feat: Add getCurrentActivity tool for unified app/activity detection

src/android.ts

@@ -463,6 +463,37 @@ export class AndroidRobot implements Robot {
 		return rotation === "0" ? "portrait" : "landscape";
 	}
 
+	public async getCurrentActivity(): Promise<{ id: string; isCanonical: boolean }> {
+		try {
+			const dumpsysOutput = this.adb("shell", "dumpsys", "activity", "activities").toString();
+
+			// Try to find mResumedActivity first (preferred for foreground app)
+			let focusMatch = dumpsysOutput.match(/mResumedActivity=ActivityRecord\{[^\s]+ u\d+ ([^\s/]+)\//);
+			if (focusMatch && focusMatch[1]) {
+				return { id: focusMatch[1], isCanonical: true };
+			}
+
+			// Fallback to mCurrentFocus (may point to IME/system windows)
+			focusMatch = dumpsysOutput.match(/mCurrentFocus=Window\{[^\s]+ u\d+ ([^\s/]+)\//);
+			if (focusMatch && focusMatch[1]) {
+				return { id: focusMatch[1], isCanonical: true };
+			}
+
+			// Fallback to mFocusedActivity (legacy, dead on Android 9+)
+			focusMatch = dumpsysOutput.match(/mFocusedActivity=ActivityRecord\{[^\s]+ u\d+ ([^\s/]+)\//);
+			if (focusMatch && focusMatch[1]) {
+				return { id: focusMatch[1], isCanonical: true };
+			}
+
+			throw new ActionableError("No activity is currently in focus. Please launch an app and try again.");
+		} catch (error) {
+			if (error instanceof ActionableError) {
+				throw error;
+			}
+			throw new ActionableError("Failed to get current activity. Please ensure the device is properly connected.");
+		}
+	}
+
 	private async getUiAutomatorDump(): Promise<string> {
 		for (let tries = 0; tries < 10; tries++) {
 			const dump = this.adb("exec-out", "uiautomator", "dump", "/dev/tty").toString();

src/ios.ts

@@ -1,7 +1,7 @@
 import { Socket } from "node:net";
 import { execFileSync } from "node:child_process";
 
-import { WebDriverAgent } from "./webdriver-agent";
+import { WebDriverAgent, parseWdaPageSourceForAppId } from "./webdriver-agent";
 import { ActionableError, Button, InstalledApp, Robot, ScreenSize, SwipeDirection, ScreenElement, Orientation } from "./robot";
 
 const WDA_PORT = 8100;
@@ -229,6 +229,22 @@ export class IosRobot implements Robot {
 		const wda = await this.wda();
 		return await wda.getOrientation();
 	}
+
+	public async getCurrentActivity(): Promise<{ id: string; isCanonical: boolean }> {
+		try {
+			const wda = await this.wda();
+			const [source, sessionBundleId] = await Promise.all([
+				wda.getPageSource(),
+				wda.getActiveSessionBundleId()
+			]);
+			return parseWdaPageSourceForAppId(source, sessionBundleId);
+		} catch (error) {
+			if (error instanceof ActionableError) {
+				throw error;
+			}
+			throw new ActionableError("Failed to get current app. Please ensure the device is properly connected and WebDriver Agent is running.");
+		}
+	}
 }
 
 export class IosManager {

src/iphone-simulator.ts

@@ -4,7 +4,7 @@ import { tmpdir } from "node:os";
 import { join, basename, extname } from "node:path";
 
 import { trace } from "./logger";
-import { WebDriverAgent } from "./webdriver-agent";
+import { WebDriverAgent, parseWdaPageSourceForAppId } from "./webdriver-agent";
 import { ActionableError, Button, InstalledApp, Robot, ScreenElement, ScreenSize, SwipeDirection, Orientation } from "./robot";
 
 export interface Simulator {
@@ -269,4 +269,20 @@ export class Simctl implements Robot {
 		const wda = await this.wda();
 		return wda.getOrientation();
 	}
+
+	public async getCurrentActivity(): Promise<{ id: string; isCanonical: boolean }> {
+		try {
+			const wda = await this.wda();
+			const [source, sessionBundleId] = await Promise.all([
+				wda.getPageSource(),
+				wda.getActiveSessionBundleId()
+			]);
+			return parseWdaPageSourceForAppId(source, sessionBundleId);
+		} catch (error) {
+			if (error instanceof ActionableError) {
+				throw error;
+			}
+			throw new ActionableError("Failed to get current app. Please ensure WebDriver Agent is running.");
+		}
+	}
 }

src/mobile-device.ts

@@ -1,5 +1,5 @@
 import { Mobilecli } from "./mobilecli";
-import { Button, InstalledApp, Orientation, Robot, ScreenElement, ScreenSize, SwipeDirection } from "./robot";
+import { ActionableError, Button, InstalledApp, Orientation, Robot, ScreenElement, ScreenSize, SwipeDirection } from "./robot";
 
 interface InstalledAppsResponse {
 	status: "ok",
@@ -213,4 +213,22 @@ export class MobileDevice implements Robot {
 		const response = JSON.parse(this.runCommand(["device", "orientation", "get"])) as OrientationResponse;
 		return response.data.orientation;
 	}
+
+	public async getCurrentActivity(): Promise<{ id: string; isCanonical: boolean }> {
+		try {
+			const response = JSON.parse(this.runCommand(["device", "get-current-activity"])) as any;
+			if (response.status === "ok" && response.data?.id) {
+				return {
+					id: response.data.id,
+					isCanonical: response.data.isCanonical ?? true
+				};
+			}
+			throw new ActionableError("No activity is currently in focus. Please launch an app and try again.");
+		} catch (error) {
+			if (error instanceof ActionableError) {
+				throw error;
+			}
+			throw new ActionableError("Failed to get current activity. Please ensure the device is properly connected.");
+		}
+	}
 }

src/robot.ts

@@ -143,4 +143,12 @@ export interface Robot {
 	 * Get the current screen orientation.
 	 */
 	getOrientation(): Promise<Orientation>;
+
+	/**
+	 * Get the currently running activity (Android) or foreground app (iOS).
+	 * @returns An object with the `id` field containing the package name (Android) or bundle ID (iOS),
+	 * and `isCanonical` indicating if the ID is a canonical identifier.
+	 * @throws ActionableError if no activity is currently in focus
+	 */
+	getCurrentActivity(): Promise<{ id: string; isCanonical: boolean }>;
 }

src/server.ts

@@ -637,5 +637,20 @@ export const createMcpServer = (): McpServer => {
 		}
 	);
 
+	tool(
+		"mobile_get_current_activity",
+		"Get Current Activity",
+		"Get the currently running activity (Android) or foreground app (iOS). Returns the package name for Android or bundle ID for iOS.",
+		{
+			device: z.string().describe("The device identifier to use. Use mobile_list_available_devices to find which devices are available to you.")
+		},
+		{ readOnlyHint: true },
+		async ({ device }) => {
+			const robot = getRobotFromDevice(device);
+			const activity = await robot.getCurrentActivity();
+			return JSON.stringify(activity);
+		}
+	);
+
 	return server;
 };

src/webdriver-agent.ts

@@ -451,4 +451,67 @@ export class WebDriverAgent {
 			return json.value.toLowerCase() as Orientation;
 		});
 	}
+
+	public async getActiveSessionBundleId(): Promise<string | null> {
+		try {
+			const sessionsUrl = `http://${this.host}:${this.port}/sessions`;
+			const sessionsResponse = await fetch(sessionsUrl);
+			const sessionsJson = await sessionsResponse.json();
+			const sessions = sessionsJson.value || [];
+
+			if (sessions[0]?.id) {
+				const sessionId = sessions[0].id;
+				const sessionUrl = `http://${this.host}:${this.port}/session/${sessionId}`;
+				const response = await fetch(sessionUrl);
+				const json = await response.json();
+				// Some WDA versions put it in capabilities, some in the root
+				return json.value?.capabilities?.bundleId || json.value?.bundleId || null;
+			}
+		} catch (error) {
+			// ignore and fall back to tree parsing
+		}
+		return null;
+	}
+}
+
+/**
+ * Extracts the app bundle identifier from the WDA page source.
+ *
+ * Resolution order:
+ * 1. `rootElement.rawIdentifier`: Canonical bundle ID if present.
+ * 2. `sessionBundleId`: Canonical bundle ID derived from active session capabilities.
+ * 3. `rootElement.type`: Fallback if it matches bundle ID format (contains dots).
+ * 4. `rootElement.name`: Last-resort fallback. Note that this returns the app's
+ *    accessibility label / display name (e.g., "Safari") which is non-canonical.
+ *
+ * @param source The WDA page source tree.
+ * @param sessionBundleId Optional bundle ID from active session capabilities.
+ * @returns An object containing the resolved app identifier and whether it is canonical.
+ */
+export function parseWdaPageSourceForAppId(source: SourceTree, sessionBundleId?: string | null): { id: string; isCanonical: boolean } {
+	const rootElement = source.value;
+
+	// 1. Prefer rawIdentifier (usually contains bundle ID)
+	if (rootElement.rawIdentifier) {
+		return { id: rootElement.rawIdentifier, isCanonical: true };
+	}
+
+	// 2. Fallback to bundleId from active session if provided
+	if (sessionBundleId) {
+		return { id: sessionBundleId, isCanonical: true };
+	}
+
+	// 3. Fallback to type if it looks like a bundle ID (contains a dot)
+	// We check for dots to avoid generic types like "XCUIElementTypeApplication"
+	if (rootElement.type && rootElement.type.includes(".")) {
+		return { id: rootElement.type, isCanonical: true };
+	}
+
+	// 4. Fallback to name (accessibility label/display name) as a last resort
+	// NOTE: This may return "Safari" instead of "com.apple.mobilesafari"
+	if (rootElement.name) {
+		return { id: rootElement.name, isCanonical: false };
+	}
+
+	throw new ActionableError("No app is in foreground. Please launch an app and try again.");
 }

test/android.ts

@@ -5,14 +5,14 @@ import { AndroidRobot, AndroidDeviceManager } from "../src/android";
 
 const manager = new AndroidDeviceManager();
 const devices = manager.getConnectedDevices();
-const hasOneAndroidDevice = devices.length === 1;
+const hasAndroidDevice = devices.length > 0;
 
 describe("android", () => {
 
 	const android = new AndroidRobot(devices?.[0]?.deviceId || "");
 
 	it("should be able to get the screen size", async function() {
-		hasOneAndroidDevice || this.skip();
+		hasAndroidDevice || this.skip();
 		const screenSize = await android.getScreenSize();
 		assert.ok(screenSize.width > 1024);
 		assert.ok(screenSize.height > 1024);
@@ -21,7 +21,7 @@ describe("android", () => {
 	});
 
 	it("should be able to take screenshot", async function() {
-		hasOneAndroidDevice || this.skip();
+		hasAndroidDevice || this.skip();
 
 		const screenSize = await android.getScreenSize();
 		const screenshot = await android.getScreenshot();
@@ -35,20 +35,20 @@ describe("android", () => {
 	});
 
 	it("should be able to list apps", async function() {
-		hasOneAndroidDevice || this.skip();
+		hasAndroidDevice || this.skip();
 		const apps = await android.listApps();
 		const packages = apps.map(app => app.packageName);
 		assert.ok(packages.includes("com.android.settings"));
 	});
 
 	it("should be able to open a url", async function() {
-		hasOneAndroidDevice || this.skip();
+		hasAndroidDevice || this.skip();
 		await android.adb("shell", "input", "keyevent", "HOME");
 		await android.openUrl("https://www.example.com");
 	});
 
 	it("should be able to list elements on screen", async function() {
-		hasOneAndroidDevice || this.skip();
+		hasAndroidDevice || this.skip();
 		await android.terminateApp("com.android.chrome");
 		await android.adb("shell", "input", "keyevent", "HOME");
 		await android.openUrl("https://www.example.com");
@@ -68,7 +68,7 @@ describe("android", () => {
 	});
 
 	it("should be able to send keys and tap", async function() {
-		hasOneAndroidDevice || this.skip();
+		hasAndroidDevice || this.skip();
 		await android.terminateApp("com.google.android.deskclock");
 		await android.adb("shell", "pm", "clear", "com.google.android.deskclock");
 		await android.launchApp("com.google.android.deskclock");
@@ -97,7 +97,7 @@ describe("android", () => {
 	});
 
 	it("should be able to launch and terminate an app", async function() {
-		hasOneAndroidDevice || this.skip();
+		hasAndroidDevice || this.skip();
 
 		// kill if running
 		await android.terminateApp("com.android.chrome");
@@ -113,7 +113,7 @@ describe("android", () => {
 	});
 
 	it("should handle orientation changes", async function() {
-		hasOneAndroidDevice || this.skip();
+		hasAndroidDevice || this.skip();
 
 		// assume we start in portrait
 		const originalOrientation = await android.getOrientation();
@@ -136,4 +136,17 @@ describe("android", () => {
 		// screen size should not have changed
 		assert.deepEqual(screenSize1, screenSize2);
 	});
+
+	it("should get current activity", async function() {
+		hasAndroidDevice || this.skip();
+
+		// Go to home screen
+		await android.pressButton("HOME");
+		await new Promise(resolve => setTimeout(resolve, 500));
+
+		// Get current activity
+		const activity = await android.getCurrentActivity();
+		assert.ok(activity.id, "Activity id should be defined");
+		assert.ok(typeof activity.id === "string", "Activity id should be a string");
+	});
 });