diff --git a/frontend/src/components/ai/AIBuiltinToolsCatalog.test.tsx b/frontend/src/components/ai/AIBuiltinToolsCatalog.test.tsx
index a754347..8ef364d 100644
--- a/frontend/src/components/ai/AIBuiltinToolsCatalog.test.tsx
+++ b/frontend/src/components/ai/AIBuiltinToolsCatalog.test.tsx
@@ -26,6 +26,8 @@ describe('AIBuiltinToolsCatalog', () => {
     expect(markup).toContain('inspect_table_bundle');
     expect(markup).toContain('全库快速摸底');
     expect(markup).toContain('inspect_database_bundle');
+    expect(markup).toContain('AI 应用健康总览');
+    expect(markup).toContain('inspect_app_health');
     expect(markup).toContain('一键体检 AI 配置');
     expect(markup).toContain('inspect_ai_setup_health');
     expect(markup).toContain('查看 AI 当前能力');
diff --git a/frontend/src/components/ai/AIBuiltinToolsCatalog.tsx b/frontend/src/components/ai/AIBuiltinToolsCatalog.tsx
index af30012..033fbf2 100644
--- a/frontend/src/components/ai/AIBuiltinToolsCatalog.tsx
+++ b/frontend/src/components/ai/AIBuiltinToolsCatalog.tsx
@@ -40,6 +40,11 @@ const BUILTIN_TOOL_FLOWS = [
     steps: 'inspect_database_bundle → inspect_table_bundle',
     description: '适合先看整库有哪些表、每张表大概有哪些字段，再对目标表继续做深挖快照。',
   },
+  {
+    title: 'AI 应用健康总览',
+    steps: 'inspect_app_health → inspect_ai_setup_health / inspect_app_logs / inspect_recent_connection_failures',
+    description: '适合用户反馈 AI 不稳定、连接和 MCP 问题交织、或需要先看整体健康状态时，一次汇总配置、日志、连接失败和工作区现场。',
+  },
   {
     title: '一键体检 AI 配置',
     steps: 'inspect_ai_setup_health → inspect_ai_providers / inspect_mcp_setup / inspect_ai_guidance',
diff --git a/frontend/src/components/ai/aiAppHealthInsights.test.ts b/frontend/src/components/ai/aiAppHealthInsights.test.ts
new file mode 100644
index 0000000..fb1b9e7
--- /dev/null
+++ b/frontend/src/components/ai/aiAppHealthInsights.test.ts
@@ -0,0 +1,132 @@
+import { describe, expect, it } from 'vitest';
+
+import { buildAIAppHealthSnapshot } from './aiAppHealthInsights';
+
+describe('buildAIAppHealthSnapshot', () => {
+  it('marks the app health as degraded when logs and connection failures show runtime problems', () => {
+    const snapshot = buildAIAppHealthSnapshot({
+      providers: [{
+        id: 'provider-1',
+        type: 'openai',
+        name: 'OpenAI 主账号',
+        apiKey: '',
+        hasSecret: true,
+        baseUrl: 'https://api.openai.com/v1',
+        model: 'gpt-5.4',
+        models: ['gpt-5.4'],
+        maxTokens: 32000,
+        temperature: 0.2,
+      }],
+      activeProviderId: 'provider-1',
+      safetyLevel: 'readonly',
+      contextLevel: 'schema_only',
+      builtinToolNames: ['inspect_app_health', 'inspect_ai_setup_health', 'inspect_app_logs'],
+      mcpServers: [{
+        id: 'server-1',
+        name: 'GoNavi MCP',
+        transport: 'stdio',
+        command: 'gonavi-mcp-server',
+        args: ['stdio'],
+        env: {},
+        enabled: true,
+        timeoutSeconds: 20,
+      }],
+      mcpClientStatuses: [],
+      mcpTools: [],
+      userPromptSettings: {
+        global: '',
+        database: '',
+        jvm: '',
+        jvmDiagnostic: '',
+      },
+      activeContext: {
+        connectionId: 'conn-1',
+        dbName: 'crm',
+      },
+      aiContexts: {
+        'conn-1:crm': [],
+      },
+      connections: [{
+        id: 'conn-1',
+        name: '主库',
+        config: {
+          type: 'mysql',
+          host: '127.0.0.1',
+          port: 3306,
+          user: 'root',
+        },
+      }],
+      tabs: [{
+        id: 'query-1',
+        title: '订单查询',
+        type: 'query',
+        connectionId: 'conn-1',
+        dbName: 'crm',
+        query: 'select * from orders',
+      }],
+      activeTabId: 'query-1',
+      appLogReadResult: {
+        success: true,
+        data: {
+          logPath: 'C:/Users/demo/.GoNavi/Logs/gonavi.log',
+          requestedLineLimit: 120,
+          lines: [
+            '2026/06/10 09:00:00.000000 [INFO] started',
+            '2026/06/10 09:00:01.000000 [ERROR] MCP server boot failed',
+          ],
+        },
+      },
+      connectionFailureReadResult: {
+        success: true,
+        data: {
+          logPath: 'C:/Users/demo/.GoNavi/Logs/gonavi.log',
+          requestedLineLimit: 120,
+          lines: [
+            '2026/06/10 09:01:00.000000 [ERROR] 建立数据库连接失败：类型=mysql 地址=127.0.0.1:3306 数据库=crm 用户=root；错误链：连接建立后验证失败：127.0.0.1:3306 验证失败: Error 1064 (42000): syntax error',
+            '2026/06/10 09:01:01.000000 [WARN] 命中数据库连接失败冷却：类型=mysql 地址=127.0.0.1:3306 数据库=crm 缓存Key=abc 剩余=29s 原因=连接建立后验证失败：127.0.0.1:3306 验证失败: Error 1064 (42000): syntax error',
+          ],
+        },
+      },
+    });
+
+    expect(snapshot.status).toBe('degraded');
+    expect(snapshot.summary.appLogErrorCount).toBe(1);
+    expect(snapshot.summary.recentConnectionFailureCount).toBe(2);
+    expect(snapshot.summary.activeTabTitle).toBe('订单查询');
+    expect(snapshot.warnings).toContain('最近应用日志里有 1 条 ERROR，需要优先查看 inspect_app_logs');
+    expect(snapshot.nextActions).toContain('调用 inspect_recent_connection_failures 查看最新连接失败根因，再决定是否检查当前连接或保存连接配置');
+    expect(snapshot.appLog.lines).toHaveLength(0);
+    expect(snapshot.appLog.linesOmitted).toBe(true);
+  });
+
+  it('marks missing provider pieces as blocked even when logs are clean', () => {
+    const snapshot = buildAIAppHealthSnapshot({
+      providers: [{
+        id: 'provider-1',
+        type: 'openai',
+        name: 'OpenAI 主账号',
+        apiKey: '',
+        hasSecret: false,
+        baseUrl: '',
+        model: '',
+        models: [],
+        maxTokens: 32000,
+        temperature: 0.2,
+      }],
+      activeProviderId: 'provider-1',
+      appLogReadResult: {
+        success: true,
+        data: { lines: ['2026/06/10 09:00:00.000000 [INFO] started'] },
+      },
+      connectionFailureReadResult: {
+        success: true,
+        data: { lines: [] },
+      },
+    });
+
+    expect(snapshot.status).toBe('blocked');
+    expect(snapshot.blockers).toContain('当前活动供应商缺少 API Key / Secret');
+    expect(snapshot.blockers).toContain('当前活动供应商缺少接口地址');
+    expect(snapshot.summary.chatReady).toBe(false);
+  });
+});
diff --git a/frontend/src/components/ai/aiAppHealthInsights.ts b/frontend/src/components/ai/aiAppHealthInsights.ts
new file mode 100644
index 0000000..2225960
--- /dev/null
+++ b/frontend/src/components/ai/aiAppHealthInsights.ts
@@ -0,0 +1,297 @@
+import type {
+  AIContextItem,
+  AIMCPClientInstallStatus,
+  AIMCPServerConfig,
+  AIMCPToolDescriptor,
+  AIProviderConfig,
+  AISafetyLevel,
+  AISkillConfig,
+  AIUserPromptSettings,
+  SavedConnection,
+  TabData,
+} from '../../types';
+import { buildAISetupHealthSnapshot } from './aiSetupHealthInsights';
+import { buildAppLogSnapshot } from './aiAppLogInsights';
+import { buildRecentConnectionFailureSnapshot } from './aiConnectionFailureInsights';
+import { buildActiveTabSnapshot, buildWorkspaceTabsSnapshot } from './aiWorkspaceInsights';
+
+type AIAppHealthStatus = 'ready' | 'needs_attention' | 'degraded' | 'blocked';
+
+const DEFAULT_APP_HEALTH_LOG_LIMIT = 120;
+const MAX_APP_HEALTH_LOG_LIMIT = 240;
+
+const appendUnique = (items: string[], value: string) => {
+  const trimmed = String(value || '').trim();
+  if (!trimmed || items.includes(trimmed)) {
+    return;
+  }
+  items.push(trimmed);
+};
+
+const normalizeAppHealthLogLimit = (value: unknown): number => {
+  const normalized = Math.floor(Number(value) || DEFAULT_APP_HEALTH_LOG_LIMIT);
+  if (normalized < 1) return 1;
+  if (normalized > MAX_APP_HEALTH_LOG_LIMIT) return MAX_APP_HEALTH_LOG_LIMIT;
+  return normalized;
+};
+
+const resolveActiveContextKey = (activeContext?: { connectionId?: string | null; dbName?: string | null } | null): string =>
+  activeContext?.connectionId ? `${activeContext.connectionId}:${activeContext.dbName || ''}` : 'default';
+
+const buildUnreadLogSnapshot = (message: string, lineLimit: number) => ({
+  readable: false,
+  logPath: '',
+  requestedLineLimit: lineLimit,
+  returnedLineCount: 0,
+  fileWindowTruncated: false,
+  matchedLinesTruncated: false,
+  levelBreakdown: {
+    INFO: 0,
+    WARN: 0,
+    ERROR: 0,
+    OTHER: 0,
+  },
+  hasWarnings: false,
+  hasErrors: false,
+  lines: [] as string[],
+  linesOmitted: false,
+  message,
+});
+
+const summarizeAppLogSnapshot = (
+  readResult: any,
+  options: {
+    keyword?: unknown;
+    lineLimit: number;
+    includeLogLines?: boolean;
+  },
+) => {
+  if (!readResult?.success) {
+    return buildUnreadLogSnapshot(
+      `GoNavi 应用日志暂不可读: ${readResult?.message || '当前环境未提供日志读取能力'}`,
+      options.lineLimit,
+    );
+  }
+
+  const snapshot = buildAppLogSnapshot({
+    readResult,
+    keyword: options.keyword,
+    lineLimit: options.lineLimit,
+  });
+  return {
+    readable: true,
+    ...snapshot,
+    lines: options.includeLogLines ? snapshot.lines : [],
+    linesOmitted: !options.includeLogLines && snapshot.lines.length > 0,
+  };
+};
+
+const summarizeConnectionFailures = (
+  readResult: any,
+  options: {
+    keyword?: unknown;
+    lineLimit: number;
+  },
+) => {
+  if (!readResult?.success) {
+    return {
+      readable: false,
+      logPath: '',
+      keyword: String(options.keyword || '').trim(),
+      requestedLineLimit: options.lineLimit,
+      returnedLineCount: 0,
+      fileWindowTruncated: false,
+      matchedLinesTruncated: false,
+      failureEventCount: 0,
+      hasRecentFailures: false,
+      primaryCategory: '',
+      primaryCategoryLabel: '',
+      cooldownHitCount: 0,
+      validationFailureCount: 0,
+      sshFailureCount: 0,
+      categorySummary: [],
+      addresses: [],
+      latestFailureAt: '',
+      latestFailure: null,
+      recentFailures: [],
+      nextActions: [] as string[],
+      message: `连接失败日志暂不可读: ${readResult?.message || '当前环境未提供日志读取能力'}`,
+    };
+  }
+
+  return {
+    readable: true,
+    ...buildRecentConnectionFailureSnapshot({
+      readResult,
+      keyword: options.keyword,
+      lineLimit: options.lineLimit,
+    }),
+  };
+};
+
+export const buildAIAppHealthSnapshot = (params: {
+  providers?: AIProviderConfig[];
+  activeProviderId?: string | null;
+  safetyLevel?: AISafetyLevel | string;
+  contextLevel?: string;
+  skills?: AISkillConfig[];
+  mcpServers?: AIMCPServerConfig[];
+  mcpClientStatuses?: AIMCPClientInstallStatus[];
+  mcpTools?: AIMCPToolDescriptor[];
+  dynamicModels?: string[];
+  builtinToolNames?: string[];
+  userPromptSettings?: AIUserPromptSettings;
+  activeContext?: { connectionId?: string | null; dbName?: string | null } | null;
+  aiContexts?: Record<string, AIContextItem[]>;
+  connections?: SavedConnection[];
+  tabs?: TabData[];
+  activeTabId?: string | null;
+  appLogReadResult?: any;
+  connectionFailureReadResult?: any;
+  keyword?: unknown;
+  connectionKeyword?: unknown;
+  lineLimit?: unknown;
+  includeLogLines?: boolean;
+}) => {
+  const connections = Array.isArray(params.connections) ? params.connections : [];
+  const tabs = Array.isArray(params.tabs) ? params.tabs : [];
+  const lineLimit = normalizeAppHealthLogLimit(params.lineLimit);
+  const activeContextKey = resolveActiveContextKey(params.activeContext);
+  const activeContextItems = params.aiContexts?.[activeContextKey] || [];
+  const setupHealth = buildAISetupHealthSnapshot({
+    providers: params.providers,
+    activeProviderId: params.activeProviderId,
+    safetyLevel: params.safetyLevel,
+    contextLevel: params.contextLevel,
+    skills: params.skills,
+    mcpServers: params.mcpServers,
+    mcpClientStatuses: params.mcpClientStatuses,
+    mcpTools: params.mcpTools,
+    dynamicModels: params.dynamicModels,
+    builtinToolNames: params.builtinToolNames,
+    userPromptSettings: params.userPromptSettings,
+    activeContext: params.activeContext,
+    activeContextItems,
+  });
+  const appLog = summarizeAppLogSnapshot(params.appLogReadResult, {
+    keyword: params.keyword,
+    lineLimit,
+    includeLogLines: params.includeLogLines === true,
+  });
+  const connectionFailures = summarizeConnectionFailures(params.connectionFailureReadResult, {
+    keyword: params.connectionKeyword ?? params.keyword,
+    lineLimit,
+  });
+  const workspace = buildWorkspaceTabsSnapshot({
+    tabs,
+    activeTabId: params.activeTabId,
+    connections,
+    includeContent: false,
+    limit: 8,
+  });
+  const activeTab = buildActiveTabSnapshot({
+    tabs,
+    activeTabId: params.activeTabId,
+    connections,
+    includeContent: false,
+  });
+  const activeTabTitle = activeTab.hasActiveTab && 'title' in activeTab ? activeTab.title : '';
+  const activeTabType = activeTab.hasActiveTab && 'type' in activeTab ? activeTab.type : '';
+
+  const blockers = [...setupHealth.blockers];
+  const warnings = [...setupHealth.warnings];
+  const nextActions = [...setupHealth.nextActions];
+
+  if (!appLog.readable) {
+    appendUnique(warnings, '当前无法读取 GoNavi 应用日志，启动异常和 MCP/连接错误缺少日志证据');
+    appendUnique(nextActions, '确认当前运行环境支持读取 gonavi.log 后，再调用 inspect_app_logs 下钻日志细节');
+  } else {
+    const errorCount = Number(appLog.levelBreakdown.ERROR) || 0;
+    const warnCount = Number(appLog.levelBreakdown.WARN) || 0;
+    if (errorCount > 0) {
+      appendUnique(warnings, `最近应用日志里有 ${errorCount} 条 ERROR，需要优先查看 inspect_app_logs`);
+      appendUnique(nextActions, '调用 inspect_app_logs 查看最近 ERROR/WARN 原文，确认是否影响 AI、MCP 或数据库连接');
+    } else if (warnCount > 0) {
+      appendUnique(warnings, `最近应用日志里有 ${warnCount} 条 WARN，建议确认是否为已知可忽略警告`);
+      appendUnique(nextActions, '如用户反馈不稳定，先调用 inspect_app_logs 查看 WARN 是否集中在 AI/MCP/连接链路');
+    }
+  }
+
+  if (!connectionFailures.readable) {
+    appendUnique(warnings, '当前无法读取连接失败日志，数据库连接冷却和验证失败缺少结构化证据');
+  } else if (connectionFailures.failureEventCount > 0) {
+    appendUnique(warnings, `最近识别到 ${connectionFailures.failureEventCount} 条连接失败/冷却记录`);
+    appendUnique(nextActions, '调用 inspect_recent_connection_failures 查看最新连接失败根因，再决定是否检查当前连接或保存连接配置');
+    connectionFailures.nextActions.forEach((action: string) => appendUnique(nextActions, action));
+  }
+
+  if (workspace.totalTabs === 0) {
+    appendUnique(warnings, '当前工作区没有打开任何页签，AI 缺少可直接读取的活动编辑器上下文');
+    appendUnique(nextActions, '如果要分析当前 SQL，先打开或选中目标 SQL 页签，再调用 inspect_active_tab');
+  }
+
+  const status: AIAppHealthStatus = blockers.length > 0
+    ? 'blocked'
+    : connectionFailures.failureEventCount > 0 || Number(appLog.levelBreakdown.ERROR) > 0
+      ? 'degraded'
+      : warnings.length > 0
+        ? 'needs_attention'
+        : 'ready';
+
+  const message = status === 'ready'
+    ? '当前 AI 应用健康总览通过，AI 配置、日志、连接失败和工作区上下文都没有明显异常'
+    : status === 'blocked'
+      ? `当前 AI 应用健康存在 ${blockers.length} 个阻塞项，优先修复供应商和发送前置条件`
+      : status === 'degraded'
+        ? '当前 AI 应用健康存在运行期异常信号，建议先下钻日志或连接失败记录'
+        : `当前 AI 应用健康整体可用，但还有 ${warnings.length} 个建议项`;
+
+  return {
+    status,
+    ready: status === 'ready',
+    message,
+    blockers,
+    warnings,
+    nextActions,
+    summary: {
+      aiSetupStatus: setupHealth.status,
+      chatReady: setupHealth.summary.chatReady,
+      hasActiveProvider: setupHealth.summary.hasActiveProvider,
+      activeProviderName: setupHealth.summary.activeProviderName,
+      safetyLevel: setupHealth.summary.safetyLevel,
+      contextLevel: setupHealth.summary.contextLevel,
+      providerCount: setupHealth.summary.providerCount,
+      enabledSkillCount: setupHealth.summary.enabledSkillCount,
+      customPromptCount: setupHealth.summary.customPromptCount,
+      mcpServerCount: setupHealth.summary.mcpServerCount,
+      enabledMCPServerCount: setupHealth.summary.enabledMCPServerCount,
+      discoveredMCPToolCount: setupHealth.summary.discoveredMCPToolCount,
+      totalAvailableToolCount: setupHealth.summary.totalAvailableToolCount,
+      connectionCount: connections.length,
+      activeContextConnectionId: params.activeContext?.connectionId || '',
+      activeContextDbName: params.activeContext?.dbName || '',
+      workspaceTabCount: workspace.totalTabs,
+      activeTabId: params.activeTabId || '',
+      activeTabTitle,
+      activeTabType,
+      appLogReadable: appLog.readable,
+      appLogErrorCount: Number(appLog.levelBreakdown.ERROR) || 0,
+      appLogWarnCount: Number(appLog.levelBreakdown.WARN) || 0,
+      recentConnectionFailureCount: connectionFailures.failureEventCount,
+      primaryConnectionFailureLabel: connectionFailures.primaryCategoryLabel,
+    },
+    aiSetup: {
+      status: setupHealth.status,
+      ready: setupHealth.ready,
+      message: setupHealth.message,
+      blockers: setupHealth.blockers,
+      warnings: setupHealth.warnings,
+      nextActions: setupHealth.nextActions,
+      summary: setupHealth.summary,
+    },
+    appLog,
+    connectionFailures,
+    workspace,
+    activeTab,
+  };
+};
diff --git a/frontend/src/components/ai/aiLocalToolExecutor.appHealthInspection.test.ts b/frontend/src/components/ai/aiLocalToolExecutor.appHealthInspection.test.ts
new file mode 100644
index 0000000..c448ab0
--- /dev/null
+++ b/frontend/src/components/ai/aiLocalToolExecutor.appHealthInspection.test.ts
@@ -0,0 +1,106 @@
+import { describe, expect, it, vi } from 'vitest';
+
+import type { AIToolCall, SavedConnection } from '../../types';
+import { executeLocalAIToolCall } from './aiLocalToolExecutor';
+
+const buildConnection = (): SavedConnection => ({
+  id: 'conn-1',
+  name: '主库',
+  config: {
+    type: 'mysql',
+    host: '127.0.0.1',
+    port: 3306,
+    user: 'root',
+  },
+});
+
+const buildToolCall = (name: string, args: Record<string, unknown>): AIToolCall => ({
+  id: `call-${name}`,
+  type: 'function',
+  function: {
+    name,
+    arguments: JSON.stringify(args),
+  },
+});
+
+describe('aiLocalToolExecutor inspect_app_health', () => {
+  it('returns an app-level health snapshot across ai setup, logs, connection failures, and workspace tabs', async () => {
+    const readAppLogTail = vi.fn()
+      .mockResolvedValueOnce({
+        success: true,
+        data: {
+          logPath: 'C:/Users/demo/.GoNavi/Logs/gonavi.log',
+          requestedLineLimit: 120,
+          lines: [
+            '2026/06/10 09:00:00.000000 [INFO] started',
+            '2026/06/10 09:00:01.000000 [ERROR] MCP server boot failed',
+          ],
+        },
+      })
+      .mockResolvedValueOnce({
+        success: true,
+        data: {
+          logPath: 'C:/Users/demo/.GoNavi/Logs/gonavi.log',
+          requestedLineLimit: 120,
+          lines: [
+            '2026/06/10 09:01:00.000000 [ERROR] 建立数据库连接失败：类型=mysql 地址=127.0.0.1:3306 数据库=crm 用户=root；错误链：连接建立后验证失败：127.0.0.1:3306 验证失败: Error 1064 (42000): syntax error',
+          ],
+        },
+      });
+
+    const result = await executeLocalAIToolCall({
+      toolCall: buildToolCall('inspect_app_health', {
+        lineLimit: 120,
+      }),
+      connections: [buildConnection()],
+      activeContext: {
+        connectionId: 'conn-1',
+        dbName: 'crm',
+      },
+      tabs: [{
+        id: 'query-1',
+        title: '订单查询',
+        type: 'query',
+        connectionId: 'conn-1',
+        dbName: 'crm',
+        query: 'select * from orders',
+      }],
+      activeTabId: 'query-1',
+      mcpTools: [],
+      toolContextMap: new Map(),
+      runtime: {
+        getDatabases: vi.fn(),
+        getTables: vi.fn(),
+        getAIRuntimeState: vi.fn().mockResolvedValue({
+          activeProviderId: 'provider-1',
+          providers: [{
+            id: 'provider-1',
+            type: 'openai',
+            name: 'OpenAI 主账号',
+            apiKey: '',
+            hasSecret: true,
+            baseUrl: 'https://api.openai.com/v1',
+            model: 'gpt-5.4',
+            models: ['gpt-5.4'],
+            maxTokens: 32000,
+            temperature: 0.2,
+          }],
+          safetyLevel: 'readonly',
+          contextLevel: 'schema_only',
+        }),
+        getMCPServers: vi.fn().mockResolvedValue([]),
+        getMCPClientInstallStatuses: vi.fn().mockResolvedValue([]),
+        readAppLogTail,
+      },
+    });
+
+    expect(result.success).toBe(true);
+    expect(result.content).toContain('"status":"degraded"');
+    expect(result.content).toContain('"activeProviderName":"OpenAI 主账号"');
+    expect(result.content).toContain('"appLogErrorCount":1');
+    expect(result.content).toContain('"recentConnectionFailureCount":1');
+    expect(result.content).toContain('"activeTabTitle":"订单查询"');
+    expect(result.content).toContain('inspect_recent_connection_failures');
+    expect(readAppLogTail).toHaveBeenCalledWith(120, '');
+  });
+});
diff --git a/frontend/src/components/ai/aiSnapshotInspectionAppHealthToolExecutor.ts b/frontend/src/components/ai/aiSnapshotInspectionAppHealthToolExecutor.ts
new file mode 100644
index 0000000..ab7d9d2
--- /dev/null
+++ b/frontend/src/components/ai/aiSnapshotInspectionAppHealthToolExecutor.ts
@@ -0,0 +1,136 @@
+import type {
+  AIContextItem,
+  AIMCPToolDescriptor,
+  AISkillConfig,
+  AIUserPromptSettings,
+  SavedConnection,
+  TabData,
+} from '../../types';
+import { BUILTIN_AI_TOOL_INFO } from '../../utils/aiToolRegistry';
+import { buildAIAppHealthSnapshot } from './aiAppHealthInsights';
+import type {
+  AISnapshotInspectionRuntime,
+  AISnapshotInspectionRuntimeState,
+  SnapshotInspectionResult,
+} from './aiSnapshotInspectionToolTypes';
+
+const BUILTIN_AI_TOOL_NAMES = BUILTIN_AI_TOOL_INFO.map((item) => item.name);
+const DEFAULT_APP_HEALTH_LOG_LIMIT = 120;
+const MAX_APP_HEALTH_LOG_LIMIT = 240;
+
+interface ExecuteAppHealthSnapshotToolCallOptions {
+  toolName: string;
+  args: Record<string, any>;
+  activeContext?: { connectionId: string; dbName: string } | null;
+  aiContexts?: Record<string, AIContextItem[]>;
+  connections: SavedConnection[];
+  tabs?: TabData[];
+  activeTabId?: string | null;
+  mcpTools: AIMCPToolDescriptor[];
+  skills?: AISkillConfig[];
+  userPromptSettings?: AIUserPromptSettings;
+  dynamicModels?: string[];
+  runtime?: AISnapshotInspectionRuntime;
+}
+
+const loadRuntimeState = async (
+  runtime: AISnapshotInspectionRuntime | undefined,
+): Promise<AISnapshotInspectionRuntimeState | undefined> =>
+  typeof runtime?.getAIRuntimeState === 'function'
+    ? runtime.getAIRuntimeState()
+    : undefined;
+
+const loadMCPSetupState = async (runtime: AISnapshotInspectionRuntime | undefined) =>
+  Promise.all([
+    typeof runtime?.getMCPServers === 'function' ? runtime.getMCPServers() : Promise.resolve(undefined),
+    typeof runtime?.getMCPClientInstallStatuses === 'function'
+      ? runtime.getMCPClientInstallStatuses()
+      : Promise.resolve(undefined),
+  ]);
+
+const readLogTail = async (
+  runtime: AISnapshotInspectionRuntime | undefined,
+  lineLimit: number,
+  keyword: string,
+) => {
+  if (typeof runtime?.readAppLogTail !== 'function') {
+    return { success: false, message: '当前环境暂不支持读取 GoNavi 应用日志' };
+  }
+  return runtime.readAppLogTail(lineLimit, keyword);
+};
+
+const normalizeLineLimit = (value: unknown): number => {
+  const normalized = Math.floor(Number(value) || DEFAULT_APP_HEALTH_LOG_LIMIT);
+  if (normalized < 1) return 1;
+  if (normalized > MAX_APP_HEALTH_LOG_LIMIT) return MAX_APP_HEALTH_LOG_LIMIT;
+  return normalized;
+};
+
+export async function executeAppHealthSnapshotToolCall(
+  options: ExecuteAppHealthSnapshotToolCallOptions,
+): Promise<SnapshotInspectionResult | null> {
+  const {
+    toolName,
+    args,
+    activeContext = null,
+    aiContexts = {},
+    connections,
+    tabs = [],
+    activeTabId = null,
+    mcpTools,
+    skills = [],
+    userPromptSettings,
+    dynamicModels = [],
+    runtime,
+  } = options;
+
+  if (toolName !== 'inspect_app_health') {
+    return null;
+  }
+
+  try {
+    const lineLimit = normalizeLineLimit(args.lineLimit);
+    const keyword = String(args.keyword || '').trim();
+    const connectionKeyword = String(args.connectionKeyword ?? args.keyword ?? '').trim();
+    const [runtimeState, mcpState, appLogReadResult, connectionFailureReadResult] = await Promise.all([
+      loadRuntimeState(runtime),
+      loadMCPSetupState(runtime),
+      readLogTail(runtime, lineLimit, keyword),
+      readLogTail(runtime, lineLimit, connectionKeyword),
+    ]);
+    const [mcpServers, mcpClientInstallStatuses] = mcpState;
+
+    return {
+      content: JSON.stringify(buildAIAppHealthSnapshot({
+        providers: Array.isArray(runtimeState?.providers) ? runtimeState.providers : [],
+        activeProviderId: runtimeState?.activeProviderId || '',
+        safetyLevel: runtimeState?.safetyLevel,
+        contextLevel: runtimeState?.contextLevel,
+        skills,
+        mcpServers: Array.isArray(mcpServers) ? mcpServers : [],
+        mcpClientStatuses: Array.isArray(mcpClientInstallStatuses) ? mcpClientInstallStatuses : [],
+        mcpTools,
+        dynamicModels,
+        builtinToolNames: BUILTIN_AI_TOOL_NAMES,
+        userPromptSettings,
+        activeContext,
+        aiContexts,
+        connections,
+        tabs,
+        activeTabId,
+        appLogReadResult,
+        connectionFailureReadResult,
+        keyword,
+        connectionKeyword,
+        lineLimit,
+        includeLogLines: args.includeLogLines === true,
+      })),
+      success: true,
+    };
+  } catch (error: any) {
+    return {
+      content: `读取 AI 应用健康总览失败: ${error?.message || error}`,
+      success: false,
+    };
+  }
+}
diff --git a/frontend/src/components/ai/aiSnapshotInspectionToolExecutor.ts b/frontend/src/components/ai/aiSnapshotInspectionToolExecutor.ts
index b7bedd0..66f2fa1 100644
--- a/frontend/src/components/ai/aiSnapshotInspectionToolExecutor.ts
+++ b/frontend/src/components/ai/aiSnapshotInspectionToolExecutor.ts
@@ -36,6 +36,7 @@ import { buildShortcutSnapshot } from './aiShortcutInsights';
 import { buildAILastRenderErrorSnapshot } from './aiLastRenderErrorInsights';
 import { buildRecentConnectionFailureSnapshot } from './aiConnectionFailureInsights';
 import { executeAIConfigSnapshotToolCall } from './aiSnapshotInspectionAIConfigToolExecutor';
+import { executeAppHealthSnapshotToolCall } from './aiSnapshotInspectionAppHealthToolExecutor';
 import type {
   AISnapshotInspectionRuntime,
   SnapshotInspectionResult,
@@ -90,6 +91,24 @@ export async function executeSnapshotInspectionToolCall(
   } = options;
 
   try {
+    const appHealthResult = await executeAppHealthSnapshotToolCall({
+      toolName,
+      args,
+      activeContext,
+      aiContexts,
+      connections,
+      tabs,
+      activeTabId,
+      mcpTools,
+      skills,
+      userPromptSettings,
+      dynamicModels,
+      runtime,
+    });
+    if (appHealthResult) {
+      return appHealthResult;
+    }
+
     const aiConfigResult = await executeAIConfigSnapshotToolCall({
       toolName,
       activeContext,
@@ -374,6 +393,7 @@ export async function executeSnapshotInspectionToolCall(
       inspect_saved_queries: '读取已保存查询失败',
       inspect_sql_snippets: '读取 SQL 片段失败',
       inspect_shortcuts: '读取快捷键配置失败',
+      inspect_app_health: '读取 AI 应用健康总览失败',
     }[toolName] || '读取本地探针快照失败';
     return {
       content: `${label}: ${error?.message || error}`,
diff --git a/frontend/src/components/ai/aiSystemContextMessages.test.ts b/frontend/src/components/ai/aiSystemContextMessages.test.ts
index e0cc8cb..3c14bd3 100644
--- a/frontend/src/components/ai/aiSystemContextMessages.test.ts
+++ b/frontend/src/components/ai/aiSystemContextMessages.test.ts
@@ -68,13 +68,14 @@ describe('buildAISystemContextMessages', () => {
       connections: [connections[0]],
       tabs: [],
       activeTabId: null,
-      availableToolNames: ['inspect_workspace_tabs', 'inspect_ai_setup_health', 'inspect_ai_runtime', 'inspect_ai_safety', 'inspect_ai_providers', 'inspect_ai_chat_readiness', 'inspect_mcp_setup', 'inspect_mcp_authoring_guide', 'inspect_ai_guidance', 'inspect_ai_context', 'inspect_current_connection', 'inspect_connection_capabilities', 'inspect_saved_connections', 'inspect_external_sql_directories', 'inspect_external_sql_file', 'inspect_recent_sql_activity', 'inspect_sql_risk', 'inspect_recent_connection_failures', 'inspect_app_logs', 'inspect_ai_last_render_error', 'inspect_saved_queries', 'inspect_ai_sessions', 'inspect_sql_snippets', 'inspect_shortcuts', 'get_columns'],
+      availableToolNames: ['inspect_workspace_tabs', 'inspect_app_health', 'inspect_ai_setup_health', 'inspect_ai_runtime', 'inspect_ai_safety', 'inspect_ai_providers', 'inspect_ai_chat_readiness', 'inspect_mcp_setup', 'inspect_mcp_authoring_guide', 'inspect_ai_guidance', 'inspect_ai_context', 'inspect_current_connection', 'inspect_connection_capabilities', 'inspect_saved_connections', 'inspect_external_sql_directories', 'inspect_external_sql_file', 'inspect_recent_sql_activity', 'inspect_sql_risk', 'inspect_recent_connection_failures', 'inspect_app_logs', 'inspect_ai_last_render_error', 'inspect_saved_queries', 'inspect_ai_sessions', 'inspect_sql_snippets', 'inspect_shortcuts', 'get_columns'],
       skills,
       userPromptSettings,
     });
 
     const joined = messages.map((message) => message.content).join('\n');
     expect(joined).toContain('inspect_workspace_tabs 盘点当前工作区');
+    expect(joined).toContain('inspect_app_health 获取 AI 配置、应用日志、连接失败和工作区页签的全局健康总览');
     expect(joined).toContain('inspect_ai_setup_health 先拿到整体现状');
     expect(joined).toContain('inspect_ai_runtime 读取当前 AI 运行状态');
     expect(joined).toContain('inspect_ai_safety 读取真实安全边界');
diff --git a/frontend/src/components/ai/aiSystemInspectionGuidance.ts b/frontend/src/components/ai/aiSystemInspectionGuidance.ts
index ceacb49..353b03b 100644
--- a/frontend/src/components/ai/aiSystemInspectionGuidance.ts
+++ b/frontend/src/components/ai/aiSystemInspectionGuidance.ts
@@ -55,6 +55,12 @@ export const appendDatabaseInspectionGuidanceMessages = (
     '如果用户提到“当前 AI 上下文”“当前关联了哪些表”“现在带了哪些表结构”，优先调用 inspect_ai_context 读取当前挂载的表结构上下文，不要凭记忆复述。',
   );
   appendAIRuntimeInspectionGuidance(messages, availableToolNames);
+  appendGuidanceIfToolAvailable(
+    messages,
+    availableToolNames,
+    'inspect_app_health',
+    '如果用户提到“AI 不稳定”“整体帮我看看”“GoNavi AI 现在还有哪些明显问题”“连接、MCP、日志一起排查”，优先调用 inspect_app_health 获取 AI 配置、应用日志、连接失败和工作区页签的全局健康总览，再决定下钻 inspect_ai_setup_health、inspect_app_logs 或 inspect_recent_connection_failures。',
+  );
   appendGuidanceIfToolAvailable(
     messages,
     availableToolNames,
diff --git a/frontend/src/utils/aiBuiltinInspectionToolInfo.ts b/frontend/src/utils/aiBuiltinInspectionToolInfo.ts
index 063579d..983f41d 100644
--- a/frontend/src/utils/aiBuiltinInspectionToolInfo.ts
+++ b/frontend/src/utils/aiBuiltinInspectionToolInfo.ts
@@ -1,6 +1,31 @@
 import type { AIBuiltinToolInfo } from "./aiBuiltinToolInfo.types";
 
 export const BUILTIN_AI_INSPECTION_TOOL_INFO: AIBuiltinToolInfo[] = [
+  {
+    name: "inspect_app_health",
+    icon: "🧭",
+    desc: "一键查看 AI 应用健康总览",
+    detail:
+      "汇总 AI 配置、供应商发送前置、MCP 接入、应用日志 ERROR/WARN、最近连接失败/冷却和当前工作区页签，给出阻塞项、运行期异常信号和下一步探针建议。适合用户说“AI 不稳定”“整体帮我看看”“连接和 MCP 一起排查”时先做一次全局摸底。",
+    params: "keyword?, connectionKeyword?, lineLimit?(默认 120), includeLogLines?(默认 false)",
+    tool: {
+      type: "function",
+      function: {
+        name: "inspect_app_health",
+        description:
+          "读取 GoNavi AI 应用健康总览，汇总 AI 供应商与发送前置、MCP 接入、应用日志 ERROR/WARN、最近连接失败/冷却和当前工作区页签，并返回阻塞项、运行期异常信号与下一步探针建议。适用于用户提到 AI 不稳定、整体不成熟、连接/MCP/日志需要一起排查或要求先看全局状态时，优先调用该工具。",
+        parameters: {
+          type: "object",
+          properties: {
+            keyword: { type: "string", description: "可选，读取应用日志时按关键词过滤，例如 ai、mcp、mysql、error；不传则读取最近日志窗口" },
+            connectionKeyword: { type: "string", description: "可选，分析连接失败日志时按连接类型、地址或错误关键词过滤；不传时复用 keyword" },
+            lineLimit: { type: "number", description: "可选，每次最多分析多少行日志，默认 120，最大 240" },
+            includeLogLines: { type: "boolean", description: "可选，是否在结果里附带日志原文行，默认 false；需要引用原文时再开启" },
+          },
+        },
+      },
+    },
+  },
   {
     name: "inspect_ai_setup_health",
     icon: "🩺",