feat(android): add AccessibilityService, ScreenTreeBuilder, permissions

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

android/app/src/main/AndroidManifest.xml

@@ -2,6 +2,13 @@
 <manifest xmlns:android="http://schemas.android.com/apk/res/android"
     xmlns:tools="http://schemas.android.com/tools">
 
+    <uses-permission android:name="android.permission.INTERNET" />
+    <uses-permission android:name="android.permission.FOREGROUND_SERVICE" />
+    <uses-permission android:name="android.permission.FOREGROUND_SERVICE_CONNECTED_DEVICE" />
+    <uses-permission android:name="android.permission.POST_NOTIFICATIONS" />
+    <uses-permission android:name="android.permission.REQUEST_IGNORE_BATTERY_OPTIMIZATIONS" />
+    <uses-permission android:name="android.permission.WAKE_LOCK" />
+
     <application
         android:name=".DroidClawApp"
         android:allowBackup="true"
@@ -23,6 +30,23 @@
                 <category android:name="android.intent.category.LAUNCHER" />
             </intent-filter>
         </activity>
+
+        <service
+            android:name=".accessibility.DroidClawAccessibilityService"
+            android:permission="android.permission.BIND_ACCESSIBILITY_SERVICE"
+            android:exported="false">
+            <intent-filter>
+                <action android:name="android.accessibilityservice.AccessibilityService" />
+            </intent-filter>
+            <meta-data
+                android:name="android.accessibilityservice"
+                android:resource="@xml/accessibility_config" />
+        </service>
+
+        <service
+            android:name=".connection.ConnectionService"
+            android:foregroundServiceType="connectedDevice"
+            android:exported="false" />
     </application>
 
 </manifest>

android/app/src/main/java/com/thisux/droidclaw/accessibility/DroidClawAccessibilityService.kt

@@ -0,0 +1,96 @@
+package com.thisux.droidclaw.accessibility
+
+import android.accessibilityservice.AccessibilityService
+import android.util.Log
+import android.view.accessibility.AccessibilityEvent
+import android.view.accessibility.AccessibilityNodeInfo
+import com.thisux.droidclaw.model.UIElement
+import kotlinx.coroutines.delay
+import kotlinx.coroutines.flow.MutableStateFlow
+import kotlinx.coroutines.runBlocking
+
+class DroidClawAccessibilityService : AccessibilityService() {
+
+    companion object {
+        private const val TAG = "DroidClawA11y"
+        val isRunning = MutableStateFlow(false)
+        val lastScreenTree = MutableStateFlow<List<UIElement>>(emptyList())
+        var instance: DroidClawAccessibilityService? = null
+    }
+
+    override fun onServiceConnected() {
+        super.onServiceConnected()
+        Log.i(TAG, "Accessibility service connected")
+        instance = this
+        isRunning.value = true
+    }
+
+    override fun onAccessibilityEvent(event: AccessibilityEvent?) {
+        // We capture on-demand via getScreenTree(), not on every event
+    }
+
+    override fun onInterrupt() {
+        Log.w(TAG, "Accessibility service interrupted")
+    }
+
+    override fun onDestroy() {
+        super.onDestroy()
+        Log.i(TAG, "Accessibility service destroyed")
+        instance = null
+        isRunning.value = false
+    }
+
+    fun getScreenTree(): List<UIElement> {
+        val delays = longArrayOf(50, 100, 200)
+        for (delayMs in delays) {
+            val root = rootInActiveWindow
+            if (root != null) {
+                try {
+                    val elements = ScreenTreeBuilder.capture(root)
+                    lastScreenTree.value = elements
+                    return elements
+                } finally {
+                    root.recycle()
+                }
+            }
+            runBlocking { delay(delayMs) }
+        }
+        Log.w(TAG, "rootInActiveWindow null after retries")
+        return emptyList()
+    }
+
+    fun findNodeAt(x: Int, y: Int): AccessibilityNodeInfo? {
+        val root = rootInActiveWindow ?: return null
+        return findNodeAtRecursive(root, x, y)
+    }
+
+    private fun findNodeAtRecursive(
+        node: AccessibilityNodeInfo,
+        x: Int,
+        y: Int
+    ): AccessibilityNodeInfo? {
+        val rect = android.graphics.Rect()
+        node.getBoundsInScreen(rect)
+
+        if (!rect.contains(x, y)) {
+            node.recycle()
+            return null
+        }
+
+        for (i in 0 until node.childCount) {
+            val child = node.getChild(i) ?: continue
+            val found = findNodeAtRecursive(child, x, y)
+            if (found != null) {
+                node.recycle()
+                return found
+            }
+        }
+
+        return if (node.isClickable || node.isLongClickable || node.isEditable) {
+            node
+        } else {
+            node.recycle()
+            null
+        }
+    }
+}

android/app/src/main/java/com/thisux/droidclaw/accessibility/ScreenTreeBuilder.kt

@@ -0,0 +1,95 @@
+package com.thisux.droidclaw.accessibility
+
+import android.graphics.Rect
+import android.view.accessibility.AccessibilityNodeInfo
+import com.thisux.droidclaw.model.UIElement
+import java.security.MessageDigest
+
+object ScreenTreeBuilder {
+
+    fun capture(rootNode: AccessibilityNodeInfo?): List<UIElement> {
+        if (rootNode == null) return emptyList()
+        val elements = mutableListOf<UIElement>()
+        walkTree(rootNode, elements, depth = 0, parentDesc = "")
+        return elements
+    }
+
+    private fun walkTree(
+        node: AccessibilityNodeInfo,
+        elements: MutableList<UIElement>,
+        depth: Int,
+        parentDesc: String
+    ) {
+        try {
+            val rect = Rect()
+            node.getBoundsInScreen(rect)
+
+            val text = node.text?.toString() ?: ""
+            val contentDesc = node.contentDescription?.toString() ?: ""
+            val viewId = node.viewIdResourceName ?: ""
+            val className = node.className?.toString() ?: ""
+            val displayText = text.ifEmpty { contentDesc }
+
+            val isInteractive = node.isClickable || node.isLongClickable ||
+                node.isEditable || node.isScrollable || node.isFocusable
+
+            if (isInteractive || displayText.isNotEmpty()) {
+                val centerX = (rect.left + rect.right) / 2
+                val centerY = (rect.top + rect.bottom) / 2
+                val width = rect.width()
+                val height = rect.height()
+
+                val action = when {
+                    node.isEditable -> "type"
+                    node.isScrollable -> "scroll"
+                    node.isLongClickable -> "longpress"
+                    node.isClickable -> "tap"
+                    else -> "read"
+                }
+
+                elements.add(
+                    UIElement(
+                        id = viewId,
+                        text = displayText,
+                        type = className.substringAfterLast("."),
+                        bounds = "[${rect.left},${rect.top}][${rect.right},${rect.bottom}]",
+                        center = listOf(centerX, centerY),
+                        size = listOf(width, height),
+                        clickable = node.isClickable,
+                        editable = node.isEditable,
+                        enabled = node.isEnabled,
+                        checked = node.isChecked,
+                        focused = node.isFocused,
+                        selected = node.isSelected,
+                        scrollable = node.isScrollable,
+                        longClickable = node.isLongClickable,
+                        password = node.isPassword,
+                        hint = node.hintText?.toString() ?: "",
+                        action = action,
+                        parent = parentDesc,
+                        depth = depth
+                    )
+                )
+            }
+
+            for (i in 0 until node.childCount) {
+                val child = node.getChild(i) ?: continue
+                try {
+                    walkTree(child, elements, depth + 1, className)
+                } finally {
+                    child.recycle()
+                }
+            }
+        } catch (_: Exception) {
+            // Node may have been recycled during traversal
+        }
+    }
+
+    fun computeScreenHash(elements: List<UIElement>): String {
+        val digest = MessageDigest.getInstance("MD5")
+        for (el in elements) {
+            digest.update("${el.id}|${el.text}|${el.center}".toByteArray())
+        }
+        return digest.digest().joinToString("") { "%02x".format(it) }.take(12)
+    }
+}

android/app/src/main/java/com/thisux/droidclaw/connection/ConnectionService.kt

@@ -0,0 +1,18 @@
+package com.thisux.droidclaw.connection
+
+import android.app.Service
+import android.content.Intent
+import android.os.IBinder
+
+/**
+ * Foreground service for maintaining the WebSocket connection to the DroidClaw server.
+ * Full implementation will be added in Task 9.
+ */
+class ConnectionService : Service() {
+
+    override fun onBind(intent: Intent?): IBinder? = null
+
+    override fun onStartCommand(intent: Intent?, flags: Int, startId: Int): Int {
+        return START_NOT_STICKY
+    }
+}

android/app/src/main/res/xml/accessibility_config.xml

@@ -0,0 +1,9 @@
+<?xml version="1.0" encoding="utf-8"?>
+<accessibility-service xmlns:android="http://schemas.android.com/apk/res/android"
+    android:accessibilityEventTypes="typeWindowStateChanged|typeWindowContentChanged|typeViewFocused"
+    android:accessibilityFeedbackType="feedbackGeneric"
+    android:accessibilityFlags="flagDefault|flagIncludeNotImportantViews|flagReportViewIds|flagRequestEnhancedWebAccessibility"
+    android:canPerformGestures="true"
+    android:canRetrieveWindowContent="true"
+    android:notificationTimeout="100"
+    android:settingsActivity="com.thisux.droidclaw.MainActivity" />