@midscene/web 0.10.1 → 0.10.2

package/dist/lib/index.js

@@ -564,7 +564,10 @@ var PageTaskExecutor = class {
           locate: plan2.locate,
           executor: async (param, taskContext) => {
             const { task } = taskContext;
-            (0, import_node_assert.default)((param == null ? void 0 : param.prompt) || (param == null ? void 0 : param.id), "No prompt or id to locate");
+            (0, import_node_assert.default)(
+              (param == null ? void 0 : param.prompt) || (param == null ? void 0 : param.id) || (param == null ? void 0 : param.position),
+              "No prompt or id or position to locate"
+            );
             let insightDump;
             const dumpCollector = (dump2) => {
               insightDump = dump2;
@@ -719,6 +722,22 @@ var PageTaskExecutor = class {
           }
         };
         tasks.push(taskActionTap);
+      } else if (plan2.type === "Drag") {
+        const taskActionDrag = {
+          type: "Action",
+          subType: "Drag",
+          param: plan2.param,
+          thought: plan2.thought,
+          locate: plan2.locate,
+          executor: async (taskParam) => {
+            (0, import_node_assert.default)(
+              (taskParam == null ? void 0 : taskParam.start_box) && (taskParam == null ? void 0 : taskParam.end_box),
+              "No start_box or end_box to drag"
+            );
+            await this.page.mouse.drag(taskParam.start_box, taskParam.end_box);
+          }
+        };
+        tasks.push(taskActionDrag);
       } else if (plan2.type === "Hover") {
         const taskActionHover = {
           type: "Action",
@@ -753,9 +772,9 @@ var PageTaskExecutor = class {
             } else if (scrollToEventName === "untilLeft") {
               await this.page.scrollUntilLeft(startingPoint);
             } else if (scrollToEventName === "once" || !scrollToEventName) {
-              if (taskParam.direction === "down" || !taskParam.direction) {
+              if ((taskParam == null ? void 0 : taskParam.direction) === "down" || !taskParam || !taskParam.direction) {
                 await this.page.scrollDown(
-                  taskParam.distance || void 0,
+                  (taskParam == null ? void 0 : taskParam.distance) || void 0,
                   startingPoint
                 );
               } else if (taskParam.direction === "up") {
@@ -950,6 +969,7 @@ var PageTaskExecutor = class {
         userPrompt
       },
       executor: async (param, executorContext) => {
+        var _a;
         const shotTime = Date.now();
         const pageContext = await this.insight.contextRetrieverFn("locate");
         const recordItem = {
@@ -986,7 +1006,7 @@ var PageTaskExecutor = class {
         return {
           output: {
             actions,
-            thought: actions[0].thought,
+            thought: (_a = actions[0]) == null ? void 0 : _a.thought,
             actionType: actions[0].type,
             taskWillBeAccomplished: false,
             furtherPlan: {
@@ -1819,15 +1839,49 @@ var Page = class {
           );
         }
       },
-      move: async (x, y) => this.underlyingPage.mouse.move(x, y)
+      move: async (x, y) => this.underlyingPage.mouse.move(x, y),
+      drag: async (from, to) => {
+        if (this.pageType === "puppeteer") {
+          await this.underlyingPage.mouse.drag(
+            {
+              x: from.x,
+              y: from.y
+            },
+            {
+              x: to.x,
+              y: to.y
+            }
+          );
+        } else if (this.pageType === "playwright") {
+          await this.underlyingPage.mouse.move(
+            from.x,
+            from.y
+          );
+          await this.underlyingPage.mouse.down();
+          await this.underlyingPage.mouse.move(to.x, to.y);
+          await this.underlyingPage.mouse.up();
+        }
+      }
     };
   }
   get keyboard() {
     return {
       type: async (text) => this.underlyingPage.keyboard.type(text, { delay: 80 }),
-      press: async (key) => this.underlyingPage.keyboard.press(key),
-      down: async (key) => this.underlyingPage.keyboard.down(key),
-      up: async (key) => this.underlyingPage.keyboard.up(key)
+      press: async (key) => {
+        const keys = Array.isArray(key) ? key : [key];
+        for (const key2 of keys) {
+          await this.underlyingPage.keyboard.down(key2);
+        }
+        for (const key2 of [...keys].reverse()) {
+          await this.underlyingPage.keyboard.up(key2);
+        }
+      },
+      down: async (key) => {
+        this.underlyingPage.keyboard.down(key);
+      },
+      up: async (key) => {
+        this.underlyingPage.keyboard.up(key);
+      }
     };
   }
   async clearInput(element) {
@@ -1946,6 +2000,23 @@ var PuppeteerAgent = class extends PageAgent {
   constructor(page, opts) {
     const webPage = new WebPage2(page);
     super(webPage, opts);
+    if (opts == null ? void 0 : opts.trackingActiveTab) {
+      const browser = this.page.underlyingPage.browser();
+      browser.on("targetcreated", async (target) => {
+        if (target.type() === "page") {
+          const targetPage = await target.page();
+          if (!targetPage) {
+            console.warn(
+              "got a targetPage event, but the page is not ready yet, skip"
+            );
+            return;
+          }
+          const midscenePage = new WebPage2(targetPage);
+          this.page = midscenePage;
+          this.taskExecutor.page = midscenePage;
+        }
+      });
+    }
   }
 };
 
@@ -2179,7 +2250,8 @@ var Page2 = class {
     return {
       click: (x, y, options) => this.mouseClick(x, y, (options == null ? void 0 : options.button) || "left"),
       wheel: (deltaX, deltaY) => this.mouseWheel(deltaX, deltaY),
-      move: (x, y) => this.mouseMove(x, y)
+      move: (x, y) => this.mouseMove(x, y),
+      drag: (from, to) => this.mouseDrag(from, to)
     };
   }
   // Object that includes keyboard and mouse operations
@@ -2326,6 +2398,21 @@ var Page2 = class {
       }
     ]);
   }
+  async mouseDrag(from, to) {
+    await this.browser.performActions([
+      {
+        type: "pointer",
+        id: "mouse",
+        parameters: { pointerType: "mouse" },
+        actions: [
+          { type: "pointerMove", duration: 0, x: from.x, y: from.y },
+          { type: "pointerDown", button: 0 },
+          { type: "pointerMove", duration: 500, x: to.x, y: to.y },
+          { type: "pointerUp", button: 0 }
+        ]
+      }
+    ]);
+  }
   async mouseWheel(deltaX, deltaY, duration = 1e3) {
     const n = 4;
     const windowSize = await this.browser.getWindowSize();

package/dist/lib/midscene-playground.js

@@ -1012,7 +1012,10 @@ var PageTaskExecutor = class {
           locate: plan2.locate,
           executor: async (param, taskContext) => {
             const { task } = taskContext;
-            (0, import_node_assert2.default)((param == null ? void 0 : param.prompt) || (param == null ? void 0 : param.id), "No prompt or id to locate");
+            (0, import_node_assert2.default)(
+              (param == null ? void 0 : param.prompt) || (param == null ? void 0 : param.id) || (param == null ? void 0 : param.position),
+              "No prompt or id or position to locate"
+            );
             let insightDump;
             const dumpCollector = (dump) => {
               insightDump = dump;
@@ -1167,6 +1170,22 @@ var PageTaskExecutor = class {
           }
         };
         tasks.push(taskActionTap);
+      } else if (plan2.type === "Drag") {
+        const taskActionDrag = {
+          type: "Action",
+          subType: "Drag",
+          param: plan2.param,
+          thought: plan2.thought,
+          locate: plan2.locate,
+          executor: async (taskParam) => {
+            (0, import_node_assert2.default)(
+              (taskParam == null ? void 0 : taskParam.start_box) && (taskParam == null ? void 0 : taskParam.end_box),
+              "No start_box or end_box to drag"
+            );
+            await this.page.mouse.drag(taskParam.start_box, taskParam.end_box);
+          }
+        };
+        tasks.push(taskActionDrag);
       } else if (plan2.type === "Hover") {
         const taskActionHover = {
           type: "Action",
@@ -1201,9 +1220,9 @@ var PageTaskExecutor = class {
             } else if (scrollToEventName === "untilLeft") {
               await this.page.scrollUntilLeft(startingPoint);
             } else if (scrollToEventName === "once" || !scrollToEventName) {
-              if (taskParam.direction === "down" || !taskParam.direction) {
+              if ((taskParam == null ? void 0 : taskParam.direction) === "down" || !taskParam || !taskParam.direction) {
                 await this.page.scrollDown(
-                  taskParam.distance || void 0,
+                  (taskParam == null ? void 0 : taskParam.distance) || void 0,
                   startingPoint
                 );
               } else if (taskParam.direction === "up") {
@@ -1398,6 +1417,7 @@ var PageTaskExecutor = class {
         userPrompt
       },
       executor: async (param, executorContext) => {
+        var _a;
         const shotTime = Date.now();
         const pageContext = await this.insight.contextRetrieverFn("locate");
         const recordItem = {
@@ -1434,7 +1454,7 @@ var PageTaskExecutor = class {
         return {
           output: {
             actions,
-            thought: actions[0].thought,
+            thought: (_a = actions[0]) == null ? void 0 : _a.thought,
             actionType: actions[0].type,
             taskWillBeAccomplished: false,
             furtherPlan: {
@@ -1946,7 +1966,8 @@ var StaticPage = class {
     this.mouse = {
       click: ThrowNotImplemented.bind(null, "mouse.click"),
       wheel: ThrowNotImplemented.bind(null, "mouse.wheel"),
-      move: ThrowNotImplemented.bind(null, "mouse.move")
+      move: ThrowNotImplemented.bind(null, "mouse.move"),
+      drag: ThrowNotImplemented.bind(null, "mouse.drag")
     };
     this.keyboard = {
       type: ThrowNotImplemented.bind(null, "keyboard.type"),

package/dist/lib/playground.js

@@ -675,7 +675,10 @@ var PageTaskExecutor = class {
           locate: plan2.locate,
           executor: async (param, taskContext) => {
             const { task } = taskContext;
-            (0, import_node_assert2.default)((param == null ? void 0 : param.prompt) || (param == null ? void 0 : param.id), "No prompt or id to locate");
+            (0, import_node_assert2.default)(
+              (param == null ? void 0 : param.prompt) || (param == null ? void 0 : param.id) || (param == null ? void 0 : param.position),
+              "No prompt or id or position to locate"
+            );
             let insightDump;
             const dumpCollector = (dump) => {
               insightDump = dump;
@@ -830,6 +833,22 @@ var PageTaskExecutor = class {
           }
         };
         tasks.push(taskActionTap);
+      } else if (plan2.type === "Drag") {
+        const taskActionDrag = {
+          type: "Action",
+          subType: "Drag",
+          param: plan2.param,
+          thought: plan2.thought,
+          locate: plan2.locate,
+          executor: async (taskParam) => {
+            (0, import_node_assert2.default)(
+              (taskParam == null ? void 0 : taskParam.start_box) && (taskParam == null ? void 0 : taskParam.end_box),
+              "No start_box or end_box to drag"
+            );
+            await this.page.mouse.drag(taskParam.start_box, taskParam.end_box);
+          }
+        };
+        tasks.push(taskActionDrag);
       } else if (plan2.type === "Hover") {
         const taskActionHover = {
           type: "Action",
@@ -864,9 +883,9 @@ var PageTaskExecutor = class {
             } else if (scrollToEventName === "untilLeft") {
               await this.page.scrollUntilLeft(startingPoint);
             } else if (scrollToEventName === "once" || !scrollToEventName) {
-              if (taskParam.direction === "down" || !taskParam.direction) {
+              if ((taskParam == null ? void 0 : taskParam.direction) === "down" || !taskParam || !taskParam.direction) {
                 await this.page.scrollDown(
-                  taskParam.distance || void 0,
+                  (taskParam == null ? void 0 : taskParam.distance) || void 0,
                   startingPoint
                 );
               } else if (taskParam.direction === "up") {
@@ -1061,6 +1080,7 @@ var PageTaskExecutor = class {
         userPrompt
       },
       executor: async (param, executorContext) => {
+        var _a;
         const shotTime = Date.now();
         const pageContext = await this.insight.contextRetrieverFn("locate");
         const recordItem = {
@@ -1097,7 +1117,7 @@ var PageTaskExecutor = class {
         return {
           output: {
             actions,
-            thought: actions[0].thought,
+            thought: (_a = actions[0]) == null ? void 0 : _a.thought,
             actionType: actions[0].type,
             taskWillBeAccomplished: false,
             furtherPlan: {
@@ -1609,7 +1629,8 @@ var StaticPage = class {
     this.mouse = {
       click: ThrowNotImplemented.bind(null, "mouse.click"),
       wheel: ThrowNotImplemented.bind(null, "mouse.wheel"),
-      move: ThrowNotImplemented.bind(null, "mouse.move")
+      move: ThrowNotImplemented.bind(null, "mouse.move"),
+      drag: ThrowNotImplemented.bind(null, "mouse.drag")
     };
     this.keyboard = {
       type: ThrowNotImplemented.bind(null, "keyboard.type"),

package/dist/lib/playwright.js

@@ -558,7 +558,10 @@ var PageTaskExecutor = class {
           locate: plan2.locate,
           executor: async (param, taskContext) => {
             const { task } = taskContext;
-            (0, import_node_assert.default)((param == null ? void 0 : param.prompt) || (param == null ? void 0 : param.id), "No prompt or id to locate");
+            (0, import_node_assert.default)(
+              (param == null ? void 0 : param.prompt) || (param == null ? void 0 : param.id) || (param == null ? void 0 : param.position),
+              "No prompt or id or position to locate"
+            );
             let insightDump;
             const dumpCollector = (dump) => {
               insightDump = dump;
@@ -713,6 +716,22 @@ var PageTaskExecutor = class {
           }
         };
         tasks.push(taskActionTap);
+      } else if (plan2.type === "Drag") {
+        const taskActionDrag = {
+          type: "Action",
+          subType: "Drag",
+          param: plan2.param,
+          thought: plan2.thought,
+          locate: plan2.locate,
+          executor: async (taskParam) => {
+            (0, import_node_assert.default)(
+              (taskParam == null ? void 0 : taskParam.start_box) && (taskParam == null ? void 0 : taskParam.end_box),
+              "No start_box or end_box to drag"
+            );
+            await this.page.mouse.drag(taskParam.start_box, taskParam.end_box);
+          }
+        };
+        tasks.push(taskActionDrag);
       } else if (plan2.type === "Hover") {
         const taskActionHover = {
           type: "Action",
@@ -747,9 +766,9 @@ var PageTaskExecutor = class {
             } else if (scrollToEventName === "untilLeft") {
               await this.page.scrollUntilLeft(startingPoint);
             } else if (scrollToEventName === "once" || !scrollToEventName) {
-              if (taskParam.direction === "down" || !taskParam.direction) {
+              if ((taskParam == null ? void 0 : taskParam.direction) === "down" || !taskParam || !taskParam.direction) {
                 await this.page.scrollDown(
-                  taskParam.distance || void 0,
+                  (taskParam == null ? void 0 : taskParam.distance) || void 0,
                   startingPoint
                 );
               } else if (taskParam.direction === "up") {
@@ -944,6 +963,7 @@ var PageTaskExecutor = class {
         userPrompt
       },
       executor: async (param, executorContext) => {
+        var _a;
         const shotTime = Date.now();
         const pageContext = await this.insight.contextRetrieverFn("locate");
         const recordItem = {
@@ -980,7 +1000,7 @@ var PageTaskExecutor = class {
         return {
           output: {
             actions,
-            thought: actions[0].thought,
+            thought: (_a = actions[0]) == null ? void 0 : _a.thought,
             actionType: actions[0].type,
             taskWillBeAccomplished: false,
             furtherPlan: {
@@ -1813,15 +1833,49 @@ var Page = class {
           );
         }
       },
-      move: async (x, y) => this.underlyingPage.mouse.move(x, y)
+      move: async (x, y) => this.underlyingPage.mouse.move(x, y),
+      drag: async (from, to) => {
+        if (this.pageType === "puppeteer") {
+          await this.underlyingPage.mouse.drag(
+            {
+              x: from.x,
+              y: from.y
+            },
+            {
+              x: to.x,
+              y: to.y
+            }
+          );
+        } else if (this.pageType === "playwright") {
+          await this.underlyingPage.mouse.move(
+            from.x,
+            from.y
+          );
+          await this.underlyingPage.mouse.down();
+          await this.underlyingPage.mouse.move(to.x, to.y);
+          await this.underlyingPage.mouse.up();
+        }
+      }
     };
   }
   get keyboard() {
     return {
       type: async (text) => this.underlyingPage.keyboard.type(text, { delay: 80 }),
-      press: async (key) => this.underlyingPage.keyboard.press(key),
-      down: async (key) => this.underlyingPage.keyboard.down(key),
-      up: async (key) => this.underlyingPage.keyboard.up(key)
+      press: async (key) => {
+        const keys = Array.isArray(key) ? key : [key];
+        for (const key2 of keys) {
+          await this.underlyingPage.keyboard.down(key2);
+        }
+        for (const key2 of [...keys].reverse()) {
+          await this.underlyingPage.keyboard.up(key2);
+        }
+      },
+      down: async (key) => {
+        this.underlyingPage.keyboard.down(key);
+      },
+      up: async (key) => {
+        this.underlyingPage.keyboard.up(key);
+      }
     };
   }
   async clearInput(element) {

package/dist/lib/puppeteer.js

@@ -555,7 +555,10 @@ var PageTaskExecutor = class {
           locate: plan2.locate,
           executor: async (param, taskContext) => {
             const { task } = taskContext;
-            (0, import_node_assert.default)((param == null ? void 0 : param.prompt) || (param == null ? void 0 : param.id), "No prompt or id to locate");
+            (0, import_node_assert.default)(
+              (param == null ? void 0 : param.prompt) || (param == null ? void 0 : param.id) || (param == null ? void 0 : param.position),
+              "No prompt or id or position to locate"
+            );
             let insightDump;
             const dumpCollector = (dump) => {
               insightDump = dump;
@@ -710,6 +713,22 @@ var PageTaskExecutor = class {
           }
         };
         tasks.push(taskActionTap);
+      } else if (plan2.type === "Drag") {
+        const taskActionDrag = {
+          type: "Action",
+          subType: "Drag",
+          param: plan2.param,
+          thought: plan2.thought,
+          locate: plan2.locate,
+          executor: async (taskParam) => {
+            (0, import_node_assert.default)(
+              (taskParam == null ? void 0 : taskParam.start_box) && (taskParam == null ? void 0 : taskParam.end_box),
+              "No start_box or end_box to drag"
+            );
+            await this.page.mouse.drag(taskParam.start_box, taskParam.end_box);
+          }
+        };
+        tasks.push(taskActionDrag);
       } else if (plan2.type === "Hover") {
         const taskActionHover = {
           type: "Action",
@@ -744,9 +763,9 @@ var PageTaskExecutor = class {
             } else if (scrollToEventName === "untilLeft") {
               await this.page.scrollUntilLeft(startingPoint);
             } else if (scrollToEventName === "once" || !scrollToEventName) {
-              if (taskParam.direction === "down" || !taskParam.direction) {
+              if ((taskParam == null ? void 0 : taskParam.direction) === "down" || !taskParam || !taskParam.direction) {
                 await this.page.scrollDown(
-                  taskParam.distance || void 0,
+                  (taskParam == null ? void 0 : taskParam.distance) || void 0,
                   startingPoint
                 );
               } else if (taskParam.direction === "up") {
@@ -941,6 +960,7 @@ var PageTaskExecutor = class {
         userPrompt
       },
       executor: async (param, executorContext) => {
+        var _a;
         const shotTime = Date.now();
         const pageContext = await this.insight.contextRetrieverFn("locate");
         const recordItem = {
@@ -977,7 +997,7 @@ var PageTaskExecutor = class {
         return {
           output: {
             actions,
-            thought: actions[0].thought,
+            thought: (_a = actions[0]) == null ? void 0 : _a.thought,
             actionType: actions[0].type,
             taskWillBeAccomplished: false,
             furtherPlan: {
@@ -1658,15 +1678,49 @@ var Page = class {
           );
         }
       },
-      move: async (x, y) => this.underlyingPage.mouse.move(x, y)
+      move: async (x, y) => this.underlyingPage.mouse.move(x, y),
+      drag: async (from, to) => {
+        if (this.pageType === "puppeteer") {
+          await this.underlyingPage.mouse.drag(
+            {
+              x: from.x,
+              y: from.y
+            },
+            {
+              x: to.x,
+              y: to.y
+            }
+          );
+        } else if (this.pageType === "playwright") {
+          await this.underlyingPage.mouse.move(
+            from.x,
+            from.y
+          );
+          await this.underlyingPage.mouse.down();
+          await this.underlyingPage.mouse.move(to.x, to.y);
+          await this.underlyingPage.mouse.up();
+        }
+      }
     };
   }
   get keyboard() {
     return {
       type: async (text) => this.underlyingPage.keyboard.type(text, { delay: 80 }),
-      press: async (key) => this.underlyingPage.keyboard.press(key),
-      down: async (key) => this.underlyingPage.keyboard.down(key),
-      up: async (key) => this.underlyingPage.keyboard.up(key)
+      press: async (key) => {
+        const keys = Array.isArray(key) ? key : [key];
+        for (const key2 of keys) {
+          await this.underlyingPage.keyboard.down(key2);
+        }
+        for (const key2 of [...keys].reverse()) {
+          await this.underlyingPage.keyboard.up(key2);
+        }
+      },
+      down: async (key) => {
+        this.underlyingPage.keyboard.down(key);
+      },
+      up: async (key) => {
+        this.underlyingPage.keyboard.up(key);
+      }
     };
   }
   async clearInput(element) {
@@ -1772,7 +1826,7 @@ var defaultViewportWidth = 1440;
 var defaultViewportHeight = 900;
 var defaultViewportScale = process.platform === "darwin" ? 2 : 1;
 var defaultWaitForNetworkIdleTimeout = 10 * 1e3;
-async function puppeteerAgentForTarget(target, preference) {
+async function launchPuppeteerPage(target, preference) {
   var _a, _b, _c;
   (0, import_node_assert3.default)(target.url, "url is required");
   const freeFn = [];
@@ -1827,10 +1881,12 @@ async function puppeteerAgentForTarget(target, preference) {
   const isWindows = process.platform === "win32";
   const browser = await puppeteer.launch({
     headless: !headed,
+    defaultViewport: viewportConfig,
     args: [
       ...isWindows ? [] : ["--no-sandbox", "--disable-setuid-sandbox"],
       "--disable-features=PasswordLeakDetection",
       "--disable-save-password-bubble",
+      `--user-agent="${ua}"`,
       preferMaximizedWindow ? "--start-maximized" : `--window-size=${width},${height}`
     ]
   });
@@ -1850,15 +1906,13 @@ async function puppeteerAgentForTarget(target, preference) {
   });
   const pages = await browser.pages();
   const page = pages[0];
-  await page.setUserAgent(ua);
-  await page.setViewport(viewportConfig);
   if (target.cookie) {
     const cookieFileContent = (0, import_node_fs3.readFileSync)(target.cookie, "utf-8");
     await page.setCookie(...JSON.parse(cookieFileContent));
   }
-  await page.goto(target.url);
   const waitForNetworkIdleTimeout = typeof ((_a = target.waitForNetworkIdle) == null ? void 0 : _a.timeout) === "number" ? target.waitForNetworkIdle.timeout : defaultWaitForNetworkIdleTimeout;
   try {
+    await page.goto(target.url);
     if (waitForNetworkIdleTimeout > 0) {
       await page.waitForNetworkIdle({
         timeout: waitForNetworkIdleTimeout
@@ -1874,9 +1928,15 @@ async function puppeteerAgentForTarget(target, preference) {
     const newMessage = `failed to wait for network idle after ${waitForNetworkIdleTimeout}ms, but the script will continue.`;
     console.warn(newMessage);
   }
+  return { page, freeFn };
+}
+async function puppeteerAgentForTarget(target, preference) {
+  const { page, freeFn } = await launchPuppeteerPage(target, preference);
   const agent = new PuppeteerAgent(page, {
     autoPrintReportMsg: false,
-    testId: preference == null ? void 0 : preference.testId
+    testId: preference == null ? void 0 : preference.testId,
+    trackingActiveTab: typeof target.trackingActiveTab !== "undefined" ? target.trackingActiveTab : true
+    // true for default in yaml script
   });
   freeFn.push({
     name: "midscene_puppeteer_agent",
@@ -1890,6 +1950,23 @@ var PuppeteerAgent = class extends PageAgent {
   constructor(page, opts) {
     const webPage = new WebPage(page);
     super(webPage, opts);
+    if (opts == null ? void 0 : opts.trackingActiveTab) {
+      const browser = this.page.underlyingPage.browser();
+      browser.on("targetcreated", async (target) => {
+        if (target.type() === "page") {
+          const targetPage = await target.page();
+          if (!targetPage) {
+            console.warn(
+              "got a targetPage event, but the page is not ready yet, skip"
+            );
+            return;
+          }
+          const midscenePage = new WebPage(targetPage);
+          this.page = midscenePage;
+          this.taskExecutor.page = midscenePage;
+        }
+      });
+    }
   }
 };
 // Annotate the CommonJS export names for ESM import in node:

package/dist/types/appium.d.ts

@@ -1,6 +1,6 @@
-export { P as AppiumAgent } from './tasks-f2973dd7.js';
-export { P as AppiumPage } from './page-fc3be0ec.js';
-import '@midscene/core/dist/lib/types/types-64c4d87b';
+export { P as AppiumAgent } from './tasks-681afd8a.js';
+export { P as AppiumPage } from './page-cfd75cdf.js';
+import '@midscene/core/dist/lib/types/types-7fe32cfe';
 import '@midscene/core';
 import '@midscene/core/ai-model';
 import '@midscene/shared/fs';

package/dist/types/bridge-mode-browser.d.ts

@@ -1,8 +1,8 @@
-export { C as ChromeExtensionPageBrowserSide } from './browser-fd2e9b68.js';
-import './page-fc3be0ec.js';
+export { C as ChromeExtensionPageBrowserSide } from './browser-bc6c204c.js';
+import './page-cfd75cdf.js';
 import 'playwright';
 import '@midscene/core';
 import 'puppeteer';
 import '@midscene/shared/constants';
-import '@midscene/core/dist/lib/types/types-64c4d87b';
+import '@midscene/core/dist/lib/types/types-7fe32cfe';
 import 'webdriverio';

package/dist/types/bridge-mode.d.ts

@@ -1,7 +1,8 @@
-import { P as PageAgent } from './tasks-f2973dd7.js';
-import { B as BridgeConnectTabOptions, C as ChromeExtensionPageBrowserSide } from './browser-fd2e9b68.js';
-import '@midscene/core/dist/lib/types/types-64c4d87b';
-import './page-fc3be0ec.js';
+import { P as PageAgent, a as PageAgentOpt } from './tasks-681afd8a.js';
+import { B as BridgeConnectTabOptions, C as ChromeExtensionPageBrowserSide } from './browser-bc6c204c.js';
+export { overrideAIConfig } from '@midscene/core/env';
+import '@midscene/core/dist/lib/types/types-7fe32cfe';
+import './page-cfd75cdf.js';
 import 'playwright';
 import '@midscene/core';
 import 'puppeteer';
@@ -14,7 +15,7 @@ interface ChromeExtensionPageCliSide extends ChromeExtensionPageBrowserSide {
     showStatusMessage: (message: string) => Promise<void>;
 }
 declare class AgentOverChromeBridge extends PageAgent<ChromeExtensionPageCliSide> {
-    constructor();
+    constructor(opts?: PageAgentOpt);
     connectNewTabWithUrl(url: string, options?: BridgeConnectTabOptions): Promise<void>;
     connectCurrentTab(options?: BridgeConnectTabOptions): Promise<void>;
     aiAction(prompt: string, options?: any): Promise<void>;