alumnium 0.16.0 → 0.17.0

package/dist/Alumni.d.ts

@@ -7,6 +7,7 @@ import { Data } from "./clients/typecasting.js";
 import { BaseDriver } from "./drivers/BaseDriver.js";
 import { Element } from "./drivers/index.js";
 import { Model } from "./Model.js";
+import { DoResult } from "./result.js";
 import { ToolClass } from "./tools/BaseTool.js";
 export interface AlumniOptions {
     url?: string;
@@ -25,7 +26,7 @@ export declare class Alumni {
     private model;
     constructor(driver: WebDriver | Page | Browser, options?: AlumniOptions);
     quit(): Promise<void>;
-    do(goal: string): Promise<void>;
+    do(goal: string): Promise<DoResult>;
     check(statement: string, options?: VisionOptions): Promise<string>;
     get(data: string, options?: VisionOptions): Promise<Data>;
     find(description: string): Promise<Element>;

package/dist/Alumni.js

@@ -1,3 +1,12 @@
+var __decorate = (this && this.__decorate) || function (decorators, target, key, desc) {
+    var c = arguments.length, r = c < 3 ? target : desc === null ? desc = Object.getOwnPropertyDescriptor(target, key) : desc, d;
+    if (typeof Reflect === "object" && typeof Reflect.decorate === "function") r = Reflect.decorate(decorators, target, key, desc);
+    else for (var i = decorators.length - 1; i >= 0; i--) if (d = decorators[i]) r = (c < 3 ? d(r) : c > 3 ? d(target, key, r) : d(target, key)) || r;
+    return c > 3 && r && Object.defineProperty(target, key, r), r;
+};
+var __metadata = (this && this.__metadata) || function (k, v) {
+    if (typeof Reflect === "object" && typeof Reflect.metadata === "function") return Reflect.metadata(k, v);
+};
 import { WebDriver } from "selenium-webdriver";
 import { Area } from "./Area.js";
 import { Cache } from "./Cache.js";
@@ -8,16 +17,12 @@ import { SeleniumDriver } from "./drivers/SeleniumDriver.js";
 import { AssertionError } from "./errors/AssertionError.js";
 import { Model } from "./Model.js";
 import { BaseTool } from "./tools/BaseTool.js";
-import { ClickTool } from "./tools/ClickTool.js";
-import { DragAndDropTool } from "./tools/DragAndDropTool.js";
-import { HoverTool } from "./tools/HoverTool.js";
-import { PressKeyTool } from "./tools/PressKeyTool.js";
-import { SelectTool } from "./tools/SelectTool.js";
-import { TypeTool } from "./tools/TypeTool.js";
 import { getLogger } from "./utils/logger.js";
+import { retry } from "./utils/retry.js";
 const logger = getLogger(["Alumni"]);
 export class Alumni {
     constructor(driver, options = {}) {
+        this.tools = {};
         this.url = options.url || "http://localhost:8013";
         this.model = options.model || Model.current;
         // Wrap driver or use directly if already wrapped
@@ -35,20 +40,11 @@ export class Alumni {
         else {
             throw new Error(`Unsupported driver type '${typeof driver}'`);
         }
-        // Initialize tools
-        this.tools = {
-            ClickTool,
-            TypeTool,
-            HoverTool,
-            SelectTool,
-            PressKeyTool,
-            DragAndDropTool,
-        };
-        // Add extra tools if provided
-        if (options.extraTools) {
-            for (const tool of options.extraTools) {
-                this.tools[tool.name] = tool;
-            }
+        for (const tool of new Set([
+            ...this.driver.supportedTools,
+            ...(options.extraTools || []),
+        ])) {
+            this.tools[tool.name] = tool;
         }
         // Initialize HTTP client
         this.client = new HttpClient(this.url, this.model, this.driver.platform, this.tools);
@@ -62,7 +58,8 @@ export class Alumni {
     }
     async do(goal) {
         const initialAccessibilityTree = await this.driver.getAccessibilityTree();
-        const steps = await this.client.planActions(goal, initialAccessibilityTree.toStr());
+        const { explanation, steps } = await this.client.planActions(goal, initialAccessibilityTree.toStr());
+        const executedSteps = [];
         for (let idx = 0; idx < steps.length; idx++) {
             const step = steps[idx];
             // Use initial tree for first step, fresh tree for subsequent steps
@@ -70,11 +67,14 @@ export class Alumni {
                 ? initialAccessibilityTree
                 : await this.driver.getAccessibilityTree();
             const actorResponse = await this.client.executeAction(goal, step, accessibilityTree.toStr());
-            // Execute tool calls
+            const calledTools = [];
             for (const toolCall of actorResponse) {
-                await BaseTool.executeToolCall(toolCall, this.tools, this.driver);
+                const calledTool = await BaseTool.executeToolCall(toolCall, this.tools, this.driver);
+                calledTools.push(calledTool);
             }
+            executedSteps.push({ name: step, tools: calledTools });
         }
+        return { explanation, steps: executedSteps };
     }
     async check(statement, options = {}) {
         const screenshot = options.vision
@@ -92,9 +92,8 @@ export class Alumni {
             ? await this.driver.screenshot()
             : undefined;
         const accessibilityTree = await this.driver.getAccessibilityTree();
-        // eslint-disable-next-line @typescript-eslint/no-unused-vars
-        const [_explanation, value] = await this.client.retrieve(data, accessibilityTree.toStr(), await this.driver.title(), await this.driver.url(), screenshot);
-        return value;
+        const [explanation, value] = await this.client.retrieve(data, accessibilityTree.toStr(), await this.driver.title(), await this.driver.url(), screenshot);
+        return value === null ? explanation : value;
     }
     async find(description) {
         const accessibilityTree = await this.driver.getAccessibilityTree();
@@ -117,3 +116,27 @@ export class Alumni {
         return await this.client.getStats();
     }
 }
+__decorate([
+    retry(),
+    __metadata("design:type", Function),
+    __metadata("design:paramtypes", [String]),
+    __metadata("design:returntype", Promise)
+], Alumni.prototype, "do", null);
+__decorate([
+    retry(),
+    __metadata("design:type", Function),
+    __metadata("design:paramtypes", [String, Object]),
+    __metadata("design:returntype", Promise)
+], Alumni.prototype, "check", null);
+__decorate([
+    retry(),
+    __metadata("design:type", Function),
+    __metadata("design:paramtypes", [String, Object]),
+    __metadata("design:returntype", Promise)
+], Alumni.prototype, "get", null);
+__decorate([
+    retry(),
+    __metadata("design:type", Function),
+    __metadata("design:paramtypes", [String]),
+    __metadata("design:returntype", Promise)
+], Alumni.prototype, "find", null);

package/dist/Area.d.ts

@@ -4,6 +4,7 @@ import { HttpClient } from "./clients/HttpClient.js";
 import { Data } from "./clients/typecasting.js";
 import { BaseDriver } from "./drivers/BaseDriver.js";
 import { Element } from "./drivers/index.js";
+import { DoResult } from "./result.js";
 import { ToolClass } from "./tools/BaseTool.js";
 export declare class Area {
     id: number;
@@ -13,7 +14,7 @@ export declare class Area {
     private tools;
     private client;
     constructor(id: number, description: string, accessibilityTree: BaseAccessibilityTree, driver: BaseDriver, tools: Record<string, ToolClass>, client: HttpClient);
-    do(goal: string): Promise<void>;
+    do(goal: string): Promise<DoResult>;
     check(statement: string, options?: VisionOptions): Promise<string>;
     get(data: string, options?: VisionOptions): Promise<Data>;
     find(description: string): Promise<Element>;

package/dist/Area.js

@@ -1,5 +1,15 @@
+var __decorate = (this && this.__decorate) || function (decorators, target, key, desc) {
+    var c = arguments.length, r = c < 3 ? target : desc === null ? desc = Object.getOwnPropertyDescriptor(target, key) : desc, d;
+    if (typeof Reflect === "object" && typeof Reflect.decorate === "function") r = Reflect.decorate(decorators, target, key, desc);
+    else for (var i = decorators.length - 1; i >= 0; i--) if (d = decorators[i]) r = (c < 3 ? d(r) : c > 3 ? d(target, key, r) : d(target, key)) || r;
+    return c > 3 && r && Object.defineProperty(target, key, r), r;
+};
+var __metadata = (this && this.__metadata) || function (k, v) {
+    if (typeof Reflect === "object" && typeof Reflect.metadata === "function") return Reflect.metadata(k, v);
+};
 import { AssertionError } from "./errors/AssertionError.js";
 import { BaseTool } from "./tools/BaseTool.js";
+import { retry } from "./utils/retry.js";
 export class Area {
     constructor(id, description, accessibilityTree, driver, tools, client) {
         this.id = id;
@@ -10,14 +20,18 @@ export class Area {
         this.client = client;
     }
     async do(goal) {
-        const steps = await this.client.planActions(goal, this.accessibilityTree.toStr());
+        const { explanation, steps } = await this.client.planActions(goal, this.accessibilityTree.toStr());
+        const executedSteps = [];
         for (const step of steps) {
             const actorResponse = await this.client.executeAction(goal, step, this.accessibilityTree.toStr());
-            // Execute tool calls
+            const calledTools = [];
             for (const toolCall of actorResponse) {
-                await BaseTool.executeToolCall(toolCall, this.tools, this.driver);
+                const calledTool = await BaseTool.executeToolCall(toolCall, this.tools, this.driver);
+                calledTools.push(calledTool);
             }
+            executedSteps.push({ name: step, tools: calledTools });
         }
+        return { explanation, steps: executedSteps };
     }
     async check(statement, options = {}) {
         const screenshot = options.vision
@@ -33,12 +47,35 @@ export class Area {
         const screenshot = options.vision
             ? await this.driver.screenshot()
             : undefined;
-        // eslint-disable-next-line @typescript-eslint/no-unused-vars
-        const [_explanation, value] = await this.client.retrieve(data, this.accessibilityTree.toStr(), await this.driver.title(), await this.driver.url(), screenshot);
-        return value;
+        const [explanation, value] = await this.client.retrieve(data, this.accessibilityTree.toStr(), await this.driver.title(), await this.driver.url(), screenshot);
+        return value === null ? explanation : value;
     }
     async find(description) {
         const response = await this.client.findElement(description, this.accessibilityTree.toStr());
         return this.driver.findElement(response.id);
     }
 }
+__decorate([
+    retry(),
+    __metadata("design:type", Function),
+    __metadata("design:paramtypes", [String]),
+    __metadata("design:returntype", Promise)
+], Area.prototype, "do", null);
+__decorate([
+    retry(),
+    __metadata("design:type", Function),
+    __metadata("design:paramtypes", [String, Object]),
+    __metadata("design:returntype", Promise)
+], Area.prototype, "check", null);
+__decorate([
+    retry(),
+    __metadata("design:type", Function),
+    __metadata("design:paramtypes", [String, Object]),
+    __metadata("design:returntype", Promise)
+], Area.prototype, "get", null);
+__decorate([
+    retry(),
+    __metadata("design:type", Function),
+    __metadata("design:paramtypes", [String]),
+    __metadata("design:returntype", Promise)
+], Area.prototype, "find", null);

package/dist/Model.js

@@ -15,17 +15,17 @@ export var Provider;
 export class ModelName {
 }
 ModelName.DEFAULT = {
-    [Provider.AZURE_OPENAI]: "gpt-4o-mini",
+    [Provider.AZURE_OPENAI]: "gpt-5-nano",
     [Provider.ANTHROPIC]: "claude-haiku-4-5-20251001",
     [Provider.AWS_ANTHROPIC]: "us.anthropic.claude-haiku-4-5-20251001-v1:0",
     [Provider.AWS_META]: "us.meta.llama4-maverick-17b-instruct-v1:0",
-    [Provider.DEEPSEEK]: "deepseek-chat",
+    [Provider.DEEPSEEK]: "deepseek-reasoner",
     [Provider.GITHUB]: "gpt-4o-mini",
-    [Provider.GOOGLE]: "gemini-2.0-flash-001",
+    [Provider.GOOGLE]: "gemini-3-flash-preview",
     [Provider.MISTRALAI]: "mistral-medium-2505",
     [Provider.OLLAMA]: "mistral-small3.1",
-    [Provider.OPENAI]: "gpt-4o-mini-2024-07-18",
-    [Provider.XAI]: "grok-4-fast-non-reasoning",
+    [Provider.OPENAI]: "gpt-5-nano-2025-08-07",
+    [Provider.XAI]: "grok-4-1-fast-reasoning",
 };
 export class Model {
     constructor(provider, name) {

package/dist/accessibility/UIAutomator2AccessibilityTree.d.ts

@@ -5,6 +5,7 @@ export declare class UIAutomator2AccessibilityTree extends BaseAccessibilityTree
     private nextRawId;
     private raw;
     constructor(xmlString: string);
+    private static fromXml;
     toStr(): string;
     private addRawIds;
     elementById(rawId: number): AccessibilityElement;

package/dist/accessibility/UIAutomator2AccessibilityTree.js

@@ -12,6 +12,11 @@ export class UIAutomator2AccessibilityTree extends BaseAccessibilityTree {
         const cleanedXmlContent = cleanedLines.join("\n");
         this.xmlString = `<?xml version='1.0' encoding='UTF-8' standalone='yes' ?>\n<root>\n${cleanedXmlContent}\n</root>`;
     }
+    static fromXml(xmlString) {
+        const instance = new UIAutomator2AccessibilityTree("");
+        instance.raw = xmlString;
+        return instance;
+    }
     toStr() {
         if (this.raw !== null) {
             return this.raw;
@@ -53,7 +58,7 @@ export class UIAutomator2AccessibilityTree extends BaseAccessibilityTree {
             return this;
         }
         const scopedXml = this.elementToString(element, 0);
-        return new UIAutomator2AccessibilityTree(scopedXml);
+        return UIAutomator2AccessibilityTree.fromXml(scopedXml);
     }
     findElementByRawId(xmlString, targetRawId) {
         const elements = this.parseSimpleXml(xmlString);

package/dist/accessibility/XCUITestAccessibilityTree.d.ts

@@ -5,11 +5,13 @@ export declare class XCUITestAccessibilityTree extends BaseAccessibilityTree {
     private nextRawId;
     private raw;
     constructor(xmlString: string);
+    private static fromXml;
     toStr(): string;
     private addRawIds;
     elementById(rawId: number): AccessibilityElement;
     scopeToArea(rawId: number): XCUITestAccessibilityTree;
     private findElementByRawId;
     private elementToString;
+    private decodeHtmlEntities;
     private parseSimpleXml;
 }

package/dist/accessibility/XCUITestAccessibilityTree.js

@@ -7,6 +7,11 @@ export class XCUITestAccessibilityTree extends BaseAccessibilityTree {
         this.raw = null;
         this.xmlString = xmlString;
     }
+    static fromXml(xmlString) {
+        const instance = new XCUITestAccessibilityTree("");
+        instance.raw = xmlString;
+        return instance;
+    }
     toStr() {
         if (this.raw !== null) {
             return this.raw;
@@ -48,7 +53,7 @@ export class XCUITestAccessibilityTree extends BaseAccessibilityTree {
             return this;
         }
         const scopedXml = this.elementToString(element, 0);
-        return new XCUITestAccessibilityTree(scopedXml);
+        return XCUITestAccessibilityTree.fromXml(scopedXml);
     }
     findElementByRawId(xmlString, targetRawId) {
         const elements = this.parseSimpleXml(xmlString);
@@ -95,6 +100,29 @@ export class XCUITestAccessibilityTree extends BaseAccessibilityTree {
         }
         return result;
     }
+    decodeHtmlEntities(str) {
+        const htmlEntities = {
+            "&": "&",
+            "&lt;": "<",
+            "&gt;": ">",
+            "&quot;": '"',
+            "&apos;": "'",
+        };
+        let decoded = str;
+        let previous;
+        // Keep decoding until no more entities are found (handles double-encoded entities)
+        const MAX_DECODE_ITERATIONS = 10;
+        let iterations = 0;
+        do {
+            previous = decoded;
+            decoded = decoded.replace(/&(?:amp|lt|gt|quot|apos);/g, (entity) => htmlEntities[entity] || entity);
+            iterations++;
+            if (iterations >= MAX_DECODE_ITERATIONS) {
+                break;
+            }
+        } while (previous !== decoded);
+        return decoded;
+    }
     parseSimpleXml(xmlString) {
         const elements = [];
         const stack = [];
@@ -121,7 +149,8 @@ export class XCUITestAccessibilityTree extends BaseAccessibilityTree {
                 let attrMatch;
                 attrRegex.lastIndex = 0;
                 while ((attrMatch = attrRegex.exec(attrsString)) !== null) {
-                    attributes[attrMatch[1]] = attrMatch[2];
+                    // Decode HTML entities in attribute values
+                    attributes[attrMatch[1]] = this.decodeHtmlEntities(attrMatch[2]);
                 }
                 const elem = {
                     tag: tagName,

package/dist/clients/ApiModels.d.ts

@@ -22,6 +22,7 @@ export interface PlanRequest {
     title?: string;
 }
 export interface PlanResponse {
+    explanation: string;
     steps: string[];
 }
 export interface StepRequest {

package/dist/clients/HttpClient.d.ts

@@ -14,7 +14,10 @@ export declare class HttpClient {
     private fetchWithTimeout;
     private ensureSession;
     quit(): Promise<void>;
-    planActions(goal: string, accessibilityTree: string): Promise<string[]>;
+    planActions(goal: string, accessibilityTree: string): Promise<{
+        explanation: string;
+        steps: string[];
+    }>;
     addExample(goal: string, actions: string[]): Promise<void>;
     clearExamples(): Promise<void>;
     executeAction(goal: string, step: string, accessibilityTree: string): Promise<StepResponse["actions"]>;

package/dist/clients/HttpClient.js

@@ -71,7 +71,7 @@ export class HttpClient {
             body: JSON.stringify(requestBody),
         });
         const responseData = (await response.json());
-        return responseData.steps;
+        return { explanation: responseData.explanation, steps: responseData.steps };
     }
     async addExample(goal, actions) {
         await this.ensureSession();

package/dist/drivers/AppiumDriver.d.ts

@@ -1,11 +1,12 @@
-import type { Browser, ChainablePromiseElement } from "webdriverio";
+import type { Browser } from "webdriverio";
 import { BaseAccessibilityTree } from "../accessibility/BaseAccessibilityTree.js";
+import { ToolClass } from "../tools/BaseTool.js";
 import { BaseDriver } from "./BaseDriver.js";
 import { Key } from "./keys.js";
 export declare class AppiumDriver extends BaseDriver {
     private driver;
     platform: "xcuitest" | "uiautomator2";
-    supportedTools: Set<string>;
+    supportedTools: Set<ToolClass>;
     autoswitchContexts: boolean;
     delay: number;
     doubleFetchPageSource: boolean;
@@ -16,15 +17,17 @@ export declare class AppiumDriver extends BaseDriver {
     dragAndDrop(fromId: number, toId: number): Promise<void>;
     pressKey(key: Key): Promise<void>;
     back(): Promise<void>;
+    visit(url: string): Promise<void>;
+    scrollTo(id: number): Promise<void>;
     quit(): Promise<void>;
     screenshot(): Promise<string>;
-    select(_id: number, _option: string): Promise<void>;
-    swipe(_id: number): Promise<void>;
     title(): Promise<string>;
     type(id: number, text: string): Promise<void>;
     url(): Promise<string>;
-    findElement(id: number): Promise<ChainablePromiseElement>;
-    hover(_id: number): Promise<void>;
+    findElement(id: number): Promise<WebdriverIO.Element>;
+    executeScript(script: string): Promise<void>;
     private ensureNativeAppContext;
     private ensureWebviewContext;
+    private hideKeyboard;
+    private scrollIntoView;
 }

package/dist/drivers/AppiumDriver.js

@@ -1,25 +1,27 @@
 import { Key as SeleniumKey } from "selenium-webdriver";
 import { UIAutomator2AccessibilityTree } from "../accessibility/UIAutomator2AccessibilityTree.js";
 import { XCUITestAccessibilityTree } from "../accessibility/XCUITestAccessibilityTree.js";
+import { ClickTool } from "../tools/ClickTool.js";
+import { DragAndDropTool } from "../tools/DragAndDropTool.js";
+import { PressKeyTool } from "../tools/PressKeyTool.js";
+import { TypeTool } from "../tools/TypeTool.js";
 import { BaseDriver } from "./BaseDriver.js";
 import { Key } from "./keys.js";
 export class AppiumDriver extends BaseDriver {
     constructor(driver) {
         super();
         this.supportedTools = new Set([
-            "ClickTool",
-            "DragAndDropTool",
-            "PressKeyTool",
-            "SelectTool",
-            "TypeTool",
+            ClickTool,
+            DragAndDropTool,
+            PressKeyTool,
+            TypeTool,
         ]);
         this.autoswitchContexts = true;
         this.delay = 0;
         this.doubleFetchPageSource = false;
         this.hideKeyboardAfterTyping = false;
         this.driver = driver;
-        if (this.driver.capabilities["appium:automationName"]?.toLowerCase() ===
-            "uiautomator2") {
+        if (this.driver.capabilities.platformName?.toLowerCase() === "android") {
             this.platform = "uiautomator2";
         }
         else {
@@ -37,7 +39,7 @@ export class AppiumDriver extends BaseDriver {
             await this.driver.getPageSource();
         }
         const xmlString = await this.driver.getPageSource();
-        if (this.driver.capabilities["appium:automationName"] === "uiautomator2") {
+        if (this.platform === "uiautomator2") {
             return new UIAutomator2AccessibilityTree(xmlString);
         }
         else {
@@ -47,13 +49,14 @@ export class AppiumDriver extends BaseDriver {
     async click(id) {
         await this.ensureNativeAppContext();
         const element = await this.findElement(id);
+        await this.scrollIntoView(element);
         await element.click();
     }
     async dragAndDrop(fromId, toId) {
         await this.ensureNativeAppContext();
         const fromElement = await this.findElement(fromId);
         const toElement = await this.findElement(toId);
-        // WebdriverIO provides dragAndDrop method on elements
+        await this.scrollIntoView(fromElement);
         await fromElement.dragAndDrop(toElement);
     }
     async pressKey(key) {
@@ -79,20 +82,19 @@ export class AppiumDriver extends BaseDriver {
     async back() {
         await this.driver.back();
     }
+    async visit(url) {
+        await this.driver.url(url);
+    }
+    async scrollTo(id) {
+        const element = await this.findElement(id);
+        await this.scrollIntoView(element);
+    }
     async quit() {
         // WebdriverIO handles session termination automatically.
     }
     async screenshot() {
         return await this.driver.takeScreenshot();
     }
-    // eslint-disable-next-line @typescript-eslint/no-unused-vars
-    async select(_id, _option) {
-        // TODO: Implement select functionality and the tool
-    }
-    // eslint-disable-next-line @typescript-eslint/no-unused-vars
-    async swipe(_id) {
-        // TODO: Implement swipe functionality and the tool
-    }
     async title() {
         await this.ensureWebviewContext();
         try {
@@ -105,32 +107,10 @@ export class AppiumDriver extends BaseDriver {
     async type(id, text) {
         await this.ensureNativeAppContext();
         const element = await this.findElement(id);
-        await element.clearValue();
+        await this.scrollIntoView(element);
         await element.setValue(text);
-        if (this.hideKeyboardAfterTyping) {
-            if (this.platform === "uiautomator2") {
-                await this.driver.hideKeyboard();
-            }
-            else {
-                const location = await element.getLocation();
-                await this.driver.performActions([
-                    {
-                        type: "pointer",
-                        id: "finger1",
-                        parameters: { pointerType: "touch" },
-                        actions: [
-                            {
-                                type: "pointerMove",
-                                duration: 0,
-                                x: location.x,
-                                y: location.y - 20,
-                            },
-                            { type: "pointerDown", button: 0 },
-                            { type: "pointerUp", button: 0 },
-                        ],
-                    },
-                ]);
-            }
+        if (this.hideKeyboardAfterTyping && (await this.driver.isKeyboardShown())) {
+            await this.hideKeyboard();
         }
     }
     async url() {
@@ -161,8 +141,7 @@ export class AppiumDriver extends BaseDriver {
                 predicate += ` AND ${propsStr}`;
             }
             console.debug(`Finding element by predicate: ${predicate}`);
-            const foundElement = this.driver.$(`-ios predicate string:${predicate}`);
-            return foundElement;
+            return this.driver.$(`-ios predicate string:${predicate}`).getElement();
         }
         else {
             // Use XPath for UIAutomator2
@@ -170,23 +149,19 @@ export class AppiumDriver extends BaseDriver {
             const props = {};
             if (element.androidResourceId)
                 props["resource-id"] = element.androidResourceId;
-            if (element.androidText)
-                props["text"] = element.androidText;
-            if (element.androidContentDesc)
-                props["content-desc"] = element.androidContentDesc;
             if (element.androidBounds)
                 props["bounds"] = element.androidBounds;
             if (Object.keys(props).length > 0) {
                 const conditions = Object.entries(props).map(([k, v]) => `@${k}="${v}"`);
                 xpath += `[${conditions.join(" and ")}]`;
             }
-            return this.driver.$(xpath);
+            console.debug(`Finding element by xpath: ${xpath}`);
+            return this.driver.$(xpath).getElement();
         }
     }
-    // eslint-disable-next-line @typescript-eslint/require-await, @typescript-eslint/no-unused-vars
-    async hover(_id) {
-        // Hover is not typically supported in mobile contexts
-        throw new Error("Hover is not supported by AppiumDriver");
+    async executeScript(script) {
+        await this.ensureWebviewContext();
+        await this.driver.execute(script);
     }
     async ensureNativeAppContext() {
         if (!this.autoswitchContexts) {
@@ -212,4 +187,28 @@ export class AppiumDriver extends BaseDriver {
             }
         }
     }
+    async hideKeyboard() {
+        if (this.platform === "uiautomator2") {
+            await this.driver.hideKeyboard();
+        }
+        else {
+            // Tap to the top left corner of the keyboard to dismiss it
+            const keyboard = this.driver.$("-ios predicate string:type == 'XCUIElementTypeKeyboard'");
+            const { width, height } = await keyboard.getSize();
+            await keyboard.click({
+                x: -Math.ceil(width / 2),
+                y: -Math.ceil(height / 2),
+            });
+        }
+    }
+    async scrollIntoView(element) {
+        if (this.platform === "uiautomator2") {
+            await element.scrollIntoView();
+        }
+        else {
+            await this.driver.execute("mobile: scrollToElement", {
+                elementId: element.elementId,
+            });
+        }
+    }
 }

package/dist/drivers/BaseDriver.d.ts

@@ -1,8 +1,10 @@
 import { BaseAccessibilityTree } from "../accessibility/BaseAccessibilityTree.js";
+import { ToolClass } from "../tools/BaseTool.js";
 import { Element } from "./index.js";
 import { Key } from "./keys.js";
 export declare abstract class BaseDriver {
     abstract platform: string;
+    abstract supportedTools: Set<ToolClass>;
     abstract getAccessibilityTree(): Promise<BaseAccessibilityTree>;
     abstract click(id: number): void | Promise<void>;
     abstract dragAndDrop(fromId: number, toId: number): void | Promise<void>;
@@ -10,10 +12,11 @@ export declare abstract class BaseDriver {
     abstract quit(): void | Promise<void>;
     abstract back(): void | Promise<void>;
     abstract screenshot(): string | Promise<string>;
-    abstract select(id: number, option: string): void | Promise<void>;
     abstract title(): string | Promise<string>;
     abstract type(id: number, text: string): void | Promise<void>;
     abstract url(): string | Promise<string>;
     abstract findElement(id: number): Element | Promise<Element>;
-    abstract hover(id: number): void | Promise<void>;
+    abstract visit(url: string): void | Promise<void>;
+    abstract scrollTo(id: number): void | Promise<void>;
+    abstract executeScript(script: string): void | Promise<void>;
 }