from '@/api/types'
 
 const appSettings = useAppSettingsStore()
@@ -116,7 +115,8 @@ export function handleLifecycleEvent(event: AiChatEvent): boolean {
         const conv = state.conversations.find(c => c.id === state.activeConversationId)
         if (conv) conv.pinned_goals = event.pinned_goals
       }
-      void loadConversations()
+      // FE1(AC-EFF-F1-1/2/3):去掉显式 loadConversations(曾与下方 notify 自触发双拉),
+      // 列表刷新统一收敛到 ai-conversation-changed listener 的防抖调度,一次收尾只拉一次。
       // token 用量记录(开关开时):lastTokenUsage 供当前回复展示,convTokenTotal 累加对话总量
       if (isShowTokenUsage()) {
         state.lastTokenUsage = {
@@ -191,7 +191,7 @@ export function handleLifecycleEvent(event: AiChatEvent): boolean {
         options: event.options,
         conversationId: event.conversation_id || state.activeConversationId || null,
       }
-      void loadConversations()
+      // FE1:去掉显式 loadConversations(与下行 notify 双拉),由 listener 防抖调度统一收口。
       notifyConversationChanged()
       return true
     }