Orca is the place for working with a fleet of agents (e.g. Grok Bot, Claude Code). Run any coding agent with your own subscription. Available on desktop, mobile and remote runtime.
TypeScript adeagent-ideai-agentsautomationbotclaude-codecodexghosttygrok-bothermes-agentmobile-appmuseopenclawopencodeorchestrationpiruntimeterminalworktreesyc-backed
fix(browser): keep an agent on its own tab when the user picks or opens another (#36108)
Kelvin Amoaba324f138de21c
diff --git a/skill-guides/orca-cli/references/browser.md b/skill-guides/orca-cli/references/browser.mdindex b07f91ea782..a8435f8a8ce 100644--- a/skill-guides/orca-cli/references/browser.md+++ b/skill-guides/orca-cli/references/browser.md@@ -1,68 +1,71 @@# Built-in browser commands-Use a snapshot-interact-re-snapshot loop:+Pin one page, then use a snapshot-interact-re-snapshot loop on it:```text-ORCA goto --url https://example.com --json-ORCA snapshot --json-ORCA click --element @e3 --json-ORCA snapshot --json+ORCA tab current --json+ORCA goto --page <browserPageId> --url https://example.com --json+ORCA snapshot --page <browserPageId> --json+ORCA click --page <browserPageId> --element @e3 --json+ORCA snapshot --page <browserPageId> --json```Common commands:```text-ORCA goto --url <url> --json-ORCA back --json-ORCA reload --json-ORCA snapshot --json-ORCA screenshot --json-ORCA full-screenshot --json-ORCA pdf --json-ORCA click --element <ref> --json-ORCA fill --element <ref> --value <text> --json-ORCA type --input <text> --json-ORCA select --element <ref> --value <value> --json-ORCA check --element <ref> --json-ORCA scroll --direction down --amount 1000 --json-ORCA hover --element <ref> --json-ORCA focus --element <ref> --json-ORCA keypress --key Enter --json-ORCA upload --element <ref> --files <paths> --json-ORCA wait --text <text> --json-ORCA wait --url <substring> --json-ORCA wait --selector <css> --json-ORCA wait --load networkidle --json-ORCA eval --expression <js> --json-ORCA set device --name "iPhone 15" --json-ORCA viewport --width 390 --height 844 --scale 3 --mobile --json+ORCA goto --page <browserPageId> --url <url> --json+ORCA back --page <browserPageId> --json+ORCA reload --page <browserPageId> --json+ORCA snapshot --page <browserPageId> --json+ORCA screenshot --page <browserPageId> --json+ORCA full-screenshot --page <browserPageId> --json+ORCA pdf --page <browserPageId> --json+ORCA click --page <browserPageId> --element <ref> --json+ORCA fill --page <browserPageId> --element <ref> --value <text> --json+ORCA type --page <browserPageId> --input <text> --json+ORCA select --page <browserPageId> --element <ref> --value <value> --json+ORCA check --page <browserPageId> --element <ref> --json+ORCA scroll --page <browserPageId> --direction down --amount 1000 --json+ORCA hover --page <browserPageId> --element <ref> --json+ORCA focus --page <browserPageId> --element <ref> --json+ORCA keypress --page <browserPageId> --key Enter --json+ORCA upload --page <browserPageId> --element <ref> --files <paths> --json+ORCA wait --page <browserPageId> --text <text> --json+ORCA wait --page <browserPageId> --url <substring> --json+ORCA wait --page <browserPageId> --selector <css> --json+ORCA wait --page <browserPageId> --load networkidle --json+ORCA eval --page <browserPageId> --expression <js> --json+ORCA set device --page <browserPageId> --name "iPhone 15" --json+ORCA viewport --page <browserPageId> --width 390 --height 844 --scale 3 --mobile --jsonORCA tab list --jsonORCA tab create --url <url> --jsonORCA tab switch --index <n> --jsonORCA tab close --index <n> --json-ORCA cookie get --json-ORCA capture start --json-ORCA console --limit 50 --json-ORCA network --limit 50 --json-ORCA exec --command "help" --json+ORCA cookie get --page <browserPageId> --json+ORCA capture start --page <browserPageId> --json+ORCA console --page <browserPageId> --limit 50 --json+ORCA network --page <browserPageId> --limit 50 --json+ORCA exec --page <browserPageId> --command "help" --json```Browser rules:- Re-snapshot after navigation, tab switches, clicks that change the page, and any `browser_stale_ref`.- Refs like `@e1` are assigned by `snapshot`, scoped to one tab, and invalidated by navigation or tab switch.-- Browser commands default to the current worktree and its active tab. Use `--worktree all` only intentionally.-- For concurrent browser work, run `ORCA tab list --json`, read `tabs[].browserPageId`, and pass `--page <browserPageId>` on later commands.+- Pass `--page <browserPageId>` on every command that acts on one page; `ORCA <command> --help` lists `--page` when it does. `tab list`, `tab current`, `tab create` and `open-url` take no `--page`, and `tab switch`/`tab close` pick a tab by `--index` or `--page`, never both, since `--page` wins. Take the ID from `result.tab.browserPageId` in `ORCA tab current --json` or `result.browserPageId` in `ORCA tab create --json` and keep it for the task; `tab current` shows the routed tab, not the one you pin. When the user means a specific tab, find it by URL or title in `ORCA tab list --json`.+- Without `--page`, a command goes to the worktree's current tab. That stays the agent's tab while it has had a command in the last 20 seconds or so, even after the user picks another (`tab current` reports it too), then moves to the user's pick.+- After `ORCA tab switch`, pass the new page's `browserPageId`.+- Browser commands default to the current worktree. `--page` reaches the page in any worktree unless you also pass `--worktree <selector>`. Use `--worktree all` only intentionally.- Use typed tab commands (`ORCA tab list/create/close/switch`), not `ORCA exec --command "tab ..."`, so Orca keeps UI state synchronized.- Use `ORCA set device` or `ORCA viewport --mobile`, not `ORCA exec --command "set device ..."`, so the page also gets touch input and `pointer: coarse`. Each page load (reload, `goto`, link click) resets it, so run it again after navigating.- Prefer `wait --text`, `--url`, `--selector`, or `--load` after async page changes instead of bare timeouts.-- Anything not listed above goes through `ORCA exec --command "<agent-browser command>"`.-- If `fill` or `type` fails on a custom input, try `ORCA focus --element @e1 --json` then `ORCA inserttext --text "text" --json`.+- Anything not listed above goes through `ORCA exec --page <browserPageId> --command "<agent-browser command>"`.+- If `fill` or `type` fails on a custom input, try `ORCA focus --page <browserPageId> --element @e1 --json` then `ORCA inserttext --page <browserPageId> --text "text" --json`.- A client-hosted page renders in the paired desktop's browser engine, so every command against it needs that desktop online and returns `browser_host_unavailable` while it is closed, asleep, or disconnected. Server-hosted pages run with no desktop attached; prefer them for long or unattended automation.Common recoveries:- `browser_no_tab`: open a tab with `ORCA tab create --url <url> --json`.-- `browser_stale_ref`: run `ORCA snapshot --json` and retry with fresh refs.+- `browser_stale_ref`: run `ORCA snapshot --page <browserPageId> --json` and retry on the same page with fresh refs.- `browser_tab_not_found`: run `ORCA tab list --json` before switching or closing.- `browser_host_unavailable`: the desktop hosting the page is offline. Bring it back, or recreate the page with server placement if the work must outlive the desktop session.diff --git a/src/cli/bundled-skill-guides.ts b/src/cli/bundled-skill-guides.tsindex 46a8ed6a45b..2b649c28076 100644--- a/src/cli/bundled-skill-guides.ts+++ b/src/cli/bundled-skill-guides.ts@@ -811,70 +811,73 @@ const ORCA_CLI_AUTOMATIONS_REFERENCE_MARKDOWN = [const ORCA_CLI_BROWSER_REFERENCE_MARKDOWN = ["# Built-in browser commands","",- "Use a snapshot-interact-re-snapshot loop:",+ "Pin one page, then use a snapshot-interact-re-snapshot loop on it:","","```text",- "ORCA goto --url https://example.com --json",- "ORCA snapshot --json",- "ORCA click --element @e3 --json",- "ORCA snapshot --json",+ "ORCA tab current --json",+ "ORCA goto --page <browserPageId> --url https://example.com --json",+ "ORCA snapshot --page <browserPageId> --json",+ "ORCA click --page <browserPageId> --element @e3 --json",+ "ORCA snapshot --page <browserPageId> --json","```","","Common commands:","","```text",- "ORCA goto --url <url> --json",- "ORCA back --json",- "ORCA reload --json",- "ORCA snapshot --json",- "ORCA screenshot --json",- "ORCA full-screenshot --json",- "ORCA pdf --json",- "ORCA click --element <ref> --json",- "ORCA fill --element <ref> --value <text> --json",- "ORCA type --input <text> --json",- "ORCA select --element <ref> --value <value> --json",- "ORCA check --element <ref> --json",- "ORCA scroll --direction down --amount 1000 --json",- "ORCA hover --element <ref> --json",- "ORCA focus --element <ref> --json",- "ORCA keypress --key Enter --json",- "ORCA upload --element <ref> --files <paths> --json",- "ORCA wait --text <text> --json",- "ORCA wait --url <substring> --json",- "ORCA wait --selector <css> --json",- "ORCA wait --load networkidle --json",- "ORCA eval --expression <js> --json",- "ORCA set device --name \"iPhone 15\" --json",- "ORCA viewport --width 390 --height 844 --scale 3 --mobile --json",+ "ORCA goto --page <browserPageId> --url <url> --json",+ "ORCA back --page <browserPageId> --json",+ "ORCA reload --page <browserPageId> --json",+ "ORCA snapshot --page <browserPageId> --json",+ "ORCA screenshot --page <browserPageId> --json",+ "ORCA full-screenshot --page <browserPageId> --json",+ "ORCA pdf --page <browserPageId> --json",+ "ORCA click --page <browserPageId> --element <ref> --json",+ "ORCA fill --page <browserPageId> --element <ref> --value <text> --json",+ "ORCA type --page <browserPageId> --input <text> --json",+ "ORCA select --page <browserPageId> --element <ref> --value <value> --json",+ "ORCA check --page <browserPageId> --element <ref> --json",+ "ORCA scroll --page <browserPageId> --direction down --amount 1000 --json",+ "ORCA hover --page <browserPageId> --element <ref> --json",+ "ORCA focus --page <browserPageId> --element <ref> --json",+ "ORCA keypress --page <browserPageId> --key Enter --json",+ "ORCA upload --page <browserPageId> --element <ref> --files <paths> --json",+ "ORCA wait --page <browserPageId> --text <text> --json",+ "ORCA wait --page <browserPageId> --url <substring> --json",+ "ORCA wait --page <browserPageId> --selector <css> --json",+ "ORCA wait --page <browserPageId> --load networkidle --json",+ "ORCA eval --page <browserPageId> --expression <js> --json",+ "ORCA set device --page <browserPageId> --name \"iPhone 15\" --json",+ "ORCA viewport --page <browserPageId> --width 390 --height 844 --scale 3 --mobile --json","ORCA tab list --json","ORCA tab create --url <url> --json","ORCA tab switch --index <n> --json","ORCA tab close --index <n> --json",- "ORCA cookie get --json",- "ORCA capture start --json",- "ORCA console --limit 50 --json",- "ORCA network --limit 50 --json",- "ORCA exec --command \"help\" --json",+ "ORCA cookie get --page <browserPageId> --json",+ "ORCA capture start --page <browserPageId> --json",+ "ORCA console --page <browserPageId> --limit 50 --json",+ "ORCA network --page <browserPageId> --limit 50 --json",+ "ORCA exec --page <browserPageId> --command \"help\" --json","```","","Browser rules:","","- Re-snapshot after navigation, tab switches, clicks that change the page, and any `browser_stale_ref`.","- Refs like `@e1` are assigned by `snapshot`, scoped to one tab, and invalidated by navigation or tab switch.",- "- Browser commands default to the current worktree and its active tab. Use `--worktree all` only intentionally.",- "- For concurrent browser work, run `ORCA tab list --json`, read `tabs[].browserPageId`, and pass `--page <browserPageId>` on later commands.",+ "- Pass `--page <browserPageId>` on every command that acts on one page; `ORCA <command> --help` lists `--page` when it does. `tab list`, `tab current`, `tab create` and `open-url` take no `--page`, and `tab switch`/`tab close` pick a tab by `--index` or `--page`, never both, since `--page` wins. Take the ID from `result.tab.browserPageId` in `ORCA tab current --json` or `result.browserPageId` in `ORCA tab create --json` and keep it for the task; `tab current` shows the routed tab, not the one you pin. When the user means a specific tab, find it by URL or title in `ORCA tab list --json`.",+ "- Without `--page`, a command goes to the worktree's current tab. That stays the agent's tab while it has had a command in the last 20 seconds or so, even after the user picks another (`tab current` reports it too), then moves to the user's pick.",+ "- After `ORCA tab switch`, pass the new page's `browserPageId`.",+ "- Browser commands default to the current worktree. `--page` reaches the page in any worktree unless you also pass `--worktree <selector>`. Use `--worktree all` only intentionally.","- Use typed tab commands (`ORCA tab list/create/close/switch`), not `ORCA exec --command \"tab ...\"`, so Orca keeps UI state synchronized.","- Use `ORCA set device` or `ORCA viewport --mobile`, not `ORCA exec --command \"set device ...\"`, so the page also gets touch input and `pointer: coarse`. Each page load (reload, `goto`, link click) resets it, so run it again after navigating.","- Prefer `wait --text`, `--url`, `--selector`, or `--load` after async page changes instead of bare timeouts.",- "- Anything not listed above goes through `ORCA exec --command \"<agent-browser command>\"`.",- "- If `fill` or `type` fails on a custom input, try `ORCA focus --element @e1 --json` then `ORCA inserttext --text \"text\" --json`.",+ "- Anything not listed above goes through `ORCA exec --page <browserPageId> --command \"<agent-browser command>\"`.",+ "- If `fill` or `type` fails on a custom input, try `ORCA focus --page <browserPageId> --element @e1 --json` then `ORCA inserttext --page <browserPageId> --text \"text\" --json`.","- A client-hosted page renders in the paired desktop's browser engine, so every command against it needs that desktop online and returns `browser_host_unavailable` while it is closed, asleep, or disconnected. Server-hosted pages run with no desktop attached; prefer them for long or unattended automation.","","Common recoveries:","","- `browser_no_tab`: open a tab with `ORCA tab create --url <url> --json`.",- "- `browser_stale_ref`: run `ORCA snapshot --json` and retry with fresh refs.",+ "- `browser_stale_ref`: run `ORCA snapshot --page <browserPageId> --json` and retry on the same page with fresh refs.","- `browser_tab_not_found`: run `ORCA tab list --json` before switching or closing.","- `browser_host_unavailable`: the desktop hosting the page is offline. Bring it back, or recreate the page with server placement if the work must outlive the desktop session.","",diff --git a/src/main/browser/agent-browser-bridge-lifecycle.ts b/src/main/browser/agent-browser-bridge-lifecycle.tsindex d19093da513..0a5ec29ec4e 100644--- a/src/main/browser/agent-browser-bridge-lifecycle.ts+++ b/src/main/browser/agent-browser-bridge-lifecycle.ts@@ -62,6 +62,11 @@ export abstract class AgentBrowserBridgeLifecycle extends AgentBrowserBridgeRawP) {this.activeWebContentsPerWorktree.set(owningWorktreeId, newWebContentsId)}+ for (const [key, heldWebContentsId] of this.heldPersonSelections) {+ if (heldWebContentsId === oldWebContentsId) {+ this.heldPersonSelections.set(key, newWebContentsId)+ }+ }this.options.onTabsChanged?.(owningWorktreeId ?? undefined)}protected async ensureSession(diff --git a/src/main/browser/agent-browser-bridge-queue.ts b/src/main/browser/agent-browser-bridge-queue.tsindex 1a550067ebd..3617b129c80 100644--- a/src/main/browser/agent-browser-bridge-queue.ts+++ b/src/main/browser/agent-browser-bridge-queue.ts@@ -40,12 +40,9 @@ export abstract class AgentBrowserBridgeQueue extends AgentBrowserBridgeShutdown)}const [tabId, wcId] = liveEntries[switchedIndex]- this.activeWebContentsId = wcIdconst owningWorktreeId = worktreeId ?? this.browserManager.getWorktreeIdForTab(tabId)// Page-only switches must also update the owning workspace for later scoped commands.- if (owningWorktreeId) {- this.activeWebContentsPerWorktree.set(owningWorktreeId, wcId)- }+ this.pointAgentAt(wcId, owningWorktreeId ?? undefined)this.options.onTabsChanged?.(owningWorktreeId ?? undefined)return { switched: switchedIndex, browserPageId: tabId }})diff --git a/src/main/browser/agent-browser-bridge-state.ts b/src/main/browser/agent-browser-bridge-state.tsindex 0b926d036cb..da3fcdd4079 100644--- a/src/main/browser/agent-browser-bridge-state.ts+++ b/src/main/browser/agent-browser-bridge-state.ts@@ -13,6 +13,8 @@ export abstract class AgentBrowserBridgeState {// Isolate each worktree's tab selection.protected readonly activeWebContentsPerWorktree = new Map<string, number>()protected activeWebContentsId: number | null = null+ // A person's tab pick waiting for the agent's tab to go idle; the undefined key is the global target.+ protected readonly heldPersonSelections = new Map<string | undefined, number>()protected readonly sessions = new Map<string, SessionState>()protected readonly commandQueues = new Map<string, QueuedCommand[]>()protected readonly processingQueues = new Set<string>()diff --git a/src/main/browser/agent-browser-bridge-tab-routing.test.ts b/src/main/browser/agent-browser-bridge-tab-routing.test.tsindex e8d0d41f458..5911889c06e 100644--- a/src/main/browser/agent-browser-bridge-tab-routing.test.ts+++ b/src/main/browser/agent-browser-bridge-tab-routing.test.ts@@ -53,6 +53,7 @@ vi.mock('./cdp-bridge', () => ({}))import { AgentBrowserBridge } from './agent-browser-bridge'+import { runWithActiveRpcCaller } from '../runtime/rpc/rpc-active-caller'import {answerAgentBrowserCommands,mockBrowserManager,@@ -247,6 +248,71 @@ describe('AgentBrowserBridge', () => {expect(b.tabList('wt-1').tabs).toMatchObject([{ browserPageId: 'tab-a', active: true }])})+ describe('while an agent drives a tab', () => {+ const agentPages = new Set<string>()+ const agentOnTabA = (): AgentBrowserBridge => {+ const tabs = new Map([+ ['tab-a', 1],+ ['tab-b', 2],+ ['tab-c', 3]+ ])+ const worktrees = new Map([+ ['tab-a', 'wt-1'],+ ['tab-b', 'wt-1'],+ ['tab-c', 'wt-1']+ ])+ const guests = new Map(+ [...tabs].map(([tabId, id]) => [id, mockWebContents(id, `https://${tabId}.com`, tabId)])+ )+ webContentsFromIdMock.mockImplementation((id: number) => guests.get(id) ?? null)+ const b = new AgentBrowserBridge(+ mockBrowserManager(tabs, worktrees, {+ isAgentUsingPage: (id: string) => agentPages.has(id),+ beginAgentCommand: (id: string) => {+ agentPages.add(id)+ return { run: (execute) => execute(), end: () => {} }+ }+ })+ )+ b.onTabChanged(1, 'wt-1')+ agentPages.clear()+ agentPages.add('tab-a')+ return b+ }++ it('keeps unpinned commands on that tab until it goes idle when the user picks another', () => {+ const b = agentOnTabA()++ b.onTabChanged(2, 'wt-1')+ expect(b.getActivePageId('wt-1')).toBe('tab-a')++ agentPages.clear()+ expect(b.getActivePageId('wt-1')).toBe('tab-b')+ })++ it('holds a pick made before the agent switches tabs until the new tab goes idle', async () => {+ const b = agentOnTabA()++ b.onTabChanged(3, 'wt-1')+ await b.tabSwitch(undefined, 'wt-1', 'tab-b')+ expect(b.getActivePageId('wt-1')).toBe('tab-b')++ agentPages.clear()+ expect(b.getActivePageId('wt-1')).toBe('tab-c')+ })++ it.each([+ [{ kind: 'paired-device', deviceId: 'phone' } as const, 'tab-a'],+ [{ kind: 'local-cli' } as const, 'tab-b']+ ])('a tab created by %o routes unpinned commands to %s', (caller, expected) => {+ const b = agentOnTabA()++ runWithActiveRpcCaller(caller, () => b.setActiveTab(2, 'wt-1'))++ expect(b.getActivePageId('wt-1')).toBe(expected)+ })+ })+it('updates the owning worktree active tab when switching by browser page id', async () => {const tabs = new Map([['tab-a', 1],diff --git a/src/main/browser/agent-browser-bridge-tabs.ts b/src/main/browser/agent-browser-bridge-tabs.tsindex 16ffbe33113..047b7fd978b 100644--- a/src/main/browser/agent-browser-bridge-tabs.ts+++ b/src/main/browser/agent-browser-bridge-tabs.ts@@ -2,16 +2,76 @@ import type { BrowserTabInfo, BrowserTabListResult } from '../../shared/runtime-import { BrowserError } from './cdp-bridge'import type { ResolvedBrowserCommandTarget } from './agent-browser-bridge-types'import { AgentBrowserBridgeState } from './agent-browser-bridge-state'+import { isAgentRpcCaller } from '../runtime/rpc/rpc-caller-identity'+import { getActiveRpcCaller } from '../runtime/rpc/rpc-active-caller'export abstract class AgentBrowserBridgeTabs extends AgentBrowserBridgeState {setActiveTab(webContentsId: number, worktreeId?: string): void {- this.activeWebContentsId = webContentsId- if (worktreeId) {- this.activeWebContentsPerWorktree.set(worktreeId, webContentsId)+ if (isAgentRpcCaller(getActiveRpcCaller())) {+ this.pointAgentAt(webContentsId, worktreeId)+ } else {+ this.selectForPerson(webContentsId, worktreeId)}this.options.onTabsChanged?.(worktreeId)}+ protected pointAgentAt(webContentsId: number, worktreeId?: string): void {+ for (const key of routingKeys(worktreeId)) {+ this.setRoutedWebContentsId(key, webContentsId)+ }+ // Mark the new tab now, or a person's pick before the agent's next command would take it.+ const browserPageId = this.resolveTabIdSafe(webContentsId)+ if (browserPageId) {+ this.browserManager.beginAgentCommand(browserPageId).end()+ }+ }++ // A person's pick never pulls an agent off the tab it is driving; it applies once that tab goes idle.+ private selectForPerson(webContentsId: number, worktreeId?: string): void {+ for (const key of routingKeys(worktreeId)) {+ const current = this.routedWebContentsId(key)+ if (current !== null && current !== webContentsId && this.isAgentUsingTab(current)) {+ this.heldPersonSelections.set(key, webContentsId)+ } else {+ this.setRoutedWebContentsId(key, webContentsId)+ this.heldPersonSelections.delete(key)+ }+ }+ }++ protected applyHeldPersonSelections(worktreeId?: string): void {+ for (const key of routingKeys(worktreeId)) {+ const held = this.heldPersonSelections.get(key)+ const current = this.routedWebContentsId(key)+ if (held === undefined || (current !== null && this.isAgentUsingTab(current))) {+ continue+ }+ this.heldPersonSelections.delete(key)+ if (this.getWebContents(held)) {+ this.setRoutedWebContentsId(key, held)+ }+ }+ }++ private isAgentUsingTab(webContentsId: number): boolean {+ const browserPageId = this.resolveTabIdSafe(webContentsId)+ return browserPageId !== null && this.browserManager.isAgentUsingPage(browserPageId)+ }++ private routedWebContentsId(key: string | undefined): number | null {+ return key === undefined+ ? this.activeWebContentsId+ : (this.activeWebContentsPerWorktree.get(key) ?? null)+ }++ private setRoutedWebContentsId(key: string | undefined, webContentsId: number): void {+ if (key === undefined) {+ this.activeWebContentsId = webContentsId+ } else {+ this.activeWebContentsPerWorktree.set(key, webContentsId)+ }+ }+protected selectFallbackActiveWebContents(worktreeId: string,excludedWebContentsId?: number@@ -53,10 +113,7 @@ export abstract class AgentBrowserBridgeTabs extends AgentBrowserBridgeState {}}onTabChanged(webContentsId: number, worktreeId?: string): void {- this.activeWebContentsId = webContentsId- if (worktreeId) {- this.activeWebContentsPerWorktree.set(worktreeId, webContentsId)- }+ this.selectForPerson(webContentsId, worktreeId)this.options.onTabsChanged?.(worktreeId)}getRegisteredTabs(worktreeId?: string): Map<string, number> {@@ -82,6 +139,7 @@ export abstract class AgentBrowserBridgeTabs extends AgentBrowserBridgeState {}tabList(worktreeId?: string): BrowserTabListResult {+ this.applyHeldPersonSelections(worktreeId)const tabs = this.getRegisteredTabs(worktreeId)// Match command routing without changing active-tab selection during discovery.let activeWcId =@@ -162,6 +220,7 @@ export abstract class AgentBrowserBridgeTabs extends AgentBrowserBridgeState {}protected resolveActiveTab(worktreeId?: string): ResolvedBrowserCommandTarget {+ this.applyHeldPersonSelections(worktreeId)const tabs = this.getRegisteredTabs(worktreeId)if (tabs.size === 0) {@@ -233,3 +292,7 @@ export abstract class AgentBrowserBridgeTabs extends AgentBrowserBridgeState {return this.resolveActiveTab(onlyWorktreeId)}}++function routingKeys(worktreeId: string | undefined): (string | undefined)[] {+ return worktreeId ? [undefined, worktreeId] : [undefined]+}diff --git a/src/main/browser/agent-browser-bridge-test-harness.ts b/src/main/browser/agent-browser-bridge-test-harness.tsindex b1f91098095..1761b48890e 100644--- a/src/main/browser/agent-browser-bridge-test-harness.ts+++ b/src/main/browser/agent-browser-bridge-test-harness.ts@@ -42,6 +42,7 @@ export function mockBrowserManager(end: () => {}}),showAgentPointer: async () => {},+ isAgentUsingPage: () => false,...overrides} as unknown as BrowserManager}diff --git a/src/main/browser/browser-agent-activity.ts b/src/main/browser/browser-agent-activity.tsindex 5a681c917d3..7ee49ed8870 100644--- a/src/main/browser/browser-agent-activity.ts+++ b/src/main/browser/browser-agent-activity.ts@@ -105,6 +105,10 @@ export class BrowserAgentActivity {}}+ isActive(browserPageId: string): boolean {+ return this.pages.has(browserPageId)+ }+end(browserPageId: string): void {const page = this.pages.get(browserPageId)if (!page) {diff --git a/src/main/browser/browser-manager-visibility.ts b/src/main/browser/browser-manager-visibility.tsindex 951592284c2..837f4195b19 100644--- a/src/main/browser/browser-manager-visibility.ts+++ b/src/main/browser/browser-manager-visibility.ts@@ -22,6 +22,11 @@ export abstract class BrowserManagerVisibility extends BrowserManagerState {return this.agentActivity.beginCommand(browserPageId)}+ /** The page shows "An agent is using this page": a command is queued or one settled within the idle window. */+ isAgentUsingPage(browserPageId: string): boolean {+ return this.agentActivity.isActive(browserPageId)+ }+// Show the agent's pointer landing before the input it precedes reaches the page.showAgentPointer(guest: Electron.WebContents, input: AgentPointerInput): Promise<void> {return this.agentActivity.showPointer(guest, input)