优化: 对话交互修复与界面完善

- 修输入法组词回车直发/分离窗误关/出错后按钮死锁
- 新建任务失败弹窗内联提示,会话切换乐观高亮
- 状态多选联动刷新不再冲空可见行
- 测试连接保留已配开关与权重,防抖按提供商分键
- 代码块复制/长回复折叠带量/队列回填/排队时钟
- 子代理完成卡渲染,死链卡片与引用清理
- 截图渲染换 pro 版依赖,样式 token 化收尾

src/i18n/en/ai.ts

@@ -29,8 +29,14 @@ export default {
     notRetryable: 'This error cannot be resolved by retrying. Please check the configuration.',
     streamInterruptedAfterTool: '⚠ The tool finished, but the following reply was interrupted (no data stream for a long time). You can click Continue to retry.',
     streamInterrupted: '⚠ The response was interrupted (no data stream for a long time). This may be due to a network disconnection or a backend error. Please try again.',
+    // Light-convergence notice when the watchdog probe reports the backend already finished (only the closing event was lost)
+    streamFinishedConverged: 'The response was interrupted during finalization and has been converged.',
+    // First-time notice when the watchdog probe finds the backend loop alive/auto-resumed (only the first probe posts; subsequent re-arms stay silent)
+    streamAutoResuming: 'The response stalled and is auto-resuming (no action needed)...',
     // UX-2025-04 / CR-30-2 / decision a1: system notice bubble after mid-stream failure preserve (frontend mirrors backend session.messages)
     responseIncomplete: '⚠ The response is incomplete due to a network interruption. The above is the partial content received. You can resend to get a full reply.',
+    // Backend auto-finalized after reaching the max rounds (not a network interruption; reuses the incomplete bubble with distinct semantics)
+    responseIncompleteMaxRounds: 'Maximum rounds reached. The AI has finalized this round with the results so far; continue the conversation to have it refine and complete the answer.',
     queueFull: 'Send queue is full (max {limit} items)',
     forceSendConfirm: 'AI generation appears stuck. Force-stop and send anyway? (This may interrupt the current response)',
     forceSendBtn: 'Force Send',
@@ -73,5 +79,14 @@ export default {
       stepRunning: 'running {name}',
       stepPending: '⏳ {n} pending approval',
     },
+    // S1 sub-agent finished card (AiAgentFinished): lightweight terminal-state card at the
+    // end of the parent conversation message flow
+    agentFinished: {
+      title: 'Sub-agent',
+      completed: 'completed',
+      truncated: 'truncated',
+      failed: 'failed',
+      meta: '{rounds} rounds · {tokens} tokens',
+    },
   },
 }