feat(vom): add type primitive

.github/workflows/ci.yml

@@ -37,27 +37,6 @@ jobs:
       - name: Run Rust tests
         run: cargo test --workspace --locked
 
-  windows-daemon:
-    name: Windows daemon and update regression tests
-    runs-on: windows-latest
-
-    steps:
-      - uses: actions/checkout@v6
-
-      - uses: dtolnay/rust-toolchain@stable
-
-      - name: Run Windows named-pipe tests
-        run: cargo test -p bsk --lib --locked daemon::ipc::windows::tests
-
-      - name: Run Windows process liveness tests
-        run: cargo test -p bsk --lib --locked daemon::lockfile::tests
-
-      - name: Run Windows update tests
-        run: cargo test -p bsk --lib --locked cli::update::tests
-
-      - name: Run Windows update process tests
-        run: cargo test -p bsk --test windows_update --locked
-
   frontend:
     name: Frontend lint, typecheck, tests, build
     runs-on: ubuntu-latest

.gitignore

@@ -7,10 +7,6 @@ dist/
 .output/
 .wxt/
 .web-extension-id
-.pnpm-store/
-
-# Git worktrees
-.worktrees/
 
 # Logs
 *.log

CHANGELOG.md

@@ -1,92 +0,0 @@
-# Changelog
-
-All notable changes to BrowserSkill will be documented in this file.
-The format is based on [Keep a Changelog](https://keepachangelog.com/).
-
-Starting from 0.2.0, CLI / Extension / DSH Plugin share the same version number.
-
-## [Unreleased]
-
-## [0.2.0] - 2026-09-02
-
-### Added
-- File transfer: upload and download support across CLI, Extension, and DSH Plugin
-- File transfer: drag-and-drop upload (`drop-to-upload`)
-- VOM semantic graph, name enrichment, and hover perception modules
-- VOM hover probing (opt-in via `observe` parameter)
-- DSH Plugin: browser tool parity with CLI commands
-- Edge Add-ons automated publishing in CI
-- Protocol upgrade reminder when CLI / Extension protocol versions differ
-- Browser evaluation harness (`evals/browser/`)
-- Unified release script (`scripts/release.mjs`)
-
-### Changed
-- **Version scheme**: all three components now share the same semver
-- VOM rendering algorithm optimizations
-- Leaner SKILL.md agent instructions
-- DSH Plugin: simplified browser commands
-- Borrow confirmation UX — proactive focus and longer timeout
-- PiP window now has a close button
-
-### Fixed
-- Screenshot media type detection (was hard-coded to `image/png`)
-- DSH Plugin session lifecycle stability
-- DSH Plugin Cordis package ID mismatch
-- Observation thumbnail media type sniffing
-- Upload/download race conditions and layout bypass issues
-- VOM repeated name and safety policy issues
-
----
-
-*Previous releases used independent version numbers per component.*
-
-## CLI 0.1.11 / Extension 0.1.7 / DSH Plugin 0.1.2 — 2026-08-26 ~ 2026-08-29
-
-### Added
-- DSH Plugin sidebar integration, session lifecycle, and archive cleanup
-- Recorder iframe and OOPIF support
-
-### Changed
-- VOM functional refactor
-
-### Fixed
-- Recorder safety policy and bug fixes
-
-## CLI 0.1.10 / Extension 0.1.6 — 2026-08-08
-
-### Added
-- VOM observation recording and settled-state detection
-- Record overlay timer
-
-### Fixed
-- Browser keepalive disconnect handling
-
-## CLI 0.1.9 / Extension 0.1.5 — 2026-07-29
-
-### Added
-- CLI auto-update mechanism
-- Trace v3 protocol and recorder
-
-### Fixed
-- MV3 keepalive disconnect
-
-## CLI 0.1.8 / Extension 0.1.4 — 2026-07-22
-
-### Added
-- More browser interaction actions
-
-### Fixed
-- Windows named-pipe hash-only path issue
-
-## CLI 0.1.7 / Extension 0.1.3 — 2026-07-07
-
-Initial public release pair.
-
-## CLI 0.1.6 — 2026-06-30
-
-### Fixed
-- Minor CLI fixes
-
-## CLI 0.1.5 / Extension 0.1.2 — 2026-06-22
-
-First tagged releases.

Cargo.lock

@@ -134,10 +134,9 @@ dependencies = [
 
 [[package]]
 name = "bsk"
-version = "0.2.0"
+version = "0.1.11"
 dependencies = [
  "anyhow",
- "base64",
  "bsk-protocol",
  "clap",
  "console",
@@ -169,7 +168,7 @@ dependencies = [
 
 [[package]]
 name = "bsk-protocol"
-version = "0.2.0"
+version = "0.1.11"
 dependencies = [
  "schemars",
  "semver",

Cargo.toml

@@ -7,7 +7,7 @@ members = [
 default-members = ["crates/bsk-cli"]
 
 [workspace.package]
-version = "0.2.0"
+version = "0.1.11"
 edition = "2024"
 rust-version = "1.85"
 license = "MIT"
@@ -36,14 +36,13 @@ flate2 = "1.1.9"
 tar = "0.4.46"
 zip = { version = "7.2.0", default-features = false, features = ["deflate"] }
 sha2 = "0.11.0"
-base64 = "0.22"
 tempfile = "3"
 nix = { version = "0.29", features = ["signal", "process"] }
 rand = "0.8"
 uuid = { version = "1", features = ["v4", "serde"] }
 
 # Workspace crates (version required for `cargo publish` of dependents)
-bsk-protocol = { version = "0.2.0", path = "crates/bsk-protocol" }
+bsk-protocol = { version = "0.1.9", path = "crates/bsk-protocol" }
 
 [profile.release]
 strip = "symbols"

apps/extension/PRIVACY.md

@@ -1,6 +1,6 @@
 # BrowserSkill — Privacy Policy
 
-**Last updated:** August 20, 2026
+**Last updated:** May 25, 2026
 
 This Privacy Policy describes how the **BrowserSkill** browser extension (the "Extension") handles information when you install and use it. BrowserSkill is published as part of the open-source [BrowserSkill](https://github.com/Tencent/BrowserSkill) project. The source code is publicly auditable.
 
@@ -14,7 +14,7 @@ BrowserSkill is a local automation bridge that lets AI coding agents (such as Cu
 
 ## 2. Single Purpose
 
-The Extension's single purpose is to expose browser automation primitives (navigation, DOM observation, screenshots, clicks, form filling, task-scoped file transfer, and tab management) to a locally running BrowserSkill daemon over a WebSocket connection on `127.0.0.1`, so that an AI agent invoked by the user can interact with web pages on the user's behalf.
+The Extension's single purpose is to expose browser automation primitives (navigation, DOM observation, screenshots, clicks, form filling, tab management) to a locally running BrowserSkill daemon over a WebSocket connection on `127.0.0.1`, so that an AI agent invoked by the user can interact with web pages on the user's behalf.
 
 ## 3. Data the Extension Accesses
 
@@ -26,7 +26,6 @@ Depending on the commands the user (via their AI agent) sends to the local daemo
 | **User input simulated by the agent** | Mouse clicks, keystrokes, and form values that the AI agent dispatches through the Chrome DevTools Protocol (CDP). | Required to perform automation actions the user has asked the agent to do. |
 | **Tab and window metadata** | Tab IDs, URLs, titles, window IDs of the Agent Window and any tabs the user explicitly authorizes. | Required to target automation commands at the correct tab/window. |
 | **Local extension storage** | A randomly generated 8-character instance ID and an optional user-supplied label. | Used so the local daemon can recognize this browser instance across reconnects. No personal data is stored. |
-| **File transfers requested by the agent** | Local files explicitly supplied to `bsk upload`, and the file created by a single `bsk download` action. | Required to attach a task file to a web page or return a browser-generated download to the invoking local agent. |
 | **OS notifications** | Permission to display a system notification when the agent requests to "borrow" one of the user's existing tabs. | Required to obtain explicit, per-tab user consent before the agent touches any pre-existing tab. |
 
 ## 4. Data the Extension Does **Not** Collect
@@ -35,7 +34,7 @@ BrowserSkill does **not**:
 
 - Send any data to remote servers, the Extension's authors, or any third party.
 - Call any LLM, AI, or cloud API. The Extension contains no API keys, model identifiers, or remote endpoints.
-- Read or transmit cookies, browsing history, bookmarks, saved passwords, or autofill data. It observes only the download initiated by an active `bsk download` call, not download history generally.
+- Read or transmit cookies, browsing history, bookmarks, downloads, saved passwords, or autofill data.
 - Use webcam, microphone, geolocation, or any device sensor.
 - Include analytics, telemetry, crash reporting, advertising SDKs, or fingerprinting code.
 - Track users across websites or across sessions.
@@ -51,7 +50,6 @@ The Extension requests the following Chrome permissions. Each is used solely for
 - **`alarms`** — Periodically wake the service worker to keep the local WebSocket connection alive.
 - **`idle`** — Detect when the device returns from idle/locked so the Extension can promptly re-establish the local WebSocket connection after the machine wakes. No idle data is stored or transmitted.
 - **`notifications`** — Show a system notification to obtain user approval before the agent borrows a user-owned tab.
-- **`downloads`** — Correlate and route the one browser download initiated by an active `bsk download` command. If that claimed transaction fails, BrowserSkill cancels an in-progress file or removes its completed temporary browser file. It is not used to enumerate download history or alter unclaimed downloads.
 - **`storage`** — Persist a random instance ID and optional label in `chrome.storage.local`.
 - **Host permission `<all_urls>`** — Inject a small status overlay (showing "Agent Active") on pages controlled by the agent, and enable automation across whatever sites the user directs the agent to. The Extension does **not** read or transmit page content from sites the agent is not actively driving.
 
@@ -63,7 +61,6 @@ All Extension activity stays on the user's local device. The only network traffi
 
 - The instance ID and optional label persist in `chrome.storage.local` until the user uninstalls the Extension or clears extension storage.
 - Page content, screenshots, DOM snapshots, and other observed data are returned to the local daemon in response to commands and are **not retained by the Extension**. They live only as long as the agent's tool call.
-- Upload and download bytes are staged by the local daemon in a private, session-scoped directory. Download staging is removed after it is copied to the requested destination. Upload staging is retained until the session ends so a later form submission can still read the attached file. Remaining staging is removed when the session ends or disconnects, or when the daemon next starts after a crash.
 
 ## 8. User Control
 

apps/extension/e2e/fixtures/file-transfer.html

@@ -1,81 +0,0 @@
-<!doctype html>
-<html lang="en">
-  <head>
-    <meta charset="utf-8" />
-    <title>BrowserSkill file-transfer fixture</title>
-    <style>
-      body { font: 16px system-ui; max-width: 760px; margin: 48px auto; }
-      .actions { display: flex; gap: 12px; align-items: center; margin: 24px 0; }
-      button, a { padding: 10px 16px; border: 1px solid #777; border-radius: 8px; }
-      a[aria-disabled="true"] { opacity: .45; pointer-events: none; }
-      #drop { padding: 36px; border: 3px dashed #777; border-radius: 12px; text-align: center; }
-      #drop.dragging { border-color: #2563eb; background: #eff6ff; }
-      img { display: block; max-width: 480px; max-height: 360px; margin-top: 20px; }
-    </style>
-  </head>
-  <body>
-    <h1>Upload then download the same image</h1>
-    <p id="status">No image selected</p>
-    <div class="actions">
-      <button id="upload" type="button">Upload through file input</button>
-      <input id="file" type="file" accept="image/png" hidden />
-      <a id="download" aria-disabled="true">Download uploaded image</a>
-    </div>
-    <div id="drop" role="button" tabindex="0">Drop an image here</div>
-    <img id="preview" alt="Uploaded image preview" hidden />
-    <script>
-      const input = document.querySelector('#file');
-      const upload = document.querySelector('#upload');
-      const drop = document.querySelector('#drop');
-      const download = document.querySelector('#download');
-      const preview = document.querySelector('#preview');
-      const status = document.querySelector('#status');
-      let objectUrl;
-
-      window.__bskTransferFixture = { events: [], result: null };
-
-      function accept(file, mechanism) {
-        if (!file) return;
-        if (objectUrl) URL.revokeObjectURL(objectUrl);
-        objectUrl = URL.createObjectURL(file);
-        preview.src = objectUrl;
-        preview.hidden = false;
-        download.href = objectUrl;
-        download.download = file.name;
-        download.removeAttribute('aria-disabled');
-        status.textContent = `Ready via ${mechanism}: ${file.name} (${file.size} bytes)`;
-        window.__bskTransferFixture.result = {
-          mechanism,
-          name: file.name,
-          size: file.size,
-          type: file.type,
-          hasPathProperty: 'path' in file,
-        };
-      }
-
-      upload.addEventListener('click', () => input.click());
-      input.addEventListener('change', () => accept(input.files[0], 'input'));
-
-      for (const type of ['dragenter', 'dragover']) {
-        drop.addEventListener(type, event => {
-          const types = [...event.dataTransfer.types];
-          window.__bskTransferFixture.events.push({ type, types });
-          if (!types.includes('Files')) return;
-          event.preventDefault();
-          drop.classList.add('dragging');
-        });
-      }
-      drop.addEventListener('dragleave', () => drop.classList.remove('dragging'));
-      drop.addEventListener('drop', event => {
-        window.__bskTransferFixture.events.push({
-          type: 'drop',
-          types: [...event.dataTransfer.types],
-          fileCount: event.dataTransfer.files.length,
-        });
-        event.preventDefault();
-        drop.classList.remove('dragging');
-        accept(event.dataTransfer.files[0], 'drop');
-      });
-    </script>
-  </body>
-</html>

apps/extension/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@browser-skill/extension",
-  "version": "0.2.0",
+  "version": "0.1.7",
   "private": true,
   "type": "module",
   "scripts": {

apps/extension/src/browser-driver/__tests__/chromium-cdp.test.ts

@@ -254,69 +254,6 @@ describe("ChromiumCdp", () => {
     expect(cdp.isAttached(7)).toBe(false);
   });
 
-  it.each([
-    false,
-    true,
-  ])("waits for detach before reconnecting a returned tab (detach rejects: %s)", async (rejectDetach) => {
-    const { api } = fakeApi();
-    const attached = new Set<number>();
-    let finishDetach!: () => void;
-    vi.mocked(api.attach).mockImplementation(async ({ tabId }) => {
-      if (attached.has(tabId!)) throw new Error("Another debugger is already attached");
-      attached.add(tabId!);
-    });
-    vi.mocked(api.detach).mockImplementation(async ({ tabId }) => {
-      attached.delete(tabId!);
-    });
-    vi.mocked(api.detach).mockImplementationOnce(
-      ({ tabId }) =>
-        new Promise<void>((resolve, reject) => {
-          finishDetach = () => {
-            attached.delete(tabId!);
-            if (rejectDetach) reject(new Error("tab already detached"));
-            else resolve();
-          };
-        }),
-    );
-    const cdp = new ChromiumCdp(api);
-    cdp.trackSessionTab("aa11", 7);
-    await cdp.ensureAttached(7);
-    const returning = cdp.releaseSessionTab("aa11", 7);
-    await vi.waitFor(() => expect(api.detach).toHaveBeenCalledOnce());
-    let duplicateDetachFinished = false;
-    const duplicateDetach = cdp.detach(7).then(() => {
-      duplicateDetachFinished = true;
-    });
-
-    cdp.trackSessionTab("bb22", 7);
-    // Register rejection handlers immediately; the pre-fix driver rejects
-    // both commands while the browser still has the previous attachment.
-    const pending = Promise.allSettled([
-      cdp.send(7, "Runtime.evaluate", { expression: "document.title" }),
-      cdp.send(7, "DOM.getDocument"),
-    ]);
-    await cdp.send(8, "DOM.getDocument");
-    const finishedEarly = duplicateDetachFinished;
-    finishDetach();
-    await Promise.all([returning, duplicateDetach]);
-
-    expect(await pending).toEqual([
-      { status: "fulfilled", value: { ok: true } },
-      { status: "fulfilled", value: { ok: true } },
-    ]);
-    expect(finishedEarly).toBe(false);
-    expect(vi.mocked(api.attach).mock.calls.filter(([target]) => target.tabId === 7)).toHaveLength(
-      2,
-    );
-    expect(cdp.isAttached(7)).toBe(true);
-    expect(cdp.isAttached(8)).toBe(true);
-    await cdp.detachSession("bb22");
-    expect(api.detach).toHaveBeenCalledTimes(2);
-    expect(cdp.isAttached(7)).toBe(false);
-    expect(attached.has(7)).toBe(false);
-    expect(cdp.isAttached(8)).toBe(true);
-  });
-
   it("detachAll() iterates every cached tab", async () => {
     const { api } = fakeApi();
     const cdp = new ChromiumCdp(api);
@@ -714,155 +651,6 @@ describe("ChromiumCdp", () => {
     expect(cdp.networkEntriesSince(33, 0, 50, 1000).entries[0].url).toBeUndefined();
   });
 
-  it("releases only the returning session's claim on a tab", async () => {
-    const { api } = fakeApi();
-    const cdp = new ChromiumCdp(api);
-    await cdp.ensureAttached(1);
-    await cdp.ensureAttached(2);
-    cdp.trackSessionTab("aa11", 1);
-    cdp.trackSessionTab("bb22", 1);
-    cdp.trackSessionTab("aa11", 2);
-
-    await cdp.releaseSessionTab("unknown", 1);
-    await cdp.releaseSessionTab("aa11", 1);
-    expect(api.detach).not.toHaveBeenCalled();
-
-    await cdp.releaseSessionTab("bb22", 1);
-    await cdp.releaseSessionTab("bb22", 1);
-    expect(api.detach).toHaveBeenCalledExactlyOnceWith({ tabId: 1 });
-    expect(cdp.isAttached(2)).toBe(true);
-
-    await cdp.detachSession("aa11");
-    expect(api.detach).toHaveBeenCalledTimes(2);
-    expect(api.detach).toHaveBeenLastCalledWith({ tabId: 2 });
-  });
-
-  it.each([
-    false,
-    true,
-  ])("releases a tab with an attachment in flight (new owner: %s)", async (newOwner) => {
-    const { api } = fakeApi();
-    let finishAttach!: () => void;
-    vi.mocked(api.attach).mockImplementationOnce(
-      () =>
-        new Promise<void>((resolve) => {
-          finishAttach = resolve;
-        }),
-    );
-    const cdp = new ChromiumCdp(api);
-    cdp.trackSessionTab("aa11", 1);
-    const attaching = cdp.ensureAttached(1);
-    const releasing = cdp.releaseSessionTab("aa11", 1);
-    if (newOwner) cdp.trackSessionTab("bb22", 1);
-    finishAttach();
-    await Promise.all([attaching, releasing]);
-
-    expect(cdp.isAttached(1)).toBe(newOwner);
-    expect(api.detach).toHaveBeenCalledTimes(newOwner ? 0 : 1);
-  });
-
-  it("stops accepting dialogs on a returned tab even when another session keeps CDP attached", async () => {
-    const { api, onEvent } = fakeApi();
-    let controlled = true;
-    const cdp = new ChromiumCdp(api, { shouldAutoAcceptDialog: () => controlled });
-    cdp.trackSessionTab("aa11", 1);
-    cdp.trackSessionTab("bb22", 1);
-    await cdp.ensureAttached(1);
-    onEvent.fire({ tabId: 1 }, "Page.javascriptDialogOpening", {
-      type: "confirm",
-      message: "before return",
-    });
-    await vi.waitFor(() => expect(cdp.dialogsSince(1, 0)).toHaveLength(1));
-    vi.mocked(api.sendCommand).mockClear();
-
-    controlled = false;
-    await cdp.releaseSessionTab("aa11", 1);
-    expect(cdp.isAttached(1)).toBe(true);
-    onEvent.fire({ tabId: 1 }, "Page.javascriptDialogOpening", {
-      type: "confirm",
-      message: "after return",
-    });
-    await Promise.resolve();
-    expect(api.sendCommand).not.toHaveBeenCalled();
-    expect(cdp.dialogsSince(1, 0)).toHaveLength(1);
-
-    controlled = true;
-    onEvent.fire({ tabId: 1 }, "Page.javascriptDialogOpening", {
-      type: "alert",
-      message: "borrowed again",
-    });
-    await vi.waitFor(() => expect(cdp.dialogsSince(1, 0)).toHaveLength(2));
-    expect(api.sendCommand).toHaveBeenCalledExactlyOnceWith(
-      { tabId: 1 },
-      "Page.handleJavaScriptDialog",
-      { accept: true },
-    );
-  });
-
-  it("ignores dialog events after detach, including a pending eligibility check", async () => {
-    const { api, onEvent } = fakeApi();
-    let resolveEligibility!: (value: boolean) => void;
-    const shouldAutoAcceptDialog = vi.fn(
-      () =>
-        new Promise<boolean>((resolve) => {
-          resolveEligibility = resolve;
-        }),
-    );
-    const cdp = new ChromiumCdp(api, { shouldAutoAcceptDialog });
-    await cdp.ensureAttached(1);
-    onEvent.fire({ tabId: 1 }, "Page.javascriptDialogOpening", { type: "alert" });
-    await cdp.detach(1);
-    resolveEligibility(true);
-    await Promise.resolve();
-    onEvent.fire({ tabId: 1 }, "Page.javascriptDialogOpening", { type: "alert" });
-
-    expect(shouldAutoAcceptDialog).toHaveBeenCalledOnce();
-    expect(api.sendCommand).not.toHaveBeenCalledWith(
-      expect.anything(),
-      "Page.handleJavaScriptDialog",
-      expect.anything(),
-    );
-    expect(cdp.dialogsSince(1, 0)).toEqual([]);
-  });
-
-  it("does not accept dialogs when the tab's current scope cannot be determined", async () => {
-    const { api, onEvent } = fakeApi();
-    const shouldAutoAcceptDialog = vi.fn(async () => {
-      throw new Error("tab is gone");
-    });
-    const cdp = new ChromiumCdp(api, { shouldAutoAcceptDialog });
-    await cdp.ensureAttached(1);
-    vi.mocked(api.sendCommand).mockClear();
-
-    onEvent.fire({ tabId: 1 }, "Page.javascriptDialogOpening", { type: "confirm" });
-    await Promise.resolve();
-
-    expect(shouldAutoAcceptDialog).toHaveBeenCalledWith(1);
-    expect(api.sendCommand).not.toHaveBeenCalled();
-    expect(cdp.dialogsSince(1, 0)).toEqual([]);
-  });
-
-  it("does not restore dialog state when an acceptance completes after detach", async () => {
-    const { api, onEvent } = fakeApi();
-    const cdp = new ChromiumCdp(api);
-    await cdp.ensureAttached(1);
-    let finishAccept!: () => void;
-    vi.mocked(api.sendCommand).mockImplementationOnce(
-      () =>
-        new Promise<void>((resolve) => {
-          finishAccept = resolve;
-        }),
-    );
-
-    onEvent.fire({ tabId: 1 }, "Page.javascriptDialogOpening", { type: "alert" });
-    await cdp.detach(1);
-    finishAccept();
-    await Promise.resolve();
-
-    expect(cdp.dialogsSince(1, 0)).toEqual([]);
-    expect(cdp.dialogCursor(1)).toBe(0);
-  });
-
   it("detachSession only detaches tabs no other session owns", async () => {
     const { api } = fakeApi();
     const cdp = new ChromiumCdp(api);

apps/extension/src/browser-driver/chromium-cdp.ts

@@ -165,7 +165,6 @@ export class ChromiumCdp {
   private readonly api: CdpDebuggerApi;
   private readonly attachedTabs = new Set<number>();
   private readonly attachInFlight = new Map<number, Promise<void>>();
-  private readonly detachInFlight = new Map<number, Promise<void>>();
   private readonly tabOwners = new Map<number, Set<string>>();
   private readonly dialogBuffers = new Map<number, JavaScriptDialogInfo[]>();
   private readonly dialogSequences = new Map<number, number>();
@@ -183,13 +182,7 @@ export class ChromiumCdp {
   private networkSubscription: { dispose(): void } | null = null;
   private frameTargetSubscription: { dispose(): void } | null = null;
 
-  constructor(
-    api: CdpDebuggerApi = chromeDebuggerApi,
-    private readonly options: {
-      /** CDP observation alone does not authorize dismissing user dialogs. */
-      shouldAutoAcceptDialog?: (tabId: number) => boolean | Promise<boolean>;
-    } = {},
-  ) {
+  constructor(api: CdpDebuggerApi = chromeDebuggerApi) {
     this.api = api;
     this.bindAutoDetach();
     this.bindDialogHandler();
@@ -200,10 +193,6 @@ export class ChromiumCdp {
 
   /** Attach to `tabId` if we haven't already in this driver. */
   async ensureAttached(tabId: number): Promise<void> {
-    // Returning a tab clears the cache before Chrome finishes detaching.
-    // New observers must wait before opening the next connection to that tab.
-    const detaching = this.detachInFlight.get(tabId);
-    if (detaching) await detaching;
     if (this.attachedTabs.has(tabId)) return;
     const existing = this.attachInFlight.get(tabId);
     if (existing) {
@@ -436,11 +425,6 @@ export class ChromiumCdp {
 
   /** Detach if attached; never throws. */
   async detach(tabId: number): Promise<void> {
-    const existing = this.detachInFlight.get(tabId);
-    if (existing) {
-      await existing;
-      return;
-    }
     this.attachInFlight.delete(tabId);
     if (!this.attachedTabs.has(tabId)) return;
     this.attachedTabs.delete(tabId);
@@ -448,19 +432,13 @@ export class ChromiumCdp {
     this.clearConsoleState(tabId);
     this.clearNetworkState(tabId);
     this.clearFrameState(tabId);
-    const detach = (async () => {
-      try {
-        await this.api.detach({ tabId });
-      } catch (err) {
-        // Tab may already be gone — Chrome auto-detaches on close. Log
-        // at debug so production builds aren't noisy.
-        console.debug("[bsk cdp] detach failed (likely tab already closed)", err);
-      }
-    })().finally(() => {
-      this.detachInFlight.delete(tabId);
-    });
-    this.detachInFlight.set(tabId, detach);
-    await detach;
+    try {
+      await this.api.detach({ tabId });
+    } catch (err) {
+      // Tab may already be gone — Chrome auto-detaches on close. Log
+      // at debug so production builds aren't noisy.
+      console.debug("[bsk cdp] detach failed (likely tab already closed)", err);
+    }
   }
 
   /** True iff `ensureAttached(tabId)` has succeeded since the last detach. */
@@ -475,17 +453,6 @@ export class ChromiumCdp {
     this.tabOwners.set(tabId, owners);
   }
 
-  /** Release one session's claim, preserving attachments still used by another. */
-  async releaseSessionTab(sessionId: string, tabId: number): Promise<void> {
-    const owners = this.tabOwners.get(tabId);
-    if (!owners?.delete(sessionId) || owners.size > 0) return;
-    this.tabOwners.delete(tabId);
-    // An observation may still be attaching when the tab is returned. Wait
-    // for it so detach cannot miss the attachment or remove a new owner's claim.
-    await this.attachInFlight.get(tabId)?.catch(() => {});
-    if (!this.tabOwners.has(tabId)) await this.detach(tabId);
-  }
-
   /** Subscribe to all CDP events. Returned disposable removes the listener. */
   onEvent(handler: (source: CdpDebuggee, method: string, params: unknown) => void): {
     dispose(): void;
@@ -672,23 +639,13 @@ export class ChromiumCdp {
   }
 
   private async onJavaScriptDialogOpening(tabId: number, params: unknown): Promise<void> {
-    if (!this.attachedTabs.has(tabId) && !this.attachInFlight.has(tabId)) return;
     const parsed = parseDialogOpeningParams(params);
     try {
-      if (
-        this.options.shouldAutoAcceptDialog &&
-        !(await this.options.shouldAutoAcceptDialog(tabId))
-      ) {
-        return;
-      }
-      // The tab may have been returned while its current scope was checked.
-      if (!this.attachedTabs.has(tabId) && !this.attachInFlight.has(tabId)) return;
       const handleParams: { accept: boolean; promptText?: string } = { accept: true };
       if (parsed.type === "prompt") {
         handleParams.promptText = parsed.defaultPrompt ?? "";
       }
       await this.api.sendCommand({ tabId }, "Page.handleJavaScriptDialog", handleParams);
-      if (!this.attachedTabs.has(tabId) && !this.attachInFlight.has(tabId)) return;
       const sequence = (this.dialogSequences.get(tabId) ?? 0) + 1;
       this.dialogSequences.set(tabId, sequence);
       this.appendDialog(tabId, {
@@ -878,9 +835,15 @@ export class ChromiumCdp {
 
   /** Detach tabs only when no other live session has claimed them. */
   async detachSession(sessionId: string): Promise<void> {
-    await Promise.all(
-      Array.from(this.tabOwners.keys(), (tabId) => this.releaseSessionTab(sessionId, tabId)),
-    );
+    const tabsToDetach: number[] = [];
+    for (const [tabId, owners] of this.tabOwners) {
+      owners.delete(sessionId);
+      if (owners.size === 0) {
+        this.tabOwners.delete(tabId);
+        tabsToDetach.push(tabId);
+      }
+    }
+    await Promise.all(tabsToDetach.map((tabId) => this.detach(tabId)));
   }
 }
 

apps/extension/src/content/__tests__/BorrowConfirmationOverlay.test.tsx

@@ -23,7 +23,7 @@ describe("BorrowConfirmationOverlay", () => {
             id: "req-1",
             isActiveTab: true,
             tabTitle: "Example",
-            timeoutMs: 60_000,
+            timeoutMs: 5000,
             onAllow,
             onDeny,
           },
@@ -32,7 +32,7 @@ describe("BorrowConfirmationOverlay", () => {
     );
 
     expect(screen.getByText("允许借用标签页？")).toBeTruthy();
-    for (let i = 0; i < 60; i++) {
+    for (let i = 0; i < 5; i++) {
       await act(async () => {
         await vi.advanceTimersByTimeAsync(1000);
       });
@@ -66,7 +66,7 @@ describe("BorrowConfirmationOverlay", () => {
             id: "req-fallback",
             isActiveTab: true,
             tabTitle: "Example",
-            timeoutMs: 60_000,
+            timeoutMs: 5000,
             onAllow,
             onDeny,
           },
@@ -74,7 +74,7 @@ describe("BorrowConfirmationOverlay", () => {
       />,
     );
 
-    for (let i = 0; i < 60; i++) {
+    for (let i = 0; i < 5; i++) {
       await act(async () => {
         await vi.advanceTimersByTimeAsync(1000);
       });

apps/extension/src/content/__tests__/record-frame-agent.test.ts

@@ -45,32 +45,10 @@ function portHarness() {
 
 describe("RecordFrameAgent", () => {
   afterEach(() => {
-    vi.unstubAllGlobals();
     document.documentElement.removeAttribute(RECORD_DOCUMENT_ATTRIBUTE);
     document.body.replaceChildren();
   });
 
-  it("starts when crypto.randomUUID is unavailable", async () => {
-    const harness = portHarness();
-    const getRandomValues = vi.fn((bytes: Uint8Array) => {
-      bytes.fill(0x11);
-      return bytes;
-    });
-    vi.stubGlobal("crypto", { getRandomValues });
-    vi.stubGlobal("chrome", {
-      runtime: {
-        connect: vi.fn(() => harness.port),
-        sendMessage: vi.fn(),
-      },
-    });
-    const agent = new RecordFrameAgent();
-
-    await expect(
-      agent.start({ type: RECORD_FRAME_START, requestId: "rec-http", startedAtMs: 10 }),
-    ).resolves.toEqual({ ok: true });
-    expect(getRandomValues).toHaveBeenCalled();
-  });
-
   it("keeps a failed stop retryable and flushes the final dirty fill", async () => {
     const harness = portHarness();
     const sendMessage = vi

apps/extension/src/content/record-step-delivery.ts

@@ -1,4 +1,3 @@
-import { createRandomUuid } from "@/lib/random-uuid";
 import {
   isAcceptedRecordStepAck,
   RECORD_STEP,
@@ -19,7 +18,7 @@ export class RecordStepDelivery {
   constructor(
     requestId: string,
     send: SendRecordStepMessage = (message) => chrome.runtime.sendMessage(message),
-    producerId: string = createRandomUuid(),
+    producerId: string = crypto.randomUUID(),
   ) {
     this.#requestId = requestId;
     this.#producerId = producerId;

apps/extension/src/content/recording/frame-agent.ts

@@ -1,4 +1,3 @@
-import { createRandomUuid } from "@/lib/random-uuid";
 import {
   markRecordingDocument,
   type RecordingDocumentMarker,
@@ -68,7 +67,7 @@ export class RecordFrameAgent {
   async #start(
     message: RecordFrameStartMessage,
   ): Promise<{ ok: true } | { ok: false; error: string }> {
-    const producerId = createRandomUuid();
+    const producerId = crypto.randomUUID();
     const root = await waitForDocumentElement();
     if (this.#disposed) return { ok: false, error: "recording document is disposed" };
     const marker = markRecordingDocument(producerId, root);

apps/extension/src/entrypoints/background.ts

@@ -25,7 +25,7 @@ import { recordFrameCoordinator } from "@/lib/recording/frame-coordinator";
 import { attachSessionsLiveFlag } from "@/lib/sessions-live-flag";
 import { createDisconnectCleanup } from "@/session-manager/disconnect-cleanup";
 import { attachSessionEventHandler } from "@/session-manager/event-handler";
-import { isAgentControlledTab, SessionManager } from "@/session-manager/manager";
+import { SessionManager } from "@/session-manager/manager";
 import {
   attachBorrowNotificationButtonHandler,
   attachBorrowNotificationClickHandler,
@@ -41,8 +41,6 @@ import {
   attachRecordStepListener,
   type RecordRuntimeDeps,
 } from "@/tools/record";
-import { chromeTabsApi } from "@/tools/shared";
-import { chromeTabMutationApi } from "@/tools/tabs";
 import { detectBrowserMeta } from "@/transport/handshake";
 import type { Transport } from "@/transport/transport";
 import { WSTransport } from "@/transport/ws-transport";
@@ -51,12 +49,7 @@ export default defineBackground(() => {
   const controller = new ConnectionController();
   const transport = new WSTransport({ url: __BSK_DAEMON_WS_URL__ });
   const sessions = new SessionManager();
-  const cdp = new ChromiumCdp(undefined, {
-    shouldAutoAcceptDialog: async (tabId) => {
-      const tab = await chrome.tabs.get(tabId);
-      return sessions.findByWindowId(tab.windowId) !== null;
-    },
-  });
+  const cdp = new ChromiumCdp();
   const sessionsLive = attachSessionsLiveFlag({ manager: sessions });
   let overlayGeneration = 0;
   const controlModes = new Map<string, OverlayMode>();
@@ -87,24 +80,6 @@ export default defineBackground(() => {
     };
   }
 
-  /**
-   * Authoritative overlay state for a specific tab. Agent Window tabs are
-   * free by default; only tabs explicitly claimed through session startup,
-   * `tab_create`, or `tab_borrow` receive the control overlay.
-   */
-  function overlayStateForTab(tabId?: number, windowId?: number): OverlayAgentStateMessage {
-    if (typeof tabId === "number" && typeof windowId === "number") {
-      const ctx = sessions.findByWindowId(windowId);
-      if (ctx && isAgentControlledTab(ctx, tabId)) return overlayStateForWindow(windowId);
-    }
-    return {
-      type: OVERLAY_AGENT_STATE,
-      sessionId: null,
-      mode: "hidden",
-      generation: overlayGeneration,
-    };
-  }
-
   async function pushOverlayStateToTab(
     tabId: number,
     state: OverlayAgentStateMessage,
@@ -116,18 +91,13 @@ export default defineBackground(() => {
     }
   }
 
-  async function pushOverlayStateForTab(tabId: number, windowId?: number): Promise<void> {
-    const state = overlayStateForTab(tabId, windowId);
-    await pushOverlayStateToTab(tabId, state);
-  }
-
   async function pushOverlayStateForWindow(windowId: number): Promise<void> {
+    const state = overlayStateForWindow(windowId);
     const tabs = await chrome.tabs.query({ windowId });
     await Promise.all(
-      tabs.map((tab) => {
-        if (typeof tab.id !== "number") return Promise.resolve();
-        return pushOverlayStateForTab(tab.id, windowId);
-      }),
+      tabs.map((tab) =>
+        typeof tab.id === "number" ? pushOverlayStateToTab(tab.id, state) : Promise.resolve(),
+      ),
     );
   }
 
@@ -166,30 +136,13 @@ export default defineBackground(() => {
     if (typeof tab.windowId !== "number") return;
     pushOverlayStateForAgentWindow(tab.windowId);
   });
-  // Newly observed tabs are free unless the agent's own creation path has
-  // already claimed their concrete id. This listener only renders that
-  // state; it never infers or mutates ownership from event ordering.
-  chrome.tabs.onCreated.addListener((tab) => {
-    if (typeof tab.windowId !== "number" || typeof tab.id !== "number") return;
-    if (!sessions.findByWindowId(tab.windowId)) return;
-    void pushOverlayStateForTab(tab.id, tab.windowId);
-  });
-  chrome.tabs.onRemoved.addListener((tabId, removeInfo) => {
-    sessions.forgetClosedTab(tabId, { isWindowClosing: removeInfo.isWindowClosing });
-  });
   // Re-sync the storage.session flag on SW startup so a previous SW's
   // stale `true` does not keep waking us on every page load until the
   // first mutation (review M4/M5 round 3 m-R3-1).
   void sessionsLive.refresh();
   const cleanupAfterDisconnect = createDisconnectCleanup({
     manager: sessions,
-    // Same contract as the dispatcher's `tool.session_stop`: without these
-    // deps the agent-tab cleanup and the window-release path are dead code.
-    sessionStopDeps: {
-      cdp,
-      tabManagement: { tabs: chromeTabMutationApi },
-      tabsQuery: chromeTabsApi,
-    },
+    sessionStopDeps: { cdp },
     onSessionsChanged: () => {
       void sessionsLive.syncFromManager();
     },
@@ -223,9 +176,6 @@ export default defineBackground(() => {
     recording: recordDeps,
     onSessionsChanged: onOverlaySessionStateChanged,
     onBrowserControlResumed,
-    onAgentTabClaimed: (tabId, windowId) => {
-      void pushOverlayStateForTab(tabId, windowId);
-    },
     approveBorrow: (ctx) =>
       requestBorrowConfirmation(ctx.tabId, {
         ...(ctx.signal !== undefined ? { signal: ctx.signal } : {}),
@@ -359,11 +309,7 @@ export default defineBackground(() => {
     }
 
     if (msg.kind === OVERLAY_MSG_READY) {
-      // Resolve from explicit tab ownership before sending so free tabs never
-      // flash the control mask.
-      if (typeof sender.tab?.id === "number") {
-        void pushOverlayStateForTab(sender.tab.id, sender.tab.windowId);
-      }
+      sendResponse(overlayStateForWindow(sender.tab?.windowId));
       return false;
     }
 

apps/extension/src/entrypoints/popup/App.test.tsx

@@ -122,8 +122,8 @@ describe("App", () => {
 
     expect(screen.queryByText(/^扩展 v/)).toBeNull();
     expect(screen.queryByText(/^daemon v/)).toBeNull();
-    expect(screen.getByTitle("扩展版本").textContent).toBe(`Ext ${EXTENSION_VERSION}`);
-    expect(screen.getByTitle("bsk 版本").textContent).toBe(`CLI ${mockDaemonVersion}`);
+    expect(screen.getByTitle("扩展版本").textContent).toBe(EXTENSION_VERSION);
+    expect(screen.getByTitle("bsk 版本").textContent).toBe(mockDaemonVersion);
     expect(screen.getByText("03c3e47f")).toBeTruthy();
 
     const copyButton = screen.getByRole("button", { name: "复制实例 ID" });

apps/extension/src/entrypoints/popup/App.tsx

@@ -267,9 +267,9 @@ export function App() {
             data-slot="popup-meta"
           >
             <div className="flex shrink-0 items-center gap-1">
-              <span title={t("popup.daemonVersionHint")}>CLI {daemonVersion}</span>
+              <span title={t("popup.extensionVersionHint")}>{extensionVersion}</span>
               <span aria-hidden>/</span>
-              <span title={t("popup.extensionVersionHint")}>Ext {extensionVersion}</span>
+              <span title={t("popup.daemonVersionHint")}>{daemonVersion}</span>
             </div>
             <div className="flex min-w-0 items-center justify-end gap-1">
               <span className="shrink-0">{t("popup.instanceTitle")}</span>

apps/extension/src/lib/__tests__/capture-suppress-bridge.test.ts

@@ -2,7 +2,7 @@ import { describe, expect, it, vi } from "vitest";
 import {
   CAPTURE_SUPPRESS,
   type CaptureSuppressMessage,
-  withExtensionOverlayHidden,
+  withOverlaysHiddenForCapture,
 } from "../capture-suppress-bridge";
 
 function recordingSendToTab(events: string[]) {
@@ -12,12 +12,12 @@ function recordingSendToTab(events: string[]) {
   });
 }
 
-describe("withExtensionOverlayHidden", () => {
+describe("withOverlaysHiddenForCapture", () => {
   it("sends begin → fn → end and returns fn's result", async () => {
     const events: string[] = [];
     const sendToTab = recordingSendToTab(events);
 
-    const result = await withExtensionOverlayHidden(
+    const result = await withOverlaysHiddenForCapture(
       7,
       async () => {
         events.push("capture");
@@ -37,7 +37,7 @@ describe("withExtensionOverlayHidden", () => {
     const sendToTab = recordingSendToTab(events);
 
     await expect(
-      withExtensionOverlayHidden(
+      withOverlaysHiddenForCapture(
         7,
         async () => {
           events.push("capture");
@@ -55,7 +55,7 @@ describe("withExtensionOverlayHidden", () => {
     });
     const fn = vi.fn(async () => "shot");
 
-    const result = await withExtensionOverlayHidden(7, fn, sendToTab);
+    const result = await withOverlaysHiddenForCapture(7, fn, sendToTab);
 
     expect(result).toBe("shot");
     expect(fn).toHaveBeenCalledTimes(1);
@@ -69,7 +69,7 @@ describe("withExtensionOverlayHidden", () => {
       return { type: CAPTURE_SUPPRESS, ok: true };
     });
 
-    const result = await withExtensionOverlayHidden(7, async () => "shot", sendToTab);
+    const result = await withOverlaysHiddenForCapture(7, async () => "shot", sendToTab);
 
     expect(result).toBe("shot");
     expect(sendToTab).toHaveBeenCalledTimes(2);

apps/extension/src/lib/__tests__/random-uuid.test.ts

@@ -1,30 +0,0 @@
-import { afterEach, describe, expect, it, vi } from "vitest";
-import { createRandomUuid } from "../random-uuid";
-
-describe("createRandomUuid", () => {
-  afterEach(() => {
-    vi.unstubAllGlobals();
-  });
-
-  it("uses crypto.randomUUID when it is available", () => {
-    const randomUUID = vi.fn(() => "00000000-0000-4000-8000-000000000000");
-    vi.stubGlobal("crypto", { randomUUID, getRandomValues: vi.fn() });
-
-    expect(createRandomUuid()).toBe("00000000-0000-4000-8000-000000000000");
-    expect(randomUUID).toHaveBeenCalledOnce();
-  });
-
-  it("falls back to getRandomValues and sets UUID v4 bits", () => {
-    const bytes = new Uint8Array(16).fill(0xff);
-    const getRandomValues = vi.fn((target: Uint8Array) => {
-      target.set(bytes);
-      return target;
-    });
-    vi.stubGlobal("crypto", { getRandomValues });
-
-    const uuid = createRandomUuid();
-
-    expect(uuid).toBe("ffffffff-ffff-4fff-bfff-ffffffffffff");
-    expect(getRandomValues).toHaveBeenCalledOnce();
-  });
-});

apps/extension/src/lib/capture-suppress-bridge.ts

@@ -1,9 +1,9 @@
 /**
- * Wire protocol and background-side helper for temporarily excluding the
- * extension's in-page overlay. Screenshots use it to avoid captured chrome;
- * coordinate-driven native input uses it so Chrome hit-tests the page.
+ * Wire protocol and background-side helper for hiding the extension's
+ * in-page overlay while a screenshot is captured, so captured frames only
+ * contain the page itself.
  *
- * The background wraps each bounded operation in `withExtensionOverlayHidden`,
+ * The background wraps every capture in `withOverlaysHiddenForCapture`,
  * which sends `begin` (the content script hides the overlay host and only
  * acks once the compositor has produced an overlay-free frame) and always
  * follows up with `end`, even when the capture throws. Tabs without the
@@ -46,7 +46,7 @@ const defaultSendToTab: CaptureSuppressSendToTab = (tabId, message) =>
  * (or any other `begin` failure) simply skips suppression — there is no
  * overlay to hide in that tab.
  */
-export async function withExtensionOverlayHidden<T>(
+export async function withOverlaysHiddenForCapture<T>(
   tabId: number,
   fn: () => Promise<T>,
   sendToTab: CaptureSuppressSendToTab = defaultSendToTab,

apps/extension/src/lib/overlay-bridge.ts

@@ -3,10 +3,8 @@
  * between the content-script control overlay and the background SW.
  *
  * Content script → background:
- *  - `{ kind: "overlay.ready" }` → background decides the per-tab state first
- *      (see overlayStateForTab) and pushes the authoritative overlay state for
- *      that tab. It also proactively pushes on tab create and control-mode
- *      change, so a tab usually receives state before it ever sends ready.
+ *  - `{ kind: "overlay.ready" }` → background replies with the authoritative
+ *      overlay state for the sender's window.
  *  - `{ kind: "overlay.interrupt", sessionId }` → background asks the
  *      daemon (via a `session.user_interrupt` WS event) to cancel
  *      every inflight + queued tool call for that session with

apps/extension/src/lib/random-uuid.ts

@@ -1,26 +0,0 @@
-/**
- * Generate a UUID v4 in browser contexts where randomUUID() is unavailable.
- *
- * Crypto.randomUUID() is restricted to secure contexts, while
- * crypto.getRandomValues() is available to content scripts on ordinary HTTP
- * origins. BrowserSkill records many intranet pages served over HTTP, so the
- * fallback is required for recording to work on those pages.
- */
-export function createRandomUuid(): string {
-  const cryptoApi = globalThis.crypto;
-  if (typeof cryptoApi.randomUUID === "function") return cryptoApi.randomUUID();
-
-  const bytes = new Uint8Array(16);
-  cryptoApi.getRandomValues(bytes);
-  bytes[6] = (bytes[6]! & 0x0f) | 0x40;
-  bytes[8] = (bytes[8]! & 0x3f) | 0x80;
-
-  const hex = Array.from(bytes, (byte) => byte.toString(16).padStart(2, "0")).join("");
-  return [
-    hex.slice(0, 8),
-    hex.slice(8, 12),
-    hex.slice(12, 16),
-    hex.slice(16, 20),
-    hex.slice(20),
-  ].join("-");
-}

apps/extension/src/lib/recording/frame-coordinator.ts

@@ -1,4 +1,3 @@
-import { createRandomUuid } from "../random-uuid";
 import {
   isRecordFrameQueryMessage,
   RECORD_FRAME_PORT,
@@ -189,7 +188,7 @@ export class RecordFrameCoordinator {
     const recording = this.#recordings.get(requestId);
     if (!recording) return true;
     recording.finishing = true;
-    const commandId = createRandomUuid();
+    const commandId = crypto.randomUUID();
     const results = await Promise.all(
       [...recording.agents.values()].map((agent) => this.#stopAgent(agent, requestId, commandId)),
     );

apps/extension/src/session-manager/__tests__/disconnect-cleanup.test.ts

@@ -9,7 +9,7 @@ describe("disconnect session cleanup", () => {
     const manager = new SessionManager({
       agentWindow: {
         create: vi.fn(async () => nextWindowId++),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
         remove,
       },
     });
@@ -41,7 +41,7 @@ describe("disconnect session cleanup", () => {
     const manager = new SessionManager({
       agentWindow: {
         create: vi.fn(async () => 100),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
         remove: vi.fn(() => removeGate),
       },
     });

apps/extension/src/session-manager/__tests__/event-handler.test.ts

@@ -39,7 +39,7 @@ describe("attachSessionEventHandler", () => {
       agentWindow: {
         create: vi.fn(async () => 4242),
         remove: vi.fn(async () => {}),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     await manager.start("aa11");
@@ -73,7 +73,7 @@ describe("attachSessionEventHandler", () => {
       agentWindow: {
         create: vi.fn(async () => 4242),
         remove: vi.fn(async () => {}),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     const ctx = await manager.start("aa11");
@@ -87,7 +87,6 @@ describe("attachSessionEventHandler", () => {
       windowEvents: events.api,
     });
 
-    manager.forgetClosedTab(7, { isWindowClosing: true });
     events.emit(4242);
     for (let i = 0; i < 4; i += 1) await Promise.resolve();
 
@@ -114,7 +113,7 @@ describe("attachSessionEventHandler", () => {
       agentWindow: {
         create: vi.fn(async () => 1),
         remove: vi.fn(),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     const transport = fakeTransport();
@@ -130,7 +129,7 @@ describe("attachSessionEventHandler", () => {
       agentWindow: {
         create: vi.fn(async () => 1),
         remove: vi.fn(),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     const transport = fakeTransport();

apps/extension/src/session-manager/__tests__/manager.test.ts

@@ -1,6 +1,6 @@
 import { describe, expect, it, vi } from "vitest";
 import type { AgentWindowApi, AgentWindowCreateOptions } from "../agent-window";
-import { isAgentControlledTab, SessionManager } from "../manager";
+import { SessionManager } from "../manager";
 
 function fakeAgentWindow(): AgentWindowApi & {
   createMock: ReturnType<typeof vi.fn>;
@@ -13,7 +13,7 @@ function fakeAgentWindow(): AgentWindowApi & {
     return id;
   });
   const removeMock = vi.fn(async (_id: number) => {});
-  const ensureActiveTabMock = vi.fn(async (_windowId: number, _url: string) => 0);
+  const ensureActiveTabMock = vi.fn(async (_windowId: number, _url: string) => {});
   return {
     create: createMock,
     remove: removeMock,
@@ -151,76 +151,6 @@ describe("SessionManager", () => {
     expect(sm.list()).toEqual([]);
   });
 
-  describe("explicit tab claims", () => {
-    it("claims the exact home tab returned by AgentWindowApi", async () => {
-      const sm = new SessionManager({ agentWindow: fakeAgentWindow() });
-      const ctx = await sm.start("aa11");
-      expect(ctx.agentCreatedTabs).toEqual(new Set([0]));
-      expect(isAgentControlledTab(ctx, 0)).toBe(true);
-    });
-
-    it("keeps every other tab free without classification state", async () => {
-      const sm = new SessionManager({ agentWindow: fakeAgentWindow() });
-      const ctx = await sm.start("aa11");
-      expect(isAgentControlledTab(ctx, 99)).toBe(false);
-      expect(ctx).not.toHaveProperty("userTabs");
-      expect(ctx).not.toHaveProperty("pendingAgentTabCount");
-    });
-
-    it("treats a committed borrow as an explicit claim", async () => {
-      const sm = new SessionManager({ agentWindow: fakeAgentWindow() });
-      const ctx = await sm.start("aa11");
-      const reservation = sm.tryReserveBorrow(42, ctx.sessionId);
-      if ("borrowedBy" in reservation) throw new Error("unexpected borrow conflict");
-      reservation.commit({ tabId: 42, originalWindowId: 7, originalIndex: 3 });
-      expect(isAgentControlledTab(ctx, 42)).toBe(true);
-    });
-
-    it("forgets an agent-created tab after Chrome removes it", async () => {
-      const sm = new SessionManager({ agentWindow: fakeAgentWindow() });
-      const ctx = await sm.start("aa11");
-      ctx.agentCreatedTabs.add(42);
-      sm.forgetClosedTab(42);
-      expect(ctx.agentCreatedTabs.has(42)).toBe(false);
-      expect(isAgentControlledTab(ctx, 0)).toBe(true);
-    });
-
-    it("forgets closed borrows idempotently without affecting other tabs or sessions", async () => {
-      const sm = new SessionManager({ agentWindow: fakeAgentWindow() });
-      const a = await sm.start("aa11");
-      const b = await sm.start("bb22");
-      a.borrowedTabs.set(42, { tabId: 42, originalWindowId: 7, originalIndex: 0 });
-      a.borrowedTabs.set(43, { tabId: 43, originalWindowId: 7, originalIndex: 1 });
-      b.borrowedTabs.set(44, { tabId: 44, originalWindowId: 8, originalIndex: 0 });
-
-      sm.forgetClosedTab(42);
-      sm.forgetClosedTab(42);
-      sm.forgetClosedTab(999);
-
-      expect(isAgentControlledTab(a, 42)).toBe(false);
-      expect(sm.findBorrowingSession(42, null)).toBeNull();
-      expect([...a.borrowedTabs.keys()]).toEqual([43]);
-      expect([...b.borrowedTabs.keys()]).toEqual([44]);
-      expect(sm.list()).toHaveLength(2);
-    });
-
-    it("prevents an in-flight borrow from reclaiming a closed tab", async () => {
-      const sm = new SessionManager({ agentWindow: fakeAgentWindow() });
-      const ctx = await sm.start("aa11");
-      const reservation = sm.tryReserveBorrow(42, ctx.sessionId);
-      if ("borrowedBy" in reservation) throw new Error("unexpected borrow conflict");
-
-      sm.forgetClosedTab(42);
-
-      expect(() =>
-        reservation.commit({ tabId: 42, originalWindowId: 7, originalIndex: 0 }),
-      ).toThrow(/reservation disappeared/);
-      reservation.release();
-      expect(sm.findBorrowingSession(42, null)).toBeNull();
-      expect(ctx.borrowedTabs.has(42)).toBe(false);
-    });
-  });
-
   describe("findBorrowingSession", () => {
     it("returns null when no session has borrowed the tab", async () => {
       const sm = new SessionManager({ agentWindow: fakeAgentWindow() });

apps/extension/src/session-manager/__tests__/ref-store.test.ts

@@ -13,6 +13,8 @@ describe("RefStore", () => {
       backendNodeId: 42,
       tabId: 7,
       generation: 0,
+      kind: "dom",
+      capabilities: ["interact", "screenshot"],
     });
     expect(s.size()).toBe(1);
     expect(s.isEmpty()).toBe(false);
@@ -48,4 +50,31 @@ describe("RefStore", () => {
       cdpSessionId: "child-session",
     });
   });
+
+  it("preserves explicit capabilities for visual surface refs", () => {
+    const s = new RefStore();
+    s.set("e1", 42, {
+      tabId: 7,
+      kind: "surface",
+      capabilities: ["screenshot"],
+    });
+
+    expect(s.resolveEntry("e1")).toMatchObject({
+      kind: "surface",
+      capabilities: ["screenshot"],
+    });
+  });
+
+  it("never grants ordinary interaction to a surface through defaults or caller input", () => {
+    const s = new RefStore();
+    s.set("e1", 42, { tabId: 7, kind: "surface" });
+    s.set("e2", 43, {
+      tabId: 7,
+      kind: "surface",
+      capabilities: ["interact", "screenshot"],
+    });
+
+    expect(s.resolveEntry("e1")?.capabilities).toEqual(["screenshot"]);
+    expect(s.resolveEntry("e2")?.capabilities).toEqual(["screenshot"]);
+  });
 });

apps/extension/src/session-manager/__tests__/surface-capture-store.test.ts

@@ -0,0 +1,67 @@
+import { describe, expect, it } from "vitest";
+import { type SurfaceCaptureInput, SurfaceCaptureStore } from "../surface-capture-store";
+
+function input(): SurfaceCaptureInput {
+  return {
+    sessionId: "aa11",
+    tabId: 7,
+    navigationIdentity: "navigation",
+    surface: {
+      ref: "e3",
+      frameId: "main",
+      backendNodeId: 99,
+      observationGeneration: 2,
+    },
+    topViewportRect: { x: 10, y: 20, w: 100, h: 50 },
+    imageWidth: 200,
+    imageHeight: 100,
+    viewportSignature: "viewport",
+    frameProjectionSignature: "frame",
+  };
+}
+
+describe("SurfaceCaptureStore", () => {
+  it("creates metadata-only, short-lived captures and consumes them once", () => {
+    let now = 1_000;
+    const store = new SurfaceCaptureStore({
+      now: () => now,
+      ttlMs: 500,
+      createId: () => "sc_test",
+    });
+    const capture = store.create(input());
+
+    expect(capture).toMatchObject({
+      id: "sc_test",
+      createdAt: 1_000,
+      expiresAt: 1_500,
+      consumed: false,
+    });
+    expect(capture).not.toHaveProperty("imageBase64");
+    expect(store.consume("sc_test")).toMatchObject({ ok: true });
+    expect(store.consume("sc_test")).toEqual({ ok: false, reason: "consumed" });
+
+    now = 1_600;
+    expect(store.size()).toBe(0);
+  });
+
+  it("rejects expired and unknown captures", () => {
+    let now = 1_000;
+    const store = new SurfaceCaptureStore({
+      now: () => now,
+      ttlMs: 10,
+      createId: () => "sc_expired",
+    });
+    store.create(input());
+    now = 1_010;
+
+    expect(store.consume("sc_expired")).toEqual({ ok: false, reason: "expired" });
+    expect(store.consume("sc_missing")).toEqual({ ok: false, reason: "not_found" });
+  });
+
+  it("clears every capture on session cleanup", () => {
+    const store = new SurfaceCaptureStore({ createId: () => "sc_clear" });
+    store.create(input());
+    store.clear();
+    expect(store.size()).toBe(0);
+  });
+});

apps/extension/src/session-manager/agent-window.ts

@@ -14,11 +14,8 @@ export interface AgentWindowApi {
    * Guarantee the Agent Window has an active, CDP-navigable tab.
    * `chrome://` pages (including the New Tab page) reject `Page.navigate`,
    * so sessions bootstrap with `about:blank` instead.
-   *
-   * Resolves with the id of the activated (or newly created) tab, so callers
-   * can track the session's home tab without re-querying Chrome.
    */
-  ensureActiveTab(windowId: number, url: string): Promise<number>;
+  ensureActiveTab(windowId: number, url: string): Promise<void>;
 }
 
 /** Creation hints for a new Agent Window. */
@@ -52,19 +49,15 @@ export const chromeAgentWindowApi: AgentWindowApi = {
     // cancellation success while the Agent Window remains open.
     await chrome.windows.remove(windowId);
   },
-  async ensureActiveTab(windowId: number, url: string): Promise<number> {
+  async ensureActiveTab(windowId: number, url: string): Promise<void> {
     const tabs = await chrome.tabs.query({ windowId });
     const first = tabs.find((t) => typeof t.id === "number");
-    if (first?.id !== undefined) {
+    if (first?.id) {
       if (!first.active) {
         await chrome.tabs.update(first.id, { active: true });
       }
-      return first.id;
+      return;
     }
-    const created = await chrome.tabs.create({ windowId, url, active: true });
-    if (typeof created?.id !== "number") {
-      throw new Error("[bh] chrome.tabs.create returned no tab id");
-    }
-    return created.id;
+    await chrome.tabs.create({ windowId, url, active: true });
   },
 };

apps/extension/src/session-manager/manager.ts

@@ -1,25 +1,16 @@
 import { AGENT_WINDOW_HOME, type AgentWindowApi, chromeAgentWindowApi } from "./agent-window";
 import { RefStore } from "./ref-store";
+import { SurfaceCaptureStore } from "./surface-capture-store";
 
 export interface SessionContext {
   sessionId: string;
   agentWindowId: number;
   refStore: RefStore;
+  surfaceCaptures: SurfaceCaptureStore;
   borrowedTabs: Map<number, BorrowedTab>;
-  /**
-   * Tabs explicitly claimed by the agent because it created them. This
-   * includes the Agent Window's home tab and tabs created by `tool.tab_create`.
-   * Tabs opened through Chrome UI never enter this set.
-   */
-  agentCreatedTabs: Set<number>;
   createdAtMs: number;
 }
 
-/** Whether this session has explicitly claimed control of `tabId`. */
-export function isAgentControlledTab(ctx: SessionContext, tabId: number): boolean {
-  return ctx.agentCreatedTabs.has(tabId) || ctx.borrowedTabs.has(tabId);
-}
-
 export interface BorrowedTab {
   tabId: number;
   originalWindowId: number;
@@ -117,19 +108,6 @@ export class SessionManager {
     return Array.from(this.sessions.values());
   }
 
-  /**
-   * Forget a tab Chrome has removed, including any uncommitted borrow.
-   * Whole-window closures keep committed borrows until the window-removed
-   * handler reports which user tabs could not be returned.
-   */
-  forgetClosedTab(tabId: number, { isWindowClosing = false } = {}): void {
-    this.borrowReservations.delete(tabId);
-    for (const ctx of this.sessions.values()) {
-      ctx.agentCreatedTabs.delete(tabId);
-      if (!isWindowClosing) ctx.borrowedTabs.delete(tabId);
-    }
-  }
-
   /**
    * Look up whether `tabId` is currently borrowed by some *other*
    * session than the one calling. Used by M8 `tab_borrow` to refuse
@@ -203,18 +181,15 @@ export class SessionManager {
       const { signal: _signal, ...createOptions } = opts;
       windowId = await this.agentWindow.create(AGENT_WINDOW_HOME, createOptions);
       throwIfSessionStartAborted(opts.signal);
-      const homeTabId = await this.agentWindow.ensureActiveTab(windowId, AGENT_WINDOW_HOME);
+      await this.agentWindow.ensureActiveTab(windowId, AGENT_WINDOW_HOME);
       throwIfSessionStartAborted(opts.signal);
 
       const ctx: SessionContext = {
         sessionId,
         agentWindowId: windowId,
         refStore: new RefStore(),
+        surfaceCaptures: new SurfaceCaptureStore({ now: this.now }),
         borrowedTabs: new Map(),
-        // The home tab is the session's first explicit claim. Every other
-        // tab remains free until `tab_create` or `tab_borrow` identifies it
-        // by its concrete Chrome tab id.
-        agentCreatedTabs: new Set([homeTabId]),
         createdAtMs: this.now(),
       };
       this.sessions.set(sessionId, ctx);

apps/extension/src/session-manager/ref-store.ts

@@ -1,3 +1,5 @@
+import type { Rect } from "@browser-skill/vom";
+
 /**
  * Per-session map from `@e<N>` snapshot refs to a CDP node address.
  *
@@ -12,12 +14,17 @@
  * inside the `SessionContext`.
  */
 export type BackendNodeId = number;
+export type RefCapability = "interact" | "screenshot";
+export type RefTargetKind = "dom" | "surface";
 
 export interface RefEntry {
   backendNodeId: BackendNodeId;
   tabId: number | null;
   frameId?: string;
   cdpSessionId?: string;
+  visibleRect?: Rect;
+  kind: RefTargetKind;
+  capabilities: RefCapability[];
   generation: number;
 }
 
@@ -28,8 +35,19 @@ export type RefInput =
       tabId: number;
       frameId?: string;
       cdpSessionId?: string;
+      visibleRect?: Rect;
+      kind?: RefTargetKind;
+      capabilities?: RefCapability[];
     };
 
+function capabilitiesFor(
+  kind: RefTargetKind,
+  capabilities: RefCapability[] | undefined,
+): RefCapability[] {
+  if (kind === "surface") return ["screenshot"];
+  return capabilities ?? ["interact", "screenshot"];
+}
+
 export class RefStore {
   private readonly map = new Map<string, RefEntry>();
   private generation = 0;
@@ -70,13 +88,20 @@ export class RefStore {
       tabId?: number;
       frameId?: string;
       cdpSessionId?: string;
+      visibleRect?: Rect;
+      kind?: RefTargetKind;
+      capabilities?: RefCapability[];
     } = {},
   ): void {
+    const kind = opts.kind ?? "dom";
     this.map.set(normaliseRef(ref), {
       backendNodeId: id,
       tabId: opts.tabId ?? null,
       ...(opts.frameId ? { frameId: opts.frameId } : {}),
       ...(opts.cdpSessionId ? { cdpSessionId: opts.cdpSessionId } : {}),
+      ...(opts.visibleRect ? { visibleRect: opts.visibleRect } : {}),
+      kind,
+      capabilities: capabilitiesFor(kind, opts.capabilities),
       generation: this.generation,
     });
   }
@@ -94,14 +119,20 @@ export class RefStore {
       return {
         backendNodeId: input,
         tabId: null,
+        kind: "dom",
+        capabilities: ["interact", "screenshot"],
         generation: this.generation,
       };
     }
+    const kind = input.kind ?? "dom";
     return {
       backendNodeId: input.backendNodeId,
       tabId: input.tabId,
       ...(input.frameId ? { frameId: input.frameId } : {}),
       ...(input.cdpSessionId ? { cdpSessionId: input.cdpSessionId } : {}),
+      ...(input.visibleRect ? { visibleRect: input.visibleRect } : {}),
+      kind,
+      capabilities: capabilitiesFor(kind, input.capabilities),
       generation: this.generation,
     };
   }

apps/extension/src/session-manager/surface-capture-store.ts

@@ -0,0 +1,100 @@
+import type { Rect } from "@browser-skill/vom";
+
+export const SURFACE_CAPTURE_TTL_MS = 30_000;
+
+export interface SurfaceCapture {
+  id: string;
+  sessionId: string;
+  tabId: number;
+  navigationIdentity: string;
+  surface: {
+    ref: string;
+    frameId?: string;
+    backendNodeId: number;
+    observationGeneration: number;
+  };
+  topViewportRect: Rect;
+  imageWidth: number;
+  imageHeight: number;
+  viewportSignature: string;
+  frameProjectionSignature: string;
+  createdAt: number;
+  expiresAt: number;
+  consumed: boolean;
+}
+
+export type SurfaceCaptureInput = Omit<
+  SurfaceCapture,
+  "id" | "createdAt" | "expiresAt" | "consumed"
+>;
+
+export type SurfaceCaptureConsumeResult =
+  | { ok: true; capture: SurfaceCapture }
+  | { ok: false; reason: "not_found" | "expired" | "consumed" };
+
+export interface SurfaceCaptureStoreOptions {
+  now?: () => number;
+  ttlMs?: number;
+  createId?: () => string;
+}
+
+function randomCaptureId(): string {
+  const bytes = crypto.getRandomValues(new Uint8Array(12));
+  return `sc_${Array.from(bytes, (byte) => byte.toString(16).padStart(2, "0")).join("")}`;
+}
+
+/** Per-session, metadata-only store for short-lived screenshot coordinate transactions. */
+export class SurfaceCaptureStore {
+  private readonly captures = new Map<string, SurfaceCapture>();
+  private readonly now: () => number;
+  private readonly ttlMs: number;
+  private readonly createId: () => string;
+
+  constructor(options: SurfaceCaptureStoreOptions = {}) {
+    this.now = options.now ?? Date.now;
+    this.ttlMs = options.ttlMs ?? SURFACE_CAPTURE_TTL_MS;
+    this.createId = options.createId ?? randomCaptureId;
+  }
+
+  create(input: SurfaceCaptureInput): SurfaceCapture {
+    this.purgeExpired();
+    const createdAt = this.now();
+    const capture: SurfaceCapture = {
+      ...input,
+      id: this.createId(),
+      createdAt,
+      expiresAt: createdAt + this.ttlMs,
+      consumed: false,
+    };
+    this.captures.set(capture.id, capture);
+    return capture;
+  }
+
+  consume(id: string): SurfaceCaptureConsumeResult {
+    const capture = this.captures.get(id);
+    if (!capture) return { ok: false, reason: "not_found" };
+    if (capture.expiresAt <= this.now()) {
+      this.captures.delete(id);
+      return { ok: false, reason: "expired" };
+    }
+    if (capture.consumed) return { ok: false, reason: "consumed" };
+    capture.consumed = true;
+    return { ok: true, capture };
+  }
+
+  clear(): void {
+    this.captures.clear();
+  }
+
+  size(): number {
+    this.purgeExpired();
+    return this.captures.size;
+  }
+
+  private purgeExpired(): void {
+    const now = this.now();
+    for (const [id, capture] of this.captures) {
+      if (capture.expiresAt <= now) this.captures.delete(id);
+    }
+  }
+}

apps/extension/src/tools/__tests__/borrow-confirmation.test.ts

@@ -127,38 +127,6 @@ describe("requestBorrowConfirmation", () => {
     expect(message.requestId.length).toBeGreaterThan(0);
   });
 
-  it("proactively focuses the user window before sending the overlay message", async () => {
-    tabs.get.mockResolvedValueOnce({ id: 42, title: "Tab" });
-    windows.getLastFocused.mockResolvedValueOnce(
-      userWindowWithActiveTab({ windowId: 77, tabId: 7, url: "https://app.example/" }),
-    );
-    tabs.sendMessage.mockResolvedValueOnce({ type: "borrow-response", allowed: true });
-
-    const pending = requestBorrowConfirmation(42, {
-      deps: { tabs, windows, notifications },
-    });
-    await vi.runAllTimersAsync();
-    await pending;
-
-    expect(windows.update).toHaveBeenCalledWith(77, { focused: true });
-  });
-
-  it("still proceeds when proactive window focus fails", async () => {
-    tabs.get.mockResolvedValueOnce({ id: 42, title: "Tab" });
-    windows.getLastFocused.mockResolvedValueOnce(
-      userWindowWithActiveTab({ windowId: 77, tabId: 7, url: "https://app.example/" }),
-    );
-    windows.update.mockRejectedValueOnce(new Error("window gone"));
-    tabs.sendMessage.mockResolvedValueOnce({ type: "borrow-response", allowed: true });
-
-    const pending = requestBorrowConfirmation(42, {
-      deps: { tabs, windows, notifications },
-    });
-    await vi.runAllTimersAsync();
-
-    expect(await pending).toBe(true);
-  });
-
   it("sends confirmation to the borrow target when it is the active tab", async () => {
     tabs.get.mockResolvedValueOnce({ id: 42, title: "Active Tab" });
     windows.getLastFocused.mockResolvedValueOnce(

apps/extension/src/tools/__tests__/console.test.ts

@@ -13,7 +13,7 @@ function fakeAgentWindow(ids: number[]) {
       return id;
     }),
     remove: vi.fn(async () => {}),
-    ensureActiveTab: vi.fn(async () => 1),
+    ensureActiveTab: vi.fn(async () => {}),
   };
 }
 

apps/extension/src/tools/__tests__/dispatcher.test.ts

@@ -49,54 +49,13 @@ describe("ToolDispatcher", () => {
     vi.unstubAllGlobals();
   });
 
-  it.each([
-    "tool.snapshot",
-    "tool.evaluate",
-    "tool.get_html",
-    "tool.console",
-  ])("annotates Chrome extension-access failures from %s", async (method) => {
-    const tab = { id: 7, windowId: 4242, active: true, url: "https://example.test" };
-    vi.stubGlobal("chrome", {
-      tabs: {
-        get: vi.fn(async () => tab),
-        query: vi.fn(async () => [tab]),
-      },
-    });
-    const { transport, sent, deliver } = fakeTransport();
-    const sessions = new SessionManager({
-      agentWindow: {
-        create: vi.fn(async () => 4242),
-        remove: vi.fn(),
-        ensureActiveTab: vi.fn(),
-      },
-    });
-    await sessions.start("test");
-    const error = new Error("Cannot access a chrome-extension:// URL of different extension");
-    const cdp = {
-      send: vi.fn().mockRejectedValue(error),
-      ensureConsoleCapture: vi.fn().mockRejectedValue(error),
-      consoleEntriesSince: vi.fn(),
-    } as unknown as TestDispatcherCdp;
-    const dispatcher = new ToolDispatcher({ transport, sessions, cdp });
-    dispatcher.start();
-    deliver(makeRequest(method, { session_id: "test", expression: "document.title" }));
-    await vi.waitFor(() => expect(sent).toHaveLength(1));
-    expect(sent[0]).toMatchObject({
-      error: {
-        code: "cdp_failed",
-        data: { reason: "cdp_extension_access_denied" },
-      },
-    });
-    dispatcher.stop();
-  });
-
   it("uses the configured recording runtime for start and stop", async () => {
     const { transport, sent, deliver } = fakeTransport();
     const sessions = new SessionManager({
       agentWindow: {
         create: vi.fn(async () => 4242),
         remove: vi.fn(),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     await sessions.start("aa11");
@@ -158,7 +117,7 @@ describe("ToolDispatcher", () => {
       agentWindow: {
         create: vi.fn(async () => 4242),
         remove: vi.fn(),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     const dispatcher = new ToolDispatcher({ transport, sessions });
@@ -181,7 +140,7 @@ describe("ToolDispatcher", () => {
       agentWindow: {
         create,
         remove: vi.fn(),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     const dispatcher = new ToolDispatcher({ transport, sessions });
@@ -199,7 +158,7 @@ describe("ToolDispatcher", () => {
       agentWindow: {
         create: vi.fn(async () => 4242),
         remove: vi.fn(async () => {}),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     await sessions.start("aa11");
@@ -213,47 +172,6 @@ describe("ToolDispatcher", () => {
     expect(sent[0]).toEqual({ id: "r-1", result: {} });
   });
 
-  it("wires the production tab APIs into tool.session_stop so a surviving user tab releases the window", async () => {
-    // Regression guard for the deps that session_stop reads directly: when
-    // `tabManagement.tabs` / `tabsQuery` are not injected by the dispatcher,
-    // the agent-tab cleanup and the release decision silently no-op and the
-    // Agent Window is closed even though a user tab is still open (issue #57).
-    const removeTab = vi.fn(async () => {});
-    const closeWindow = vi.fn(async () => {});
-    // After cleanup the window still holds one tab (id 99) that is neither
-    // the home tab nor an agent-created tab — i.e. a genuine user tab.
-    vi.stubGlobal("chrome", {
-      tabs: {
-        remove: removeTab,
-        query: vi.fn(async () => [{ id: 99, windowId: 4242, active: true }]),
-      },
-    });
-    const { transport, sent, deliver } = fakeTransport();
-    const sessions = new SessionManager({
-      agentWindow: {
-        create: vi.fn(async () => 4242),
-        remove: closeWindow,
-        ensureActiveTab: vi.fn(async () => 1),
-      },
-    });
-    await sessions.start("aa11");
-    // Register the exact tab id returned by the agent's creation path.
-    sessions.get("aa11")?.agentCreatedTabs.add(7);
-
-    const dispatcher = new ToolDispatcher({ transport, sessions });
-    dispatcher.start();
-
-    deliver(makeRequest("tool.session_stop", { session_id: "aa11" }));
-    await flushMicrotasks();
-
-    // The agent tab is closed through the real chrome.tabs surface...
-    expect(removeTab).toHaveBeenCalledWith(7);
-    // ...and the window is released rather than closed, so the user's tab survives.
-    expect(closeWindow).not.toHaveBeenCalled();
-    expect(sessions.has("aa11")).toBe(false);
-    expect(sent[0]).toEqual({ id: "r-1", result: { window_released: true } });
-  });
-
   it("routes tool.console through the CDP console buffer", async () => {
     vi.stubGlobal("chrome", {
       tabs: {
@@ -265,7 +183,7 @@ describe("ToolDispatcher", () => {
       agentWindow: {
         create: vi.fn(async () => 4242),
         remove: vi.fn(async () => {}),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     await sessions.start("aa11");
@@ -314,114 +232,13 @@ describe("ToolDispatcher", () => {
     });
   });
 
-  it("bypasses and restores the control overlay for an upload trigger click", async () => {
-    const sendMessage = vi.fn(async () => undefined);
-    vi.stubGlobal("chrome", {
-      tabs: {
-        get: vi.fn(async () => ({ id: 7, windowId: 4242, active: true })),
-        query: vi.fn(async () => [{ id: 7, windowId: 4242, active: true }]),
-        sendMessage,
-      },
-    });
-    const { transport, sent, deliver } = fakeTransport();
-    const sessions = new SessionManager({
-      agentWindow: {
-        create: vi.fn(async () => 4242),
-        remove: vi.fn(async () => {}),
-        ensureActiveTab: vi.fn(async () => 1),
-      },
-    });
-    const ctx = await sessions.start("aa11");
-    ctx.refStore.set("e1", 123, { tabId: 7 });
-    const send = vi.fn(async <T>(_tabId: number, method: string, params?: object) => {
-      if (method === "Page.getLayoutMetrics") {
-        return { cssLayoutViewport: { clientWidth: 1280, clientHeight: 720 } } as T;
-      }
-      if (method === "DOM.getContentQuads") {
-        return { quads: [[0, 0, 20, 0, 20, 20, 0, 20]] } as T;
-      }
-      if (method === "DOM.resolveNode") {
-        return { object: { objectId: "trigger-object" } } as T;
-      }
-      if (method === "DOM.describeNode") return { node: { backendNodeId: 456 } } as T;
-      if (method === "Runtime.callFunctionOn") {
-        const declaration = (params as { functionDeclaration?: string }).functionDeclaration ?? "";
-        if (declaration.includes("count: state.inputs.length")) {
-          return { result: { value: { count: 1, multiple: false } } } as T;
-        }
-        if (declaration.includes("inputs[0]")) {
-          return { result: { objectId: "input-object" } } as T;
-        }
-        return { result: { value: true } } as T;
-      }
-      if (method === "Runtime.evaluate") {
-        const expression = (params as { expression?: string }).expression ?? "";
-        if (expression.includes("overlayDetails")) {
-          return { result: { value: { hitIndex: 0 } } } as T;
-        }
-        if (expression.includes("overlayHostPresent")) {
-          return {
-            result: {
-              value: { overlayHostPresent: true, overlayHostConnected: true },
-            },
-          } as T;
-        }
-        if (expression.includes("count:")) {
-          return { result: { value: { count: 1, multiple: false } } } as T;
-        }
-        if (expression.includes("?.inputs[0]")) {
-          return { result: { objectId: "input-object" } } as T;
-        }
-        return { result: { value: true } } as T;
-      }
-      return {} as T;
-    });
-    const cdp = {
-      send,
-      detachSession: vi.fn(async () => {}),
-      ensureNetworkCapture: vi.fn(async () => {}),
-      networkEntriesSince: vi.fn(() => ({
-        tab_id: 7,
-        entries: [],
-        next_since: 0,
-        truncated: false,
-      })),
-      setDeviceMetricsOverride: vi.fn(async () => {}),
-      clearDeviceMetricsOverride: vi.fn(async () => {}),
-      setUserAgentOverride: vi.fn(async () => {}),
-      setTouchEmulationEnabled: vi.fn(async () => {}),
-    };
-    const dispatcher = new ToolDispatcher({ transport, sessions, cdp: cdp as TestDispatcherCdp });
-    dispatcher.start();
-
-    deliver(
-      makeRequest("tool.upload", {
-        session_id: "aa11",
-        ref: "@e1",
-        files: [{ transfer_id: "tr_1", name: "test.png", staged_path: "/stage/test.png" }],
-      }),
-    );
-    await flushMicrotasks();
-    await vi.waitFor(() => expect(sent).toHaveLength(1));
-
-    expect(sent[0]).toMatchObject({ result: { tab_id: 7, file_names: ["test.png"] } });
-    expect(sendMessage).toHaveBeenNthCalledWith(1, 7, {
-      type: "bh-automation-bypass",
-      enabled: true,
-    });
-    expect(sendMessage).toHaveBeenNthCalledWith(2, 7, {
-      type: "bh-automation-bypass",
-      enabled: false,
-    });
-  });
-
   it("detaches CDP state before stopping a session", async () => {
     const { transport, sent, deliver } = fakeTransport();
     const sessions = new SessionManager({
       agentWindow: {
         create: vi.fn(async () => 4242),
         remove: vi.fn(async () => {}),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     await sessions.start("aa11");
@@ -457,7 +274,7 @@ describe("ToolDispatcher", () => {
       agentWindow: {
         create: vi.fn(async () => 4242),
         remove,
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     await sessions.start("aa11");
@@ -505,7 +322,7 @@ describe("ToolDispatcher", () => {
       agentWindow: {
         create: vi.fn(async () => 1),
         remove: vi.fn(),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     const dispatcher = new ToolDispatcher({ transport, sessions });
@@ -525,7 +342,7 @@ describe("ToolDispatcher", () => {
       agentWindow: {
         create: vi.fn(async () => 1),
         remove: vi.fn(),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     const dispatcher = new ToolDispatcher({ transport, sessions });
@@ -549,7 +366,7 @@ describe("ToolDispatcher", () => {
       agentWindow: {
         create: vi.fn(async () => 1),
         remove: vi.fn(),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     const dispatcher = new ToolDispatcher({ transport, sessions });
@@ -566,7 +383,7 @@ describe("ToolDispatcher", () => {
       agentWindow: {
         create: vi.fn(async () => 1),
         remove: vi.fn(),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     const dispatcher = new ToolDispatcher({ transport, sessions });
@@ -583,7 +400,7 @@ describe("ToolDispatcher", () => {
       agentWindow: {
         create: vi.fn(async () => 1),
         remove: vi.fn(async () => {}),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     const onSessionsChanged = vi.fn();
@@ -600,32 +417,16 @@ describe("ToolDispatcher", () => {
   });
 
   it("invokes onBrowserControlResumed for browser-control tools but not passive reads", async () => {
-    vi.stubGlobal("chrome", {
-      tabs: {
-        create: vi.fn(async (props: chrome.tabs.CreateProperties) => ({
-          id: 7,
-          windowId: props.windowId,
-          url: props.url,
-          active: props.active,
-        })),
-      },
-    });
     const { transport, sent, deliver } = fakeTransport();
     const sessions = new SessionManager({
       agentWindow: {
         create: vi.fn(async () => 1),
         remove: vi.fn(async () => {}),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     const onBrowserControlResumed = vi.fn();
-    const onAgentTabClaimed = vi.fn();
-    const dispatcher = new ToolDispatcher({
-      transport,
-      sessions,
-      onBrowserControlResumed,
-      onAgentTabClaimed,
-    });
+    const dispatcher = new ToolDispatcher({ transport, sessions, onBrowserControlResumed });
     dispatcher.start();
 
     deliver(makeRequest("tool.session_start", { session_id: "aa11" }));
@@ -642,7 +443,6 @@ describe("ToolDispatcher", () => {
     });
     await flushMicrotasks();
     expect(onBrowserControlResumed).toHaveBeenCalledWith("aa11");
-    expect(onAgentTabClaimed).toHaveBeenCalledWith(7, 1);
   });
 
   it("reasserts remembered hover before follow-up work and releases only after actions", async () => {
@@ -656,7 +456,7 @@ describe("ToolDispatcher", () => {
       agentWindow: {
         create: vi.fn(async () => 1),
         remove: vi.fn(async () => {}),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     const order: string[] = [];
@@ -738,7 +538,7 @@ describe("ToolDispatcher", () => {
       agentWindow: {
         create: vi.fn(async () => 1),
         remove: vi.fn(async () => {}),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     const cdp = {
@@ -788,135 +588,6 @@ describe("ToolDispatcher", () => {
     );
   });
 
-  it.each([
-    { tab_id: 7, code: "not_found" },
-    { tab_id: undefined, code: "invalid_params" },
-  ])("keeps hover when tab_return is rejected with $code", async ({ tab_id, code }) => {
-    const sendMessage = vi.fn(async () => undefined);
-    vi.stubGlobal("chrome", {
-      tabs: {
-        sendMessage,
-        get: vi.fn(async () => ({ id: 7, windowId: 100, active: true })),
-        query: vi.fn(async () => [{ id: 7, windowId: 100, active: true }]),
-      },
-    });
-    const { transport, sent, deliver } = fakeTransport();
-    const sessions = new SessionManager({
-      agentWindow: {
-        create: vi.fn(async () => 100),
-        remove: vi.fn(async () => {}),
-        ensureActiveTab: vi.fn(async () => 1),
-      },
-    });
-    await sessions.start("aa11");
-    const cdp = {
-      send: vi.fn(async <T>() => ({}) as T),
-      detachSession: vi.fn(async () => {}),
-      releaseSessionTab: vi.fn(async () => {}),
-    };
-    const dispatcher = new ToolDispatcher({
-      transport,
-      sessions,
-      cdp: cdp as unknown as TestDispatcherCdp,
-    });
-    const helpers = dispatcher as unknown as {
-      rememberHover: (sessionId: string, result: unknown) => unknown;
-      setHoverBypass: (sessionId: string, tabId: number, enabled: boolean) => Promise<void>;
-      withHoverReassert: <T>(
-        params: { session_id: string; tab_id?: number },
-        work: () => Promise<T>,
-      ) => Promise<T>;
-    };
-    helpers.rememberHover("aa11", { tab_id: 7, x: 10, y: 20 });
-    await helpers.setHoverBypass("aa11", 7, true);
-    dispatcher.start();
-    deliver(makeRequest("tool.tab_return", { session_id: "aa11", tab_id }));
-    await vi.waitFor(() => expect(sent).toHaveLength(1));
-
-    expect(sent[0]).toMatchObject({ error: { code } });
-    expect(cdp.releaseSessionTab).not.toHaveBeenCalled();
-    expect(sendMessage).not.toHaveBeenCalledWith(7, expect.objectContaining({ enabled: false }));
-    await helpers.withHoverReassert({ session_id: "aa11", tab_id: 7 }, async () => ({}));
-    expect(cdp.send).toHaveBeenCalledExactlyOnceWith(7, "Input.dispatchMouseEvent", {
-      type: "mouseMoved",
-      x: 10,
-      y: 20,
-    });
-    dispatcher.stop();
-  });
-
-  it("releases the returned tab's hover before moving it and leaves other tabs alone", async () => {
-    const sendMessage = vi.fn(async () => undefined);
-    const move = vi.fn(async () => {
-      expect(sendMessage).toHaveBeenCalledWith(7, expect.objectContaining({ enabled: false }));
-      return { id: 7, windowId: 200, index: 4 };
-    });
-    vi.stubGlobal("chrome", {
-      tabs: { sendMessage, move },
-      windows: { get: vi.fn(async () => ({ id: 200 })) },
-    });
-    const { transport, sent, deliver } = fakeTransport();
-    const sessions = new SessionManager({
-      agentWindow: {
-        create: vi.fn(async () => 100),
-        remove: vi.fn(async () => {}),
-        ensureActiveTab: vi.fn(async () => 1),
-      },
-    });
-    const ctx = await sessions.start("aa11");
-    ctx.borrowedTabs.set(7, { tabId: 7, originalWindowId: 200, originalIndex: 4 });
-    const cdp = {
-      send: vi.fn(async <T>() => ({}) as T),
-      detachSession: vi.fn(async () => {}),
-      releaseSessionTab: vi.fn(async () => {}),
-    };
-    const dispatcher = new ToolDispatcher({
-      transport,
-      sessions,
-      cdp: cdp as unknown as TestDispatcherCdp,
-    });
-    const helpers = dispatcher as unknown as {
-      rememberHover: (sessionId: string, result: unknown) => unknown;
-      setHoverBypass: (sessionId: string, tabId: number, enabled: boolean) => Promise<void>;
-      withHoverReassert: <T>(
-        params: { session_id: string; tab_id?: number },
-        work: () => Promise<T>,
-      ) => Promise<T>;
-    };
-    for (const [sessionId, tabId] of [
-      ["aa11", 7],
-      ["aa11", 8],
-      ["bb22", 9],
-    ] as const) {
-      helpers.rememberHover(sessionId, { tab_id: tabId, x: 10, y: 20 });
-      await helpers.setHoverBypass(sessionId, tabId, true);
-    }
-    dispatcher.start();
-    deliver(makeRequest("tool.tab_return", { session_id: "aa11", tab_id: 7 }));
-    await vi.waitFor(() => expect(sent).toHaveLength(1));
-
-    expect(sent[0]).toMatchObject({ result: { tab_id: 7, returned_to_window_id: 200 } });
-    expect(cdp.releaseSessionTab).toHaveBeenCalledExactlyOnceWith("aa11", 7);
-    await helpers.withHoverReassert({ session_id: "aa11", tab_id: 7 }, async () => ({}));
-    expect(cdp.send).not.toHaveBeenCalled();
-    for (const [sessionId, tabId] of [
-      ["aa11", 8],
-      ["bb22", 9],
-    ] as const) {
-      expect(sendMessage).not.toHaveBeenCalledWith(
-        tabId,
-        expect.objectContaining({ enabled: false }),
-      );
-      await helpers.withHoverReassert({ session_id: sessionId, tab_id: tabId }, async () => ({}));
-      expect(cdp.send).toHaveBeenLastCalledWith(tabId, "Input.dispatchMouseEvent", {
-        type: "mouseMoved",
-        x: 10,
-        y: 20,
-      });
-    }
-    dispatcher.stop();
-  });
-
   it("transfers hover overlay bypass ownership between sessions", async () => {
     vi.stubGlobal("chrome", {
       tabs: {
@@ -928,7 +599,7 @@ describe("ToolDispatcher", () => {
       agentWindow: {
         create: vi.fn(async () => 1),
         remove: vi.fn(async () => {}),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     const dispatcher = new ToolDispatcher({ transport, sessions });
@@ -965,7 +636,7 @@ describe("ToolDispatcher", () => {
       agentWindow: {
         create: vi.fn(async () => 4242),
         remove: vi.fn(async () => {}),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     await sessions.start("aa11");
@@ -1017,7 +688,7 @@ describe("ToolDispatcher", () => {
       agentWindow: {
         create: vi.fn(async () => 1),
         remove: vi.fn(async () => {}),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     const dispatcher = new ToolDispatcher({ transport, sessions });
@@ -1051,7 +722,7 @@ describe("ToolDispatcher", () => {
       agentWindow: {
         create: vi.fn(async () => 4242),
         remove: vi.fn(),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     transport.send = () => {
@@ -1073,7 +744,7 @@ describe("ToolDispatcher", () => {
       agentWindow: {
         create: vi.fn(async () => 5555),
         remove,
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     transport.send = () => {
@@ -1101,7 +772,7 @@ describe("ToolDispatcher", () => {
       agentWindow: {
         create: vi.fn(() => createPromise),
         remove: vi.fn(async () => {}),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     const dispatcher = new ToolDispatcher({ transport, sessions });
@@ -1152,7 +823,7 @@ describe("ToolDispatcher", () => {
       agentWindow: {
         create: vi.fn(async () => 1),
         remove: vi.fn(),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     const dispatcher = new ToolDispatcher({ transport, sessions });
@@ -1173,7 +844,7 @@ describe("ToolDispatcher", () => {
       agentWindow: {
         create: vi.fn(() => createPromise),
         remove: vi.fn(),
-        ensureActiveTab: vi.fn(async () => 1),
+        ensureActiveTab: vi.fn(async () => {}),
       },
     });
     const dispatcher = new ToolDispatcher({ transport, sessions });

apps/extension/src/tools/__tests__/editable-input.test.ts

@@ -0,0 +1,104 @@
+import { describe, expect, it, vi } from "vitest";
+import { fillEditable } from "../editable-input";
+import type { CdpRunner } from "../shared";
+
+interface FakeOptions {
+  initialValue?: string;
+  finalValue: string;
+  remount?: boolean;
+}
+
+function editableState(value: string, connected = true, focused = true) {
+  return {
+    result: {
+      value: { supported: true, connected, focused, value, kind: "input" },
+    },
+  };
+}
+
+function makeEditableCdp(options: FakeOptions) {
+  const sent: Array<{ method: string; params?: object }> = [];
+  let inserted = false;
+  const send = vi.fn(async <T>(_tabId: number, method: string, params?: object): Promise<T> => {
+    sent.push({ method, params });
+    if (method === "Input.insertText") {
+      inserted = true;
+      return {} as T;
+    }
+    if (method === "Input.dispatchKeyEvent") {
+      if ((params as { type?: string }).type === "rawKeyDown") inserted = true;
+      return {} as T;
+    }
+    if (method !== "Runtime.callFunctionOn") throw new Error(`unexpected ${method}`);
+
+    const call = params as { objectId?: string; functionDeclaration?: string };
+    const fn = call.functionDeclaration ?? "";
+    if (fn.includes("bsk-editable-selection")) {
+      return { result: { value: { selected: true } } } as T;
+    }
+    if (fn.includes("bsk-editable-sync")) return { result: { value: undefined } } as T;
+    if (fn.includes("bsk-editable-live")) {
+      return {
+        result: options.remount && inserted ? { objectId: "replacement" } : { objectId: "anchor" },
+      } as T;
+    }
+    if (fn.includes("bsk-editable-state")) {
+      if (call.objectId === "replacement") return editableState(options.finalValue) as T;
+      if (options.remount && inserted) {
+        return editableState(options.finalValue, false, false) as T;
+      }
+      return editableState(inserted ? options.finalValue : (options.initialValue ?? "")) as T;
+    }
+    throw new Error("unexpected Runtime.callFunctionOn function");
+  });
+  return { cdp: { send } as CdpRunner, sent };
+}
+
+describe("fillEditable", () => {
+  it("replaces text through the browser input path without synthetic events or value setters", async () => {
+    const fake = makeEditableCdp({ initialValue: "old", finalValue: "新值" });
+
+    const result = await fillEditable(fake.cdp, 4, "anchor", "新值", true);
+
+    expect(result).toMatchObject({ value: "新值", replacedNode: false });
+    expect(fake.sent).toContainEqual({ method: "Input.insertText", params: { text: "新值" } });
+    const declarations = fake.sent
+      .map((call) => (call.params as { functionDeclaration?: string })?.functionDeclaration ?? "")
+      .join("\n");
+    expect(declarations).not.toMatch(/dispatchEvent|\.value\s*=/);
+  });
+
+  it("reacquires the focused editable after a framework remount", async () => {
+    const fake = makeEditableCdp({ finalValue: "persisted", remount: true });
+
+    const result = await fillEditable(fake.cdp, 4, "anchor", "persisted", true);
+
+    expect(result).toMatchObject({ value: "persisted", replacedNode: true });
+  });
+
+  it("reports a page rollback after the bounded settlement", async () => {
+    const fake = makeEditableCdp({ finalValue: "" });
+
+    const result = await fillEditable(fake.cdp, 4, "anchor", "rejected", true);
+
+    expect(result).toMatchObject({
+      code: "cdp_failed",
+      data: {
+        reason: "input_not_applied",
+        expected_value_length: 8,
+        actual_value_length: 0,
+        phase: "readback",
+      },
+    });
+  });
+
+  it("clears to an empty value with a browser Backspace edit", async () => {
+    const fake = makeEditableCdp({ initialValue: "old", finalValue: "" });
+
+    const result = await fillEditable(fake.cdp, 4, "anchor", "", true);
+
+    expect(result).toMatchObject({ value: "" });
+    expect(fake.sent.some((call) => call.method === "Input.insertText")).toBe(false);
+    expect(fake.sent.filter((call) => call.method === "Input.dispatchKeyEvent")).toHaveLength(2);
+  });
+});

apps/extension/src/tools/__tests__/emulate.test.ts

@@ -20,7 +20,7 @@ function fakeAgentWindow(ids: number[]) {
       return id;
     }),
     remove: vi.fn(async () => {}),
-    ensureActiveTab: vi.fn(async () => 1),
+    ensureActiveTab: vi.fn(async () => {}),
   };
 }
 

apps/extension/src/tools/__tests__/errors.test.ts

@@ -1,39 +0,0 @@
-import { describe, expect, it } from "vitest";
-import type { RpcError } from "@/transport/types";
-import { cdpError, classifyCdpError } from "../errors";
-
-const denied = "Cannot access a chrome-extension:// URL of different extension";
-
-describe("CDP extension access errors", () => {
-  it("classifies both Chrome Error instances and string failures", () => {
-    for (const error of [new Error(denied), denied]) {
-      expect(cdpError(error)).toEqual({
-        code: "cdp_failed",
-        message: denied,
-        data: { reason: "cdp_extension_access_denied" },
-      });
-    }
-  });
-
-  it("preserves transfer outcome and cleanup instructions when CDP was denied", () => {
-    const error: RpcError = {
-      code: "cdp_failed",
-      message: denied,
-      data: {
-        reason: "transfer_outcome_unknown",
-        effect_state: "unknown",
-        cleanup_state: "failed",
-      },
-    };
-    expect(classifyCdpError(error)).toBe(error);
-  });
-
-  it("leaves other error codes and unrelated debugger failures unchanged", () => {
-    for (const error of [
-      { code: "permission_denied", message: denied },
-      { code: "cdp_failed", message: "Another debugger is already attached" },
-    ] satisfies RpcError[]) {
-      expect(classifyCdpError(error)).toBe(error);
-    }
-  });
-});

apps/extension/src/tools/__tests__/evaluate.test.ts

@@ -12,7 +12,7 @@ function fakeAgentWindow(ids: number[]) {
       return id;
     }),
     remove: vi.fn(async () => {}),
-    ensureActiveTab: vi.fn(async () => 1),
+    ensureActiveTab: vi.fn(async () => {}),
   };
 }
 

apps/extension/src/tools/__tests__/file-transfer.test.ts

@@ -1,988 +0,0 @@
-import { describe, expect, it, vi } from "vitest";
-import { SessionManager } from "@/session-manager/manager";
-import { type DownloadsApi, handleDownload } from "../download";
-import { captureBrowserDownload } from "../download-capture";
-import { uploadThroughFileDrop } from "../file-drop-transaction";
-import { uploadThroughActivatedFileInput } from "../file-input-transaction";
-import type { ResolvedActionTarget } from "../interaction";
-import type { CdpRunner } from "../shared";
-import { handleUpload } from "../upload";
-
-function sessions() {
-  return new SessionManager({
-    agentWindow: {
-      create: vi.fn(async () => 100),
-      remove: vi.fn(async () => {}),
-      ensureActiveTab: vi.fn(async () => 1),
-    },
-  });
-}
-
-function tabsApi() {
-  return {
-    get: vi.fn(
-      async (tabId: number) => ({ id: tabId, windowId: 100, active: true }) as chrome.tabs.Tab,
-    ),
-    query: vi.fn(async () => [{ id: 4, windowId: 100, active: true } as chrome.tabs.Tab]),
-  };
-}
-
-function fakeEvent<T extends (...args: never[]) => unknown>() {
-  const listeners = new Set<T>();
-  return {
-    addListener: (listener: T) => listeners.add(listener),
-    removeListener: (listener: T) => listeners.delete(listener),
-    emit: (...args: Parameters<T>) => {
-      for (const listener of listeners) listener(...args);
-    },
-  };
-}
-
-function actionTarget(frameId?: string, sessionId?: string): ResolvedActionTarget {
-  return {
-    tab: { tabId: 4, windowId: 100, active: true },
-    backendNodeId: 123,
-    cdpTarget: { tabId: 4, ...(sessionId ? { sessionId } : {}) },
-    ...(frameId ? { frameId } : {}),
-    usedRef: "e3",
-  };
-}
-
-function withoutOverlay<T>(operation: () => Promise<T>): Promise<T> {
-  return operation();
-}
-
-function uploadCdp(
-  options: {
-    inputCount?: number;
-    multiple?: boolean;
-    chooser?: { frameId?: string; backendNodeId?: number; mode?: string };
-    pendingResolve?: boolean;
-  } = {},
-) {
-  const calls: Array<{ method: string; params?: object }> = [];
-  let cdpEvent: Parameters<NonNullable<CdpRunner["onEvent"]>>[0] | undefined;
-  const send = vi.fn(async (_tabId: number, method: string, params?: object) => {
-    calls.push({ method, params });
-    if (method === "DOM.scrollIntoViewIfNeeded") return {};
-    if (method === "Page.setInterceptFileChooserDialog") return {};
-    if (method === "Page.getLayoutMetrics")
-      return { cssLayoutViewport: { clientWidth: 1280, clientHeight: 720 } };
-    if (method === "DOM.getContentQuads") return { quads: [[0, 0, 20, 0, 20, 20, 0, 20]] };
-    if (method === "DOM.resolveNode") {
-      if (options.pendingResolve) return new Promise<never>(() => {});
-      return { object: { objectId: "trigger-object" } };
-    }
-    if (method === "Runtime.callFunctionOn") {
-      const declaration = (params as { functionDeclaration?: string }).functionDeclaration ?? "";
-      if (declaration.includes("Object.defineProperty")) return { result: { value: true } };
-      if (declaration.includes("count: state.inputs.length")) {
-        return {
-          result: {
-            value: {
-              count: options.inputCount ?? 1,
-              multiple: options.multiple ?? true,
-            },
-          },
-        };
-      }
-      if (declaration.includes("inputs[0]")) {
-        return { result: { objectId: "input-object" } };
-      }
-      return { result: { value: true } };
-    }
-    if (method === "DOM.describeNode") return { node: { backendNodeId: 456 } };
-    if (
-      method === "Input.dispatchMouseEvent" &&
-      (params as { type?: string }).type === "mousePressed" &&
-      options.chooser
-    ) {
-      cdpEvent?.({ tabId: 4 }, "Page.fileChooserOpened", options.chooser);
-    }
-    return {};
-  });
-  const cdp: CdpRunner = {
-    send: send as unknown as CdpRunner["send"],
-    onEvent: (handler) => {
-      cdpEvent = handler;
-      return { dispose: vi.fn() };
-    },
-  };
-  return {
-    calls,
-    emitChooser: (event: { frameId?: string; backendNodeId?: number; mode?: string }) =>
-      cdpEvent?.({ tabId: 4 }, "Page.fileChooserOpened", event),
-    cdp,
-  };
-}
-
-function dropCdp(
-  options: {
-    targetContainsPoint?: boolean;
-    dispatchError?: (type: string) => Error | Promise<never> | undefined;
-    onCall?: (method: string) => void;
-  } = {},
-) {
-  const calls: Array<{ method: string; params?: object }> = [];
-  const detach = vi.fn(async () => {});
-  const send = vi.fn(async (_tabId: number, method: string, params?: object) => {
-    calls.push({ method, params });
-    options.onCall?.(method);
-    if (method === "DOM.scrollIntoViewIfNeeded") return {};
-    if (method === "Page.getLayoutMetrics") {
-      return { cssLayoutViewport: { clientWidth: 1280, clientHeight: 720 } };
-    }
-    if (method === "DOM.getContentQuads") {
-      return { quads: [[10, 20, 210, 20, 210, 120, 10, 120]] };
-    }
-    if (method === "DOM.getNodeForLocation") return { backendNodeId: 124 };
-    if (method === "DOM.resolveNode") {
-      const id = (params as { backendNodeId?: number }).backendNodeId;
-      return { object: { objectId: id === 123 ? "drop-target" : "hit-target" } };
-    }
-    if (method === "Runtime.callFunctionOn") {
-      return { result: { value: options.targetContainsPoint ?? true } };
-    }
-    if (method === "Input.dispatchDragEvent") {
-      const type = (params as { type?: string }).type ?? "";
-      const error = options.dispatchError?.(type);
-      if (error instanceof Promise) return error;
-      if (error) throw error;
-      return {};
-    }
-    if (method === "Runtime.releaseObjectGroup") return {};
-    throw new Error(`unexpected drop CDP command ${method}`);
-  });
-  const cdp: CdpRunner = {
-    send: send as unknown as CdpRunner["send"],
-    detach,
-  };
-  return { calls, cdp, detach };
-}
-
-describe("file transfer tools", () => {
-  it("captures the file input activated by the requested click and injects only staged paths", async () => {
-    const manager = sessions();
-    const ctx = await manager.start("s1");
-    ctx.refStore.set("e3", 123, { tabId: 4 });
-    const { cdp, calls } = uploadCdp();
-
-    const result = await handleUpload(
-      manager,
-      {
-        session_id: "s1",
-        ref: "@e3",
-        files: [
-          { transfer_id: "tr_1", name: "one.png", staged_path: "/private/stage/one" },
-          { transfer_id: "tr_2", name: "two.png", staged_path: "/private/stage/two" },
-        ],
-      },
-      { cdp, tabsApi: tabsApi() },
-    );
-
-    expect(result).toMatchObject({ tab_id: 4, file_names: ["one.png", "two.png"] });
-    expect(calls[0]).toEqual({
-      method: "Page.setInterceptFileChooserDialog",
-      params: { enabled: true },
-    });
-    expect(calls).toContainEqual({
-      method: "Page.setInterceptFileChooserDialog",
-      params: { enabled: false, cancel: true },
-    });
-    expect(calls).toContainEqual({
-      method: "DOM.setFileInputFiles",
-      params: { files: ["/private/stage/one", "/private/stage/two"], backendNodeId: 456 },
-    });
-  });
-
-  it("fails immediately when the trigger does not activate a file input", async () => {
-    const { cdp } = uploadCdp({ inputCount: 0 });
-    const result = await uploadThroughActivatedFileInput({
-      cdp,
-      actionTarget: actionTarget(),
-      files: ["/private/stage/one"],
-      timeoutMs: 100,
-      trigger: async () => ({ tab_id: 4, x: 10, y: 10 }),
-    });
-
-    expect(result).toMatchObject({
-      code: "unsupported",
-      data: { reason: "file_input_not_activated", phase: "resolve_input" },
-    });
-  });
-
-  it("fails before clicking when chooser interception is unavailable", async () => {
-    const trigger = vi.fn(async () => ({ tab_id: 4, x: 10, y: 10 }));
-    const cdp: CdpRunner = {
-      send: vi.fn(async (_tabId: number, method: string) => {
-        if (method === "Page.setInterceptFileChooserDialog") {
-          throw new Error("method unavailable");
-        }
-        return {};
-      }) as CdpRunner["send"],
-    };
-
-    const result = await uploadThroughActivatedFileInput({
-      cdp,
-      actionTarget: actionTarget(),
-      files: ["/private/stage/one"],
-      timeoutMs: 100,
-      trigger,
-    });
-
-    expect(result).toMatchObject({
-      code: "cdp_failed",
-      data: { effect_state: "none", phase: "arm_interception" },
-    });
-    expect(trigger).not.toHaveBeenCalled();
-  });
-
-  it("uses an exact chooser event as an independent input-location signal", async () => {
-    const { cdp, calls, emitChooser } = uploadCdp({ inputCount: 0 });
-    const result = await uploadThroughActivatedFileInput({
-      cdp,
-      actionTarget: actionTarget("f1"),
-      files: ["/private/stage/one"],
-      timeoutMs: 100,
-      trigger: async () => {
-        emitChooser({ frameId: "f1", backendNodeId: 789, mode: "selectSingle" });
-        return { tab_id: 4, x: 10, y: 10 };
-      },
-    });
-
-    expect(result).toMatchObject({ multiple: false });
-    expect(calls).toContainEqual({
-      method: "DOM.setFileInputFiles",
-      params: { files: ["/private/stage/one"], backendNodeId: 789 },
-    });
-  });
-
-  it("reports a File System Access picker without waiting for a timeout", async () => {
-    const { cdp, emitChooser } = uploadCdp({
-      inputCount: 0,
-    });
-    const result = await uploadThroughActivatedFileInput({
-      cdp,
-      actionTarget: actionTarget("f1"),
-      files: ["/private/stage/one"],
-      timeoutMs: 100,
-      trigger: async () => {
-        emitChooser({ frameId: "f1", mode: "selectSingle" });
-        return { tab_id: 4, x: 10, y: 10 };
-      },
-    });
-
-    expect(result).toMatchObject({
-      code: "unsupported",
-      message: "upload trigger invoked a non-input file picker",
-      data: { reason: "file_input_not_activated", phase: "resolve_input" },
-    });
-  });
-
-  it("bounds a stuck file-input probe", async () => {
-    const { cdp } = uploadCdp({ pendingResolve: true });
-
-    const result = await uploadThroughActivatedFileInput({
-      cdp,
-      actionTarget: actionTarget(),
-      files: ["/private/stage/one"],
-      timeoutMs: 5,
-      trigger: vi.fn(),
-    });
-
-    expect(result).toMatchObject({
-      code: "timeout",
-      data: { reason: "file_input_probe_failed", phase: "arm_input_probe" },
-    });
-  });
-
-  it("marks a timed-out file assignment unknown and detaches browser state", async () => {
-    const fixture = uploadCdp();
-    const originalSend = fixture.cdp.send;
-    const detach = vi.fn(async () => {});
-    fixture.cdp.detach = detach;
-    fixture.cdp.send = vi.fn((tabId: number, method: string, params?: object) => {
-      if (method === "DOM.setFileInputFiles") return new Promise<never>(() => {});
-      return originalSend(tabId, method, params);
-    }) as CdpRunner["send"];
-
-    const result = await uploadThroughActivatedFileInput({
-      cdp: fixture.cdp,
-      actionTarget: actionTarget(),
-      files: ["/private/stage/one"],
-      timeoutMs: 10,
-      trigger: async () => ({ tab_id: 4, x: 10, y: 10 }),
-    });
-
-    expect(result).toMatchObject({
-      code: "timeout",
-      data: { effect_state: "unknown", phase: "set_files" },
-    });
-    expect(detach).toHaveBeenCalledWith(4);
-  });
-
-  it("delivers staged files through one explicit native drop without clicking", async () => {
-    const manager = sessions();
-    const ctx = await manager.start("s1");
-    ctx.refStore.set("e3", 123, { tabId: 4 });
-    const lifecycle: string[] = [];
-    const { cdp, calls } = dropCdp({ onCall: (method) => lifecycle.push(method) });
-    const overlayPhases: string[] = [];
-
-    const result = await handleUpload(
-      manager,
-      {
-        session_id: "s1",
-        ref: "@e3",
-        mode: "drop",
-        files: [{ transfer_id: "tr_1", name: "one.png", staged_path: "/stage/one.png" }],
-      },
-      {
-        cdp,
-        tabsApi: tabsApi(),
-        sendToTab: vi.fn(async (_tabId, message) => {
-          overlayPhases.push(message.phase);
-          lifecycle.push(`overlay:${message.phase}`);
-        }),
-      },
-    );
-
-    expect(result).toMatchObject({ tab_id: 4, used_ref: "e3", file_names: ["one.png"] });
-    expect(calls.filter((call) => call.method === "Input.dispatchDragEvent")).toEqual([
-      {
-        method: "Input.dispatchDragEvent",
-        params: {
-          type: "dragEnter",
-          x: 110,
-          y: 70,
-          data: { items: [], files: ["/stage/one.png"], dragOperationsMask: 1 },
-        },
-      },
-      {
-        method: "Input.dispatchDragEvent",
-        params: {
-          type: "dragOver",
-          x: 110,
-          y: 70,
-          data: { items: [], files: ["/stage/one.png"], dragOperationsMask: 1 },
-        },
-      },
-      {
-        method: "Input.dispatchDragEvent",
-        params: {
-          type: "drop",
-          x: 110,
-          y: 70,
-          data: { items: [], files: ["/stage/one.png"], dragOperationsMask: 1 },
-        },
-      },
-    ]);
-    expect(calls.some((call) => call.method === "Input.dispatchMouseEvent")).toBe(false);
-    expect(overlayPhases).toEqual(["begin", "end"]);
-    expect(lifecycle.indexOf("DOM.getContentQuads")).toBeLessThan(
-      lifecycle.indexOf("overlay:begin"),
-    );
-    expect(lifecycle.indexOf("overlay:begin")).toBeLessThan(
-      lifecycle.indexOf("DOM.getNodeForLocation"),
-    );
-    expect(lifecycle.lastIndexOf("Input.dispatchDragEvent")).toBeLessThan(
-      lifecycle.indexOf("overlay:end"),
-    );
-    expect(lifecycle.indexOf("overlay:end")).toBeLessThan(
-      lifecycle.indexOf("Runtime.releaseObjectGroup"),
-    );
-  });
-
-  it("fails closed before dragging when another element owns the action point", async () => {
-    const { cdp, calls } = dropCdp({ targetContainsPoint: false });
-    const result = await uploadThroughFileDrop({
-      cdp,
-      actionTarget: actionTarget(),
-      files: ["/stage/one.png"],
-      timeoutMs: 100,
-      withOverlayHidden: withoutOverlay,
-    });
-
-    expect(result).toMatchObject({
-      code: "permission_denied",
-      data: {
-        reason: "file_drop_target_unavailable",
-        mechanism: "drop",
-        effect_state: "none",
-        phase: "resolve_target",
-      },
-    });
-    expect(calls.some((call) => call.method === "Input.dispatchDragEvent")).toBe(false);
-  });
-
-  it("dispatches an OOPIF file drop in the child target's local coordinates", async () => {
-    const rootCalls: Array<{ method: string; params?: object }> = [];
-    const childCalls: Array<{ method: string; params?: object }> = [];
-    const cdp: CdpRunner = {
-      send: vi.fn(async (_tabId, method, params) => {
-        rootCalls.push({ method, params });
-        if (method === "DOM.scrollIntoViewIfNeeded") return {};
-        if (method === "DOM.getBoxModel") {
-          return { model: { content: [100, 100, 300, 100, 300, 200, 100, 200] } };
-        }
-        if (method === "Page.getLayoutMetrics") {
-          return { cssLayoutViewport: { clientWidth: 800, clientHeight: 600 } };
-        }
-        throw new Error(`unexpected root command ${method}`);
-      }) as CdpRunner["send"],
-      sendToTarget: vi.fn(async (_target, method, params) => {
-        childCalls.push({ method, params });
-        if (method === "DOM.scrollIntoViewIfNeeded") return {};
-        if (method === "DOM.getContentQuads") {
-          return { quads: [[10, 20, 210, 20, 210, 120, 10, 120]] };
-        }
-        if (method === "Page.getLayoutMetrics") {
-          return { cssLayoutViewport: { clientWidth: 400, clientHeight: 300 } };
-        }
-        if (method === "DOM.getNodeForLocation") {
-          return { backendNodeId: 124, frameId: "child" };
-        }
-        if (method === "DOM.resolveNode") {
-          const id = (params as { backendNodeId?: number }).backendNodeId;
-          return { object: { objectId: id === 123 ? "drop-target" : "hit-target" } };
-        }
-        if (method === "Runtime.callFunctionOn") return { result: { value: true } };
-        if (method === "Input.dispatchDragEvent" || method === "Runtime.releaseObjectGroup") {
-          return {};
-        }
-        throw new Error(`unexpected child command ${method}`);
-      }) as CdpRunner["sendToTarget"],
-      getFrameGraph: vi.fn(async () => ({
-        rootFrameId: "main",
-        frames: [
-          { frameId: "main", target: { tabId: 4 } },
-          {
-            frameId: "child",
-            parentFrameId: "main",
-            ownerBackendNodeId: 10,
-            target: { tabId: 4, sessionId: "child-session" },
-          },
-        ],
-      })),
-    };
-
-    const result = await uploadThroughFileDrop({
-      cdp,
-      actionTarget: actionTarget("child", "child-session"),
-      files: ["/stage/one.png"],
-      timeoutMs: 100,
-      withOverlayHidden: withoutOverlay,
-    });
-
-    expect(result).toMatchObject({ x: 110, y: 70 });
-    expect(
-      childCalls
-        .filter((call) => call.method === "Input.dispatchDragEvent")
-        .map((call) => call.params),
-    ).toEqual([
-      expect.objectContaining({ type: "dragEnter", x: 110, y: 70 }),
-      expect.objectContaining({ type: "dragOver", x: 110, y: 70 }),
-      expect.objectContaining({ type: "drop", x: 110, y: 70 }),
-    ]);
-    expect(rootCalls.some((call) => call.method === "Input.dispatchDragEvent")).toBe(false);
-  });
-
-  it("reports a missing native drag command as unsupported before commit", async () => {
-    const { cdp } = dropCdp({
-      dispatchError: (type) =>
-        type === "dragEnter" ? new Error("'Input.dispatchDragEvent' wasn't found") : undefined,
-    });
-    const result = await uploadThroughFileDrop({
-      cdp,
-      actionTarget: actionTarget(),
-      files: ["/stage/one.png"],
-      timeoutMs: 100,
-      withOverlayHidden: withoutOverlay,
-    });
-
-    expect(result).toMatchObject({
-      code: "unsupported",
-      data: {
-        reason: "upload_mechanism_unsupported",
-        effect_state: "none",
-        phase: "drag_enter",
-      },
-    });
-  });
-
-  it("cancels an unresolved drag before returning a pre-commit timeout", async () => {
-    const { cdp, calls } = dropCdp({
-      dispatchError: (type) => (type === "dragEnter" ? new Promise<never>(() => {}) : undefined),
-    });
-    const result = await uploadThroughFileDrop({
-      cdp,
-      actionTarget: actionTarget(),
-      files: ["/stage/one.png"],
-      timeoutMs: 10,
-      withOverlayHidden: withoutOverlay,
-    });
-
-    expect(result).toMatchObject({
-      code: "timeout",
-      data: { reason: "file_drop_failed", effect_state: "none", phase: "drag_enter" },
-    });
-    expect(calls).toContainEqual({
-      method: "Input.dispatchDragEvent",
-      params: expect.objectContaining({ type: "dragCancel" }),
-    });
-  });
-
-  it("marks an unresolved drop unknown and detaches browser state", async () => {
-    const { cdp, detach } = dropCdp({
-      dispatchError: (type) => (type === "drop" ? new Promise<never>(() => {}) : undefined),
-    });
-    const result = await uploadThroughFileDrop({
-      cdp,
-      actionTarget: actionTarget(),
-      files: ["/stage/one.png"],
-      timeoutMs: 10,
-      withOverlayHidden: withoutOverlay,
-    });
-
-    expect(result).toMatchObject({
-      code: "timeout",
-      data: { reason: "file_drop_failed", effect_state: "unknown", phase: "drop" },
-    });
-    expect(detach).toHaveBeenCalledWith(4);
-  });
-
-  it("routes one exact-target download through a browser-relative capability", async () => {
-    const manager = sessions();
-    const ctx = await manager.start("s1");
-    ctx.refStore.set("e3", 123, { tabId: 4 });
-    const onCreated = fakeEvent<(item: chrome.downloads.DownloadItem) => void>();
-    const onChanged = fakeEvent<(delta: chrome.downloads.DownloadDelta) => void>();
-    const onDeterminingFilename =
-      fakeEvent<
-        (
-          item: chrome.downloads.DownloadItem,
-          suggest: (suggestion?: chrome.downloads.DownloadFilenameSuggestion) => void,
-        ) => void | true
-      >();
-    const initial = {
-      id: 9,
-      url: "https://example.test/result.zip",
-      finalUrl: "https://example.test/result.zip",
-      filename: "result.zip",
-      state: "in_progress",
-      fileSize: -1,
-      totalBytes: 12,
-      mime: "application/zip",
-      danger: "safe",
-    } as chrome.downloads.DownloadItem;
-    const completed = {
-      ...initial,
-      filename: "/profile/Downloads/BrowserSkill/tr_1/result.zip",
-      state: "complete",
-      fileSize: 12,
-    } as chrome.downloads.DownloadItem;
-    const downloads: DownloadsApi = {
-      onCreated,
-      onChanged,
-      onDeterminingFilename,
-      search: vi.fn(async () => [completed]),
-      cancel: vi.fn(async () => {}),
-      removeFile: vi.fn(async () => {}),
-    };
-    let cdpEvent: Parameters<NonNullable<CdpRunner["onEvent"]>>[0] | undefined;
-    let suggested: chrome.downloads.DownloadFilenameSuggestion | undefined;
-    const send = vi.fn(async (_tabId: number, method: string, params?: object) => {
-      if (method === "Page.getLayoutMetrics")
-        return { cssLayoutViewport: { clientWidth: 1280, clientHeight: 720 } };
-      if (method === "DOM.getContentQuads") return { quads: [[0, 0, 20, 0, 20, 20, 0, 20]] };
-      if (
-        method === "Input.dispatchMouseEvent" &&
-        (params as { type?: string }).type === "mousePressed"
-      ) {
-        cdpEvent?.({ tabId: 4 }, "Page.downloadWillBegin", {
-          url: initial.url,
-          suggestedFilename: "result.zip",
-        });
-        await new Promise<void>((resolve) => {
-          onDeterminingFilename.emit(initial, (value) => {
-            suggested = value;
-            resolve();
-          });
-        });
-        onCreated.emit(completed);
-      }
-      return {};
-    });
-    const cdp: CdpRunner = {
-      send: send as unknown as CdpRunner["send"],
-      onEvent: (handler) => {
-        cdpEvent = handler;
-        return { dispose: vi.fn() };
-      },
-    };
-
-    const result = await handleDownload(
-      manager,
-      { session_id: "s1", ref: "@e3", browser_relative_dir: "BrowserSkill/tr_1" },
-      { cdp, tabsApi: tabsApi(), downloads },
-    );
-
-    expect(suggested).toEqual({
-      filename: "BrowserSkill/tr_1/result.zip",
-      conflictAction: "overwrite",
-    });
-    expect(result).toMatchObject({
-      tab_id: 4,
-      suggested_filename: "result.zip",
-      byte_size: 12,
-      browser_path: completed.filename,
-    });
-  });
-
-  it("does not claim a download without an intent from the exact target", async () => {
-    const onCreated = fakeEvent<(item: chrome.downloads.DownloadItem) => void>();
-    const onChanged = fakeEvent<(delta: chrome.downloads.DownloadDelta) => void>();
-    const onDeterminingFilename =
-      fakeEvent<
-        (
-          item: chrome.downloads.DownloadItem,
-          suggest: (suggestion?: chrome.downloads.DownloadFilenameSuggestion) => void,
-        ) => void | true
-      >();
-    const downloads: DownloadsApi = {
-      onCreated,
-      onChanged,
-      onDeterminingFilename,
-      search: vi.fn(async () => []),
-      cancel: vi.fn(async () => {}),
-      removeFile: vi.fn(async () => {}),
-    };
-    let cdpEvent: Parameters<NonNullable<CdpRunner["onEvent"]>>[0] | undefined;
-    const cdp: CdpRunner = {
-      send: vi.fn(async () => ({})) as unknown as CdpRunner["send"],
-      onEvent: (handler) => {
-        cdpEvent = handler;
-        return { dispose: vi.fn() };
-      },
-    };
-    const unrelated = {
-      id: 17,
-      url: "https://example.test/unrelated.zip",
-      finalUrl: "https://example.test/unrelated.zip",
-      filename: "unrelated.zip",
-      state: "in_progress",
-    } as chrome.downloads.DownloadItem;
-    let defaultSuggestionCalled = false;
-
-    const result = await captureBrowserDownload({
-      cdp,
-      target: { tabId: 4, sessionId: "expected-child" },
-      downloads,
-      browserRelativeDir: "BrowserSkill/tr_1",
-      timeoutMs: 5,
-      trigger: async () => {
-        cdpEvent?.({ tabId: 4, sessionId: "other-child" }, "Page.downloadWillBegin", {
-          url: unrelated.url,
-          suggestedFilename: unrelated.filename,
-        });
-        onDeterminingFilename.emit(unrelated, (suggestion) => {
-          defaultSuggestionCalled = suggestion === undefined;
-        });
-        return { tab_id: 4, x: 10, y: 10 };
-      },
-    });
-
-    expect(defaultSuggestionCalled).toBe(true);
-    expect(downloads.cancel).not.toHaveBeenCalled();
-    expect(result).toMatchObject({
-      code: "cdp_failed",
-      data: { reason: "download_capture_failed" },
-    });
-  });
-
-  it("correlates a filename candidate that arrives before the CDP intent", async () => {
-    const onCreated = fakeEvent<(item: chrome.downloads.DownloadItem) => void>();
-    const onChanged = fakeEvent<(delta: chrome.downloads.DownloadDelta) => void>();
-    const onDeterminingFilename =
-      fakeEvent<
-        (
-          item: chrome.downloads.DownloadItem,
-          suggest: (suggestion?: chrome.downloads.DownloadFilenameSuggestion) => void,
-        ) => void | true
-      >();
-    const initial = {
-      id: 21,
-      url: "https://example.test/candidate-first.bin",
-      finalUrl: "https://example.test/candidate-first.bin",
-      filename: "candidate-first.bin",
-      state: "in_progress",
-      fileSize: -1,
-      totalBytes: 4,
-      bytesReceived: 0,
-    } as chrome.downloads.DownloadItem;
-    const complete = {
-      ...initial,
-      state: "complete",
-      fileSize: 4,
-    } as chrome.downloads.DownloadItem;
-    const downloads: DownloadsApi = {
-      onCreated,
-      onChanged,
-      onDeterminingFilename,
-      search: vi.fn(async () => [complete]),
-      cancel: vi.fn(async () => {}),
-      removeFile: vi.fn(async () => {}),
-    };
-    let cdpEvent: Parameters<NonNullable<CdpRunner["onEvent"]>>[0] | undefined;
-    const cdp: CdpRunner = {
-      send: vi.fn(async () => ({})) as CdpRunner["send"],
-      onEvent: (handler) => {
-        cdpEvent = handler;
-        return { dispose: vi.fn() };
-      },
-    };
-    let suggestion: chrome.downloads.DownloadFilenameSuggestion | undefined;
-
-    const result = await captureBrowserDownload({
-      cdp,
-      target: { tabId: 4 },
-      downloads,
-      browserRelativeDir: "BrowserSkill/tr_21",
-      timeoutMs: 1_000,
-      trigger: async () => {
-        const suggested = new Promise<void>((resolve) => {
-          onDeterminingFilename.emit(initial, (value) => {
-            suggestion = value;
-            resolve();
-          });
-        });
-        cdpEvent?.({ tabId: 4 }, "Page.downloadWillBegin", {
-          url: initial.url,
-          suggestedFilename: initial.filename,
-        });
-        await suggested;
-        onCreated.emit(complete);
-        return { tab_id: 4, x: 10, y: 10 };
-      },
-    });
-
-    expect(suggestion).toEqual({
-      filename: "BrowserSkill/tr_21/candidate-first.bin",
-      conflictAction: "overwrite",
-    });
-    expect(result).toMatchObject({ item: { id: 21, state: "complete" } });
-  });
-
-  it("removes a completed download when the final size exceeds the limit", async () => {
-    const onCreated = fakeEvent<(item: chrome.downloads.DownloadItem) => void>();
-    const onChanged = fakeEvent<(delta: chrome.downloads.DownloadDelta) => void>();
-    const onDeterminingFilename =
-      fakeEvent<
-        (
-          item: chrome.downloads.DownloadItem,
-          suggest: (suggestion?: chrome.downloads.DownloadFilenameSuggestion) => void,
-        ) => void | true
-      >();
-    const initial = {
-      id: 22,
-      url: "https://example.test/oversized.bin",
-      finalUrl: "https://example.test/oversized.bin",
-      filename: "oversized.bin",
-      state: "in_progress",
-      fileSize: -1,
-      totalBytes: -1,
-      bytesReceived: 0,
-    } as chrome.downloads.DownloadItem;
-    const complete = {
-      ...initial,
-      filename: "/profile/Downloads/BrowserSkill/tr_22/oversized.bin",
-      state: "complete",
-      fileSize: 8,
-      totalBytes: 8,
-      bytesReceived: 8,
-    } as chrome.downloads.DownloadItem;
-    const downloads: DownloadsApi = {
-      onCreated,
-      onChanged,
-      onDeterminingFilename,
-      search: vi.fn(async () => [complete]),
-      cancel: vi.fn(async () => {}),
-      removeFile: vi.fn(async () => {}),
-    };
-    let cdpEvent: Parameters<NonNullable<CdpRunner["onEvent"]>>[0] | undefined;
-    const cdp: CdpRunner = {
-      send: vi.fn(async () => ({})) as CdpRunner["send"],
-      onEvent: (handler) => {
-        cdpEvent = handler;
-        return { dispose: vi.fn() };
-      },
-    };
-
-    const result = await captureBrowserDownload({
-      cdp,
-      target: { tabId: 4 },
-      downloads,
-      browserRelativeDir: "BrowserSkill/tr_22",
-      maxByteSize: 4,
-      timeoutMs: 1_000,
-      trigger: async () => {
-        cdpEvent?.({ tabId: 4 }, "Page.downloadWillBegin", {
-          url: initial.url,
-          suggestedFilename: initial.filename,
-        });
-        await new Promise<void>((resolve) => {
-          onDeterminingFilename.emit(initial, () => resolve());
-        });
-        onCreated.emit(initial);
-        onChanged.emit({ id: initial.id, state: { current: "complete" } });
-        return { tab_id: 4, x: 10, y: 10 };
-      },
-    });
-
-    expect(result).toMatchObject({
-      code: "cdp_failed",
-      data: { reason: "download_capture_failed", effect_state: "committed" },
-    });
-    expect(downloads.removeFile).toHaveBeenCalledWith(initial.id);
-    expect(downloads.cancel).not.toHaveBeenCalled();
-  });
-
-  it("reconciles a download that completes while cancellation is being requested", async () => {
-    const onCreated = fakeEvent<(item: chrome.downloads.DownloadItem) => void>();
-    const onChanged = fakeEvent<(delta: chrome.downloads.DownloadDelta) => void>();
-    const onDeterminingFilename =
-      fakeEvent<
-        (
-          item: chrome.downloads.DownloadItem,
-          suggest: (suggestion?: chrome.downloads.DownloadFilenameSuggestion) => void,
-        ) => void | true
-      >();
-    const initial = {
-      id: 23,
-      url: "https://example.test/racing.bin",
-      finalUrl: "https://example.test/racing.bin",
-      filename: "racing.bin",
-      state: "in_progress",
-      fileSize: -1,
-      totalBytes: -1,
-      bytesReceived: 0,
-    } as chrome.downloads.DownloadItem;
-    const complete = {
-      ...initial,
-      filename: "/profile/Downloads/BrowserSkill/tr_23/racing.bin",
-      state: "complete",
-      fileSize: 4,
-      totalBytes: 4,
-      bytesReceived: 4,
-    } as chrome.downloads.DownloadItem;
-    const downloads: DownloadsApi = {
-      onCreated,
-      onChanged,
-      onDeterminingFilename,
-      search: vi.fn().mockResolvedValueOnce([initial]).mockResolvedValueOnce([complete]),
-      cancel: vi.fn(async () => {
-        throw new Error("download already complete");
-      }),
-      removeFile: vi.fn(async () => {}),
-    };
-    let cdpEvent: Parameters<NonNullable<CdpRunner["onEvent"]>>[0] | undefined;
-    const cdp: CdpRunner = {
-      send: vi.fn(async () => ({})) as CdpRunner["send"],
-      onEvent: (handler) => {
-        cdpEvent = handler;
-        return { dispose: vi.fn() };
-      },
-    };
-
-    const result = await captureBrowserDownload({
-      cdp,
-      target: { tabId: 4 },
-      downloads,
-      browserRelativeDir: "BrowserSkill/tr_23",
-      timeoutMs: 80,
-      trigger: async () => {
-        cdpEvent?.({ tabId: 4 }, "Page.downloadWillBegin", {
-          url: initial.url,
-          suggestedFilename: initial.filename,
-        });
-        await new Promise<void>((resolve) => {
-          onDeterminingFilename.emit(initial, () => resolve());
-        });
-        onCreated.emit(initial);
-        return { tab_id: 4, x: 10, y: 10 };
-      },
-    });
-
-    expect(result).toMatchObject({
-      code: "cdp_failed",
-      data: { reason: "download_capture_failed", effect_state: "committed" },
-    });
-    expect(downloads.cancel).toHaveBeenCalledWith(initial.id);
-    expect(downloads.removeFile).toHaveBeenCalledWith(initial.id);
-    expect(result).not.toMatchObject({ data: { cleanup_state: "failed" } });
-  });
-
-  it("rejects ambiguous attribution without cancelling either unclaimed download", async () => {
-    const onCreated = fakeEvent<(item: chrome.downloads.DownloadItem) => void>();
-    const onChanged = fakeEvent<(delta: chrome.downloads.DownloadDelta) => void>();
-    const onDeterminingFilename =
-      fakeEvent<
-        (
-          item: chrome.downloads.DownloadItem,
-          suggest: (suggestion?: chrome.downloads.DownloadFilenameSuggestion) => void,
-        ) => void | true
-      >();
-    const downloads: DownloadsApi = {
-      onCreated,
-      onChanged,
-      onDeterminingFilename,
-      search: vi.fn(async () => []),
-      cancel: vi.fn(async () => {}),
-      removeFile: vi.fn(async () => {}),
-    };
-    let cdpEvent: Parameters<NonNullable<CdpRunner["onEvent"]>>[0] | undefined;
-    const cdp: CdpRunner = {
-      send: vi.fn(async () => ({})) as CdpRunner["send"],
-      onEvent: (handler) => {
-        cdpEvent = handler;
-        return { dispose: vi.fn() };
-      },
-    };
-    const defaults: number[] = [];
-    const candidate = (id: number) =>
-      ({
-        id,
-        url: "https://example.test/same.bin",
-        finalUrl: "https://example.test/same.bin",
-        filename: "same.bin",
-        state: "in_progress",
-      }) as chrome.downloads.DownloadItem;
-
-    const result = await captureBrowserDownload({
-      cdp,
-      target: { tabId: 4 },
-      downloads,
-      browserRelativeDir: "BrowserSkill/tr_ambiguous",
-      timeoutMs: 100,
-      trigger: async () => {
-        cdpEvent?.({ tabId: 4 }, "Page.downloadWillBegin", {
-          url: "https://example.test/same.bin",
-          suggestedFilename: "same.bin",
-        });
-        for (const id of [31, 32]) {
-          onDeterminingFilename.emit(candidate(id), (value) => {
-            if (value === undefined) defaults.push(id);
-          });
-        }
-        return { tab_id: 4, x: 10, y: 10 };
-      },
-    });
-
-    expect(defaults.sort()).toEqual([31, 32]);
-    expect(downloads.cancel).not.toHaveBeenCalled();
-    expect(result).toMatchObject({
-      code: "cdp_failed",
-      data: { effect_state: "unknown", phase: "attribution" },
-    });
-  });
-});

apps/extension/src/tools/__tests__/fill.test.ts

@@ -1,573 +0,0 @@
-import { afterEach, describe, expect, it, vi } from "vitest";
-import { SessionManager } from "@/session-manager/manager";
-import type { CdpRunner } from "@/tools/shared";
-import { handleFill } from "../interaction";
-
-interface ScriptParams {
-  functionDeclaration: string;
-  arguments?: Array<{ value: unknown }>;
-}
-
-async function setup(markup = '<input value="old">') {
-  document.body.innerHTML = markup;
-  const element = document.body.firstElementChild as HTMLInputElement;
-  const manager = new SessionManager({
-    agentWindow: {
-      create: async () => 100,
-      remove: async () => {},
-      ensureActiveTab: async () => 4,
-    },
-  });
-  const ctx = await manager.start("aa11");
-  ctx.refStore.set("e1", 12, { tabId: 4 });
-  const script = vi.fn(async (params: ScriptParams): Promise<unknown> => {
-    try {
-      const fn = new Function(`return (${params.functionDeclaration})`)();
-      return {
-        result: {
-          value: await fn.apply(
-            element,
-            (params.arguments ?? []).map((arg) => arg.value),
-          ),
-        },
-      };
-    } catch {
-      return { exceptionDetails: { text: "test page script failed" } };
-    }
-  });
-  // Number/email controls have a browser caret even though their selection
-  // getters return null. Model that caret instead of assuming it is at the end.
-  const hiddenCaret = { start: 0, end: 0 };
-  const moveToEnd = vi.fn(async () => {
-    if (typeof element.selectionStart === "number") {
-      element.setSelectionRange(element.value.length, element.value.length);
-    } else {
-      hiddenCaret.start = hiddenCaret.end = element.value.length;
-    }
-  });
-  const insert = vi.fn(async (text: string) => {
-    const focused = document.activeElement as HTMLInputElement;
-    if (focused.readOnly || focused.disabled) return;
-    const start = focused.selectionStart ?? hiddenCaret.start;
-    const end = focused.selectionEnd ?? hiddenCaret.end;
-    const value = focused.value.slice(0, start) + text + focused.value.slice(end);
-    focused.value = focused.maxLength >= 0 ? value.slice(0, focused.maxLength) : value;
-    hiddenCaret.start = hiddenCaret.end = Math.min(start + text.length, focused.value.length);
-    focused.dispatchEvent(new Event("input", { bubbles: true }));
-  });
-  const release = vi.fn(async () => ({}));
-  const send = vi.fn(async (_tabId: number, method: string, params?: object) => {
-    switch (method) {
-      case "DOM.describeNode":
-        return {
-          node: {
-            nodeName: element.tagName,
-            attributes: Array.from(element.attributes).flatMap((attr) => [attr.name, attr.value]),
-          },
-        };
-      case "DOM.scrollIntoViewIfNeeded":
-        return {};
-      case "DOM.focus":
-        element.focus();
-        return {};
-      case "DOM.resolveNode":
-        return { object: { objectId: "fill-target" } };
-      case "Runtime.callFunctionOn":
-        return script(params as ScriptParams);
-      case "Input.insertText":
-        await insert((params as { text: string }).text);
-        return {};
-      case "Input.dispatchKeyEvent":
-        if ((params as { commands?: string[] }).commands?.includes("moveToEndOfDocument")) {
-          await moveToEnd();
-        }
-        return {};
-      case "Runtime.releaseObject":
-        return release();
-      default:
-        throw new Error(`unexpected ${method}`);
-    }
-  });
-  return {
-    element,
-    hiddenCaret,
-    moveToEnd,
-    script,
-    insert,
-    release,
-    send,
-    fill: (value = "hello", clearBefore = true, signal?: AbortSignal) =>
-      handleFill(
-        manager,
-        { session_id: "aa11", ref: "e1", value, clear_before: clearBefore },
-        {
-          cdp: { send: send as CdpRunner["send"] },
-          tabsApi: {
-            get: async (id) => ({ id, windowId: 100, active: true }) as chrome.tabs.Tab,
-            query: async () => [{ id: 4, windowId: 100, active: true } as chrome.tabs.Tab],
-          },
-          signal,
-        },
-      ),
-  };
-}
-
-afterEach(() => {
-  vi.restoreAllMocks();
-  document.body.innerHTML = "";
-});
-
-describe("fill result verification", () => {
-  it.each([
-    "readonly",
-    "disabled",
-  ])("preserves a %s field before rejecting it", async (attribute) => {
-    const h = await setup(`<input ${attribute} value="old">`);
-    expect(await h.fill()).toMatchObject({
-      code: "invalid_params",
-      data: { reason: "target_not_fillable" },
-    });
-    expect(h.element.value).toBe("old");
-    expect(h.insert).not.toHaveBeenCalled();
-    expect(h.release).toHaveBeenCalledOnce();
-  });
-
-  it.each([
-    "checkbox",
-    "radio",
-    "range",
-    "date",
-    "file",
-  ])("rejects an unsupported %s input before clearing", async (type) => {
-    const h = await setup(`<input type="${type}">`);
-    const before = h.element.value;
-    expect(await h.fill()).toMatchObject({
-      code: "invalid_params",
-      data: { reason: "target_not_fillable" },
-    });
-    expect(h.element.value).toBe(before);
-    expect(h.insert).not.toHaveBeenCalled();
-  });
-
-  it("rejects an overlong value before clearing", async () => {
-    const h = await setup('<input maxlength="3" value="old">');
-    expect(await h.fill()).toMatchObject({
-      code: "invalid_params",
-      message: /maxlength/,
-      data: { reason: "fill_value_invalid" },
-    });
-    expect(h.element.value).toBe("old");
-    expect(h.insert).not.toHaveBeenCalled();
-  });
-
-  it.each([
-    true,
-    false,
-  ])("classifies a DOM.focus rejection on a disabled=%s target", async (disabled) => {
-    const h = await setup(`<input ${disabled ? "disabled" : ""} value="old">`);
-    const send = h.send.getMockImplementation()!;
-    h.send.mockImplementation(async (tabId, method, params) => {
-      if (method === "DOM.focus") throw new Error("Element is not focusable");
-      return send(tabId, method, params);
-    });
-    expect(await h.fill()).toMatchObject({
-      code: disabled ? "invalid_params" : "cdp_failed",
-      data: { reason: disabled ? "target_not_fillable" : "fill_failed" },
-    });
-    expect(h.element.value).toBe("old");
-    expect(h.insert).not.toHaveBeenCalled();
-    expect(h.release).toHaveBeenCalledOnce();
-  });
-
-  it("fills a background document without requiring foreground browser focus", async () => {
-    const h = await setup();
-    vi.spyOn(document, "hasFocus").mockReturnValue(false);
-    expect(await h.fill()).toMatchObject({ value_length: 5 });
-    expect(h.element.value).toBe("hello");
-  });
-
-  it("restores focus once after clearing without typing into another field", async () => {
-    const h = await setup();
-    const other = document.createElement("input");
-    document.body.append(other);
-    h.element.addEventListener("input", () => other.focus());
-    expect(await h.fill()).toMatchObject({ value_length: 5 });
-    expect(h.element.value).toBe("hello");
-    expect(other.value).toBe("");
-    expect(h.send.mock.calls.filter(([, method]) => method === "DOM.focus")).toHaveLength(2);
-  });
-
-  it.each([
-    "focus",
-    "value",
-  ])("rechecks %s after restoring focus, without retrying again", async (action) => {
-    const h = await setup();
-    const other = document.createElement("input");
-    document.body.append(other);
-    h.element.addEventListener(
-      "input",
-      () => {
-        other.focus();
-        h.element.addEventListener("focus", () => {
-          if (action === "focus") queueMicrotask(() => other.focus());
-          else h.element.value = "changed";
-        });
-      },
-      { once: true },
-    );
-    expect(await h.fill()).toMatchObject({
-      code: "cdp_failed",
-      data: { reason: action === "focus" ? "fill_focus_lost" : "fill_target_changed" },
-    });
-    expect(other.value).toBe("");
-    expect(h.insert).not.toHaveBeenCalled();
-    expect(h.send.mock.calls.filter(([, method]) => method === "DOM.focus")).toHaveLength(2);
-  });
-
-  it.each([
-    "remove",
-    "readonly",
-    "restore",
-    "microtask-restore",
-  ])("stops when clearing causes the page to %s the target", async (action) => {
-    const h = await setup();
-    h.element.addEventListener("input", () => {
-      if (action === "remove") h.element.remove();
-      if (action === "readonly") h.element.readOnly = true;
-      if (action === "restore") h.element.value = "restored";
-      if (action === "microtask-restore")
-        queueMicrotask(() => {
-          h.element.value = "restored";
-        });
-    });
-    expect(await h.fill()).toMatchObject({
-      code: "cdp_failed",
-      data: { reason: "fill_target_changed" },
-    });
-    expect(h.insert).not.toHaveBeenCalled();
-  });
-
-  it("fails when insertText is acknowledged without changing the value", async () => {
-    const h = await setup();
-    h.insert.mockImplementation(async () => {});
-    expect(await h.fill()).toMatchObject({
-      code: "cdp_failed",
-      data: { reason: "fill_value_mismatch" },
-    });
-    expect(h.release).toHaveBeenCalledOnce();
-  });
-
-  it("detects truncation when maxlength changes after preparation", async () => {
-    const h = await setup();
-    h.element.addEventListener(
-      "input",
-      () => {
-        h.element.maxLength = 3;
-      },
-      { once: true },
-    );
-    expect(await h.fill()).toMatchObject({
-      code: "cdp_failed",
-      data: { reason: "fill_value_mismatch" },
-    });
-    expect(h.element.value).toBe("hel");
-  });
-
-  it.each([
-    "input",
-    "change",
-    "microtask-change",
-  ])("detects a value reverted by the %s handler", async (event) => {
-    const h = await setup();
-    h.element.addEventListener(event === "microtask-change" ? "change" : event, () => {
-      if (event === "microtask-change")
-        queueMicrotask(() => {
-          h.element.value = "";
-        });
-      else h.element.value = "";
-    });
-    expect(await h.fill()).toMatchObject({ code: "cdp_failed", message: /expected value/ });
-  });
-
-  it("compares content rather than length without including field values in the error", async () => {
-    const h = await setup('<input type="password">');
-    h.element.addEventListener("change", () => {
-      h.element.value = "other-secret";
-    });
-    const result = await h.fill("given-secret");
-    expect(result).toMatchObject({ code: "cdp_failed", message: /expected value/ });
-    expect(JSON.stringify(result)).not.toMatch(/given-secret|other-secret/);
-  });
-
-  it("does not report success for a target replaced by a change handler", async () => {
-    const h = await setup();
-    h.element.addEventListener("change", () => h.element.replaceWith(h.element.cloneNode()));
-    expect(await h.fill()).toMatchObject({
-      code: "cdp_failed",
-      data: { reason: "fill_target_changed" },
-    });
-  });
-
-  it.each([
-    1, 2, 3, 4, 5,
-  ])("rejects script exceptions at phase %s and releases the target", async (phase) => {
-    const h = await setup();
-    vi.spyOn(document, "hasFocus").mockReturnValue(false);
-    const run = h.script.getMockImplementation()!;
-    let count = 0;
-    h.script.mockImplementation(async (params) =>
-      ++count === phase ? { exceptionDetails: { text: "page secret" } } : run(params),
-    );
-    const result = await h.fill();
-    expect(result).toMatchObject({ code: "cdp_failed", data: { reason: "fill_failed" } });
-    expect(JSON.stringify(result)).not.toContain("page secret");
-    if (phase <= 3) expect(h.insert).not.toHaveBeenCalled();
-    expect(h.release).toHaveBeenCalledOnce();
-  });
-
-  it.each([1, 2, 3, 5])("rejects missing results at phase %s", async (phase) => {
-    const h = await setup();
-    vi.spyOn(document, "hasFocus").mockReturnValue(false);
-    const run = h.script.getMockImplementation()!;
-    let count = 0;
-    h.script.mockImplementation(async (params) =>
-      ++count === phase ? { result: { type: "undefined" } } : run(params),
-    );
-    expect(await h.fill()).toMatchObject({ code: "cdp_failed" });
-    if (phase <= 3) expect(h.insert).not.toHaveBeenCalled();
-  });
-
-  it.each(["input", "change"])("checks state after nested microtasks from %s", async (event) => {
-    const h = await setup();
-    h.element.addEventListener(event, () => {
-      queueMicrotask(() =>
-        queueMicrotask(() => {
-          h.element.value = "reverted";
-        }),
-      );
-    });
-    expect(await h.fill()).toMatchObject({ code: "cdp_failed" });
-    if (event === "input") expect(h.insert).not.toHaveBeenCalled();
-  });
-
-  it.each([
-    "<input>",
-    '<input type="password">',
-    "<textarea></textarea>",
-  ])("verifies successful text in %s", async (markup) => {
-    const h = await setup(markup);
-    expect(await h.fill("你好🙂")).toMatchObject({ value_length: 4 });
-    expect(h.element.value).toBe("你好🙂");
-    expect(h.release).toHaveBeenCalledOnce();
-  });
-
-  it("verifies append against the old value plus the requested text", async () => {
-    const h = await setup();
-    h.element.focus();
-    h.element.setSelectionRange(3, 3);
-    expect(await h.fill("!", false)).toMatchObject({ value_length: 4 });
-    expect(h.element.value).toBe("old!");
-  });
-
-  it.each([
-    ["<textarea></textarea>", "a\r\nb", "a\nb"],
-    ["<input>", "a\nb", "a b"],
-    ["<input>", "a\r\nb", "a b"],
-    [
-      '<input type="email" multiple>',
-      " a@example.com , b@example.com ",
-      "a@example.com,b@example.com",
-    ],
-  ])("accepts native normalization for %s", async (markup, requested, normalized) => {
-    if (markup.startsWith("<textarea")) {
-      // Happy DOM does not implement Chrome's textarea line-ending normalization.
-      const setter = Object.getOwnPropertyDescriptor(HTMLTextAreaElement.prototype, "value")!.set!;
-      vi.spyOn(HTMLTextAreaElement.prototype, "value", "set").mockImplementation(function (
-        this: HTMLTextAreaElement,
-        value,
-      ) {
-        setter.call(this, value.replace(/\r\n?/g, "\n"));
-      });
-    }
-    const h = await setup(markup);
-    h.insert.mockImplementation(async () => {
-      h.element.value = normalized;
-    });
-    expect(await h.fill(requested)).toMatchObject({ value_length: normalized.length });
-  });
-
-  it("does not count an editable padding break as an extra typed character", async () => {
-    const h = await setup('<div contenteditable="true" tabindex="0"></div>');
-    // These two layout/editability properties are missing in Happy DOM.
-    let innerText = "";
-    Object.defineProperty(h.element, "isContentEditable", { get: () => true });
-    Object.defineProperty(h.element, "innerText", { get: () => innerText });
-    h.insert.mockImplementation(async () => {
-      h.element.innerHTML = "one<div><br></div>";
-      innerText = "one\n\n";
-    });
-    expect(await h.fill("one\n")).toMatchObject({ value_length: 4 });
-  });
-
-  it.each([
-    [0, 0],
-    [1, 1],
-    [0, 3],
-  ])("appends after moving selection %s:%s to the end", async (start, end) => {
-    const h = await setup();
-    h.element.focus();
-    h.element.setSelectionRange(start, end);
-    expect(await h.fill("🙂", false)).toMatchObject({ value_length: 5 });
-    expect(h.element.value).toBe("old🙂");
-  });
-
-  it.each(
-    ["number", "email"].flatMap((type) =>
-      [
-        [0, 0],
-        [1, 1],
-        [0, 2],
-      ].map(([start, end]) => ({ type, start, end })),
-    ),
-  )("appends to $type from selection $start:$end without using its unsupported API", async ({
-    type,
-    start,
-    end,
-  }) => {
-    const h = await setup(`<input type="${type}" value="12">`);
-    Object.assign(h.hiddenCaret, { start, end });
-    vi.spyOn(h.element, "selectionStart", "get").mockReturnValue(null);
-    vi.spyOn(h.element, "selectionEnd", "get").mockReturnValue(null);
-    const selection = vi.spyOn(h.element, "setSelectionRange").mockImplementation(() => {
-      throw new DOMException("unsupported", "InvalidStateError");
-    });
-    expect(await h.fill("3", false)).toMatchObject({ value_length: 3 });
-    expect(h.element.value).toBe("123");
-    expect(selection).not.toHaveBeenCalled();
-  });
-
-  it.each([
-    "number",
-    "email",
-  ])("does not report successful %s append when the caret command is ignored", async (type) => {
-    const h = await setup(`<input type="${type}" value="12">`);
-    vi.spyOn(h.element, "selectionStart", "get").mockReturnValue(null);
-    vi.spyOn(h.element, "selectionEnd", "get").mockReturnValue(null);
-    h.moveToEnd.mockResolvedValueOnce(undefined);
-    expect(await h.fill("3", false)).toMatchObject({
-      code: "cdp_failed",
-      data: { reason: "fill_value_mismatch" },
-    });
-    expect(h.element.value).toBe("312");
-    expect(h.insert).toHaveBeenCalledOnce();
-  });
-
-  it("rejects invalid number text before clearing", async () => {
-    const h = await setup('<input type="number" value="12">');
-    expect(await h.fill("abc")).toMatchObject({
-      code: "invalid_params",
-      data: { reason: "fill_value_invalid" },
-    });
-    expect(h.element.value).toBe("12");
-    expect(h.insert).not.toHaveBeenCalled();
-  });
-
-  it("reports page formatting as unconfirmed without retrying the write", async () => {
-    const h = await setup();
-    h.element.addEventListener("change", () => {
-      h.element.value = h.element.value.toUpperCase();
-    });
-    expect(await h.fill()).toMatchObject({
-      code: "cdp_failed",
-      data: { reason: "fill_value_mismatch" },
-    });
-    expect(h.element.value).toBe("HELLO");
-    expect(h.insert).toHaveBeenCalledOnce();
-  });
-
-  it.each([
-    ["focus", false],
-    ["focus", true],
-    ["value", false],
-    ["value", true],
-    ["cancel", false],
-    ["cancel", true],
-    ["throw", false],
-    ["throw", true],
-  ] as const)("stops when positioning triggers %s (clear=%s) and releases the key", async (action, clearBefore) => {
-    const h = await setup();
-    vi.spyOn(document, "hasFocus").mockReturnValue(false);
-    const controller = new AbortController();
-    const other = document.createElement("input");
-    document.body.append(other);
-    const send = h.send.getMockImplementation()!;
-    h.send.mockImplementation(async (tabId, method, params) => {
-      const reply = await send(tabId, method, params);
-      if (
-        method === "Input.dispatchKeyEvent" &&
-        (params as { type: string }).type === "rawKeyDown"
-      ) {
-        if (action === "focus") other.focus();
-        if (action === "value") h.element.value = "changed";
-        if (action === "cancel") controller.abort();
-        if (action === "throw") throw new Error("dispatch failed");
-      }
-      return reply;
-    });
-    const result = await h.fill("!", clearBefore, controller.signal);
-    expect(result).toMatchObject({ code: action === "cancel" ? "cancelled" : "cdp_failed" });
-    expect(h.insert).not.toHaveBeenCalled();
-    expect(other.value).toBe("");
-    expect(h.send).toHaveBeenCalledWith(
-      4,
-      "Input.dispatchKeyEvent",
-      expect.objectContaining({ type: "keyUp" }),
-    );
-    expect(h.release).toHaveBeenCalledOnce();
-  });
-
-  it.each([true, false])("verifies an empty request with clear_before=%s", async (clearBefore) => {
-    const h = await setup();
-    expect(await h.fill("", clearBefore)).toMatchObject({ value_length: clearBefore ? 0 : 3 });
-    expect(h.element.value).toBe(clearBefore ? "" : "old");
-    expect(h.insert).not.toHaveBeenCalled();
-  });
-
-  it("cancels after insertion without sending change events", async () => {
-    const h = await setup();
-    const controller = new AbortController();
-    h.insert.mockImplementation(async () => controller.abort());
-    expect(await h.fill("hello", true, controller.signal)).toMatchObject({ code: "cancelled" });
-    expect(h.script).toHaveBeenCalledTimes(2);
-    expect(h.release).toHaveBeenCalledOnce();
-  });
-
-  it("cancels after resolving the remote object without clearing the field", async () => {
-    const h = await setup();
-    const controller = new AbortController();
-    const send = h.send.getMockImplementation()!;
-    h.send.mockImplementation(async (tabId, method, params) => {
-      const result = await send(tabId, method, params);
-      if (method === "DOM.resolveNode") controller.abort();
-      return result;
-    });
-    expect(await h.fill("hello", true, controller.signal)).toMatchObject({ code: "cancelled" });
-    expect(h.element.value).toBe("old");
-    expect(h.script).not.toHaveBeenCalled();
-    expect(h.release).toHaveBeenCalledOnce();
-  });
-
-  it("honors cancellation during verification", async () => {
-    const h = await setup();
-    const controller = new AbortController();
-    h.element.addEventListener("change", () => controller.abort());
-    expect(await h.fill("hello", true, controller.signal)).toMatchObject({ code: "cancelled" });
-    expect(h.release).toHaveBeenCalledOnce();
-  });
-
-  it("keeps the verified outcome when releasing the remote object fails", async () => {
-    const h = await setup();
-    h.release.mockRejectedValueOnce(new Error("target closed"));
-    expect(await h.fill()).toMatchObject({ value_length: 5 });
-  });
-});

apps/extension/src/tools/__tests__/frame-geometry.test.ts

@@ -169,7 +169,6 @@ describe("frame geometry projection", () => {
     expect(geometry).toMatchObject({
       topBounds: { x: 100, y: 100, width: 200, height: 100 },
       actionPoint: { x: 200, y: 150 },
-      targetActionPoint: { x: 100, y: 50 },
     });
   });
 
@@ -262,7 +261,6 @@ describe("frame geometry projection", () => {
     expect(geometry).toMatchObject({
       topBounds: { x: 100, y: 100, width: 200, height: 100 },
       actionPoint: { x: 200, y: 150 },
-      targetActionPoint: { x: 200, y: 150 },
     });
   });
 

apps/extension/src/tools/__tests__/interaction.test.ts

@@ -7,10 +7,12 @@ import {
   handleHover,
   handlePress,
   handleSelect,
+  handleType,
   modifiersBitfield,
   parseKeySpec,
   resolveKeyDescriptor,
 } from "../interaction";
+import { captureSurfaceEnvironment } from "../surface-target";
 
 function fakeAgentWindow(ids: number[]) {
   let i = 0;
@@ -21,7 +23,7 @@ function fakeAgentWindow(ids: number[]) {
       return id;
     }),
     remove: vi.fn(async () => {}),
-    ensureActiveTab: vi.fn(async () => 1),
+    ensureActiveTab: vi.fn(async () => {}),
   };
 }
 
@@ -33,6 +35,15 @@ function makeFakeCdp(handlers: Record<string, (params: unknown) => unknown>) {
     if (!h && method === "Page.getLayoutMetrics") {
       return { cssLayoutViewport: { clientWidth: 1280, clientHeight: 720 } };
     }
+    if (
+      !h &&
+      method === "Runtime.evaluate" &&
+      String((params as { expression?: string })?.expression ?? "").includes(
+        "bsk-focus-target-probe",
+      )
+    ) {
+      return { result: { value: { hasFocus: true, activeTag: "input" } } };
+    }
     if (!h) throw new Error(`unexpected CDP call ${method}`);
     return h(params);
   };
@@ -44,6 +55,10 @@ function makeFakeCdp(handlers: Record<string, (params: unknown) => unknown>) {
       params?: object,
     ) => Promise<T>,
     trackSessionTab: vi.fn(),
+    getFrameGraph: vi.fn(async (tabId) => ({
+      rootFrameId: "main",
+      frames: [{ frameId: "main", target: { tabId } }],
+    })),
   };
   const tabsApi = {
     get: vi.fn(
@@ -54,6 +69,84 @@ function makeFakeCdp(handlers: Record<string, (params: unknown) => unknown>) {
   return { cdp, tabsApi, sent };
 }
 
+function fillRuntimeReadback(value: string) {
+  return (params: unknown) => {
+    const fn = String((params as { functionDeclaration?: string }).functionDeclaration ?? "");
+    if (fn.includes("bsk-editable-state")) {
+      return {
+        result: {
+          value: {
+            supported: true,
+            connected: true,
+            focused: true,
+            value,
+            kind: "input",
+          },
+        },
+      };
+    }
+    if (fn.includes("bsk-editable-selection")) {
+      return { result: { value: { selected: true } } };
+    }
+    return { result: { type: "undefined" } };
+  };
+}
+
+async function makeSurfaceClickFixture(onMouse?: (params: Record<string, unknown>) => void) {
+  const manager = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
+  const context = await manager.start("aa11");
+  context.refStore.set("e3", 99, {
+    tabId: 4,
+    kind: "surface",
+    visibleRect: { x: 10, y: 20, w: 100, h: 40 },
+  });
+  const fake = makeFakeCdp({
+    "Page.getFrameTree": () => ({
+      frameTree: {
+        frame: { id: "main", loaderId: "loader-1", url: "https://fixture.test/" },
+      },
+    }),
+    "Page.getLayoutMetrics": () => ({
+      cssLayoutViewport: { clientWidth: 800, clientHeight: 600, pageX: 0, pageY: 0 },
+      cssVisualViewport: {
+        clientWidth: 800,
+        clientHeight: 600,
+        pageX: 0,
+        pageY: 0,
+        scale: 1,
+      },
+    }),
+    "DOM.getContentQuads": () => ({ quads: [[10, 20, 110, 20, 110, 60, 10, 60]] }),
+    "Runtime.evaluate": () => ({
+      result: { value: { overlayHostPresent: false, overlayHostConnected: false } },
+    }),
+    "Input.dispatchMouseEvent": (params) => {
+      onMouse?.(params as Record<string, unknown>);
+      return {};
+    },
+  });
+  const environment = await captureSurfaceEnvironment(fake.cdp, 4, undefined);
+  if ("code" in environment) throw new Error(environment.message);
+  const entry = context.refStore.resolveEntry("e3");
+  if (!entry) throw new Error("missing Surface ref");
+  const capture = context.surfaceCaptures.create({
+    sessionId: "aa11",
+    tabId: 4,
+    navigationIdentity: environment.navigationIdentity,
+    surface: {
+      ref: "e3",
+      backendNodeId: 99,
+      observationGeneration: entry.generation,
+    },
+    topViewportRect: { x: 10, y: 20, w: 100, h: 40 },
+    imageWidth: 200,
+    imageHeight: 80,
+    viewportSignature: environment.viewportSignature,
+    frameProjectionSignature: environment.frameProjectionSignature,
+  });
+  return { manager, context, fake, capture };
+}
+
 describe("modifiersBitfield", () => {
   it("matches CDP's expected bit layout", () => {
     expect(modifiersBitfield([])).toBe(0);
@@ -123,6 +216,29 @@ describe("handleClick", () => {
     expect(fake.cdp.send).not.toHaveBeenCalled();
   });
 
+  it("rejects screenshot-only visual surface refs before issuing CDP calls", async () => {
+    const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
+    const ctx = await sm.start("aa11");
+    ctx.refStore.set("e3", 1234, {
+      tabId: 4,
+      kind: "surface",
+      capabilities: ["screenshot"],
+    });
+    const fake = makeFakeCdp({});
+
+    const res = await handleClick(
+      sm,
+      { session_id: "aa11", ref: "@e3" },
+      { cdp: fake.cdp, tabsApi: fake.tabsApi },
+    );
+
+    expect(res).toMatchObject({
+      code: "permission_denied",
+      data: { reason: "ref_capability_denied", required_capability: "interact" },
+    });
+    expect(fake.cdp.send).not.toHaveBeenCalled();
+  });
+
   it("clicks by ref, computes the quad centre, dispatches three mouse events", async () => {
     const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
     const ctx = await sm.start("aa11");
@@ -167,6 +283,74 @@ describe("handleClick", () => {
     });
   });
 
+  it.each([
+    "left",
+    "middle",
+    "right",
+  ] as const)("uses the common click executor for a %s-button Surface click", async (button) => {
+    const { manager, fake, capture } = await makeSurfaceClickFixture();
+    const result = await handleClick(
+      manager,
+      {
+        session_id: "aa11",
+        ref: "@e3",
+        capture_id: capture.id,
+        image_x: 50,
+        image_y: 20,
+        button,
+        click_count: 2,
+        modifiers: ["ctrl", "shift"],
+      },
+      { cdp: fake.cdp, tabsApi: fake.tabsApi },
+    );
+
+    if ("code" in result) throw new Error(JSON.stringify(result));
+    expect(result).toMatchObject({
+      used_ref: "e3",
+      capture_id: capture.id,
+      image_x: 50,
+      image_y: 20,
+      x: 35,
+      y: 30,
+    });
+    const mouse = fake.sent.filter((call) => call.method === "Input.dispatchMouseEvent");
+    expect(mouse).toHaveLength(3);
+    expect(mouse[1].params).toMatchObject({
+      type: "mousePressed",
+      button,
+      clickCount: 2,
+      modifiers: 2 | 8,
+    });
+    expect(mouse[2].params).toMatchObject({
+      type: "mouseReleased",
+      button,
+      clickCount: 2,
+    });
+  });
+
+  it("consumes a Surface capture when the common click executor fails", async () => {
+    let mouseCalls = 0;
+    const { manager, fake, capture } = await makeSurfaceClickFixture((params) => {
+      mouseCalls += 1;
+      if (params.type === "mousePressed") throw new Error("input failed");
+    });
+    const params = {
+      session_id: "aa11",
+      ref: "@e3",
+      capture_id: capture.id,
+      image_x: 50,
+      image_y: 20,
+    };
+
+    expect(
+      await handleClick(manager, params, { cdp: fake.cdp, tabsApi: fake.tabsApi }),
+    ).toMatchObject({ code: "cdp_failed" });
+    expect(mouseCalls).toBe(2);
+    expect(
+      await handleClick(manager, params, { cdp: fake.cdp, tabsApi: fake.tabsApi }),
+    ).toMatchObject({ data: { reason: "surface_capture_consumed" } });
+  });
+
   it("resolves frame refs in their CDP session and dispatches input in top coordinates", async () => {
     const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
     const ctx = await sm.start("aa11");
@@ -540,21 +724,6 @@ describe("handleHover", () => {
   });
 });
 
-function successfulFillScript(params: unknown) {
-  const script = params as { arguments?: Array<{ value: unknown }>; functionDeclaration: string };
-  const args = script.arguments ?? [];
-  return {
-    result: {
-      value:
-        args.length === 2
-          ? { before: "", expected: args[0].value }
-          : script.functionDeclaration.startsWith("function(expected)")
-            ? { connected: true, matches: true, valueLength: String(args[0].value).length }
-            : "ready",
-    },
-  };
-}
-
 describe("handleFill", () => {
   it("returns not_found for unknown ref", async () => {
     const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
@@ -601,8 +770,7 @@ describe("handleFill", () => {
       "DOM.scrollIntoViewIfNeeded": () => ({}),
       "DOM.focus": () => ({}),
       "DOM.resolveNode": () => ({ object: { objectId: "obj-1" } }),
-      "Runtime.callFunctionOn": successfulFillScript,
-      "Runtime.releaseObject": () => ({}),
+      "Runtime.callFunctionOn": fillRuntimeReadback("hello"),
       "Input.insertText": () => ({}),
     });
     const res = await handleFill(
@@ -616,12 +784,13 @@ describe("handleFill", () => {
     expect(res.used_ref).toBe("e1");
     const insert = fake.sent.find((c) => c.method === "Input.insertText");
     expect(insert?.params).toEqual({ text: "hello" });
-    // Foreground replacement needs no extra caret-positioning round trip.
+    // clear_before defaults to true, so the edit transaction prepares a
+    // replacement selection and performs a settled live-value readback.
     const callFns = fake.sent.filter((c) => c.method === "Runtime.callFunctionOn");
-    expect(callFns).toHaveLength(4);
+    expect(callFns.length).toBeGreaterThanOrEqual(2);
   });
 
-  it("passes clear_before=false to preparation and verifies the result", async () => {
+  it("clear_before=false skips replacement selection", async () => {
     const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
     const ctx = await sm.start("aa11");
     ctx.refStore.set("e1", 1, { tabId: 4 });
@@ -633,12 +802,23 @@ describe("handleFill", () => {
       "DOM.focus": () => ({}),
       "DOM.resolveNode": () => ({ object: { objectId: "obj-2" } }),
       "Runtime.callFunctionOn": (p) => {
-        const args = (p as { arguments?: Array<{ value: unknown }> }).arguments ?? [];
-        if (args.length === 2) expect(args[1].value).toBe(false);
-        return successfulFillScript(p);
+        const fn = (p as { functionDeclaration?: string }).functionDeclaration ?? "";
+        if (fn.includes("bsk-editable-state")) {
+          return {
+            result: {
+              value: {
+                supported: true,
+                connected: true,
+                focused: true,
+                value: "existingx",
+                kind: "textarea",
+              },
+            },
+          };
+        }
+        expect(fn).not.toMatch(/dispatchEvent|\.value\s*=/);
+        return { result: { type: "undefined" } };
       },
-      "Input.dispatchKeyEvent": () => ({}),
-      "Runtime.releaseObject": () => ({}),
       "Input.insertText": () => ({}),
     });
     const res = await handleFill(
@@ -647,7 +827,7 @@ describe("handleFill", () => {
       { cdp: fake.cdp, tabsApi: fake.tabsApi },
     );
     if ("code" in res) throw new Error(`unexpected error: ${JSON.stringify(res)}`);
-    expect(res.value_length).toBe(1);
+    expect(res.value_length).toBe(9);
   });
 
   it("treats contenteditable=true as fillable", async () => {
@@ -665,8 +845,7 @@ describe("handleFill", () => {
       "DOM.scrollIntoViewIfNeeded": () => ({}),
       "DOM.focus": () => ({}),
       "DOM.resolveNode": () => ({ object: { objectId: "obj-3" } }),
-      "Runtime.callFunctionOn": successfulFillScript,
-      "Runtime.releaseObject": () => ({}),
+      "Runtime.callFunctionOn": fillRuntimeReadback("rich"),
       "Input.insertText": () => ({}),
     });
     const res = await handleFill(
@@ -677,7 +856,103 @@ describe("handleFill", () => {
     expect("code" in res).toBe(false);
   });
 
-  it("stops before Input.insertText when abort fires after clearing", async () => {
+  it("routes OOPIF focus, text, DOM events, and readback through the node target", async () => {
+    const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
+    const ctx = await sm.start("aa11");
+    ctx.refStore.set("e1", 77, {
+      tabId: 4,
+      frameId: "child-frame",
+      cdpSessionId: "child-session",
+    });
+    const fake = makeFakeCdp({ "DOM.scrollIntoViewIfNeeded": () => ({}) });
+    fake.cdp.getFrameGraph = vi.fn(async () => ({
+      rootFrameId: "main",
+      frames: [
+        { frameId: "main", target: { tabId: 4 } },
+        {
+          frameId: "child-frame",
+          parentFrameId: "main",
+          ownerBackendNodeId: 12,
+          target: { tabId: 4, sessionId: "child-session" },
+        },
+      ],
+    }));
+    const childCalls: string[] = [];
+    fake.cdp.sendToTarget = vi.fn(async (target, method, params) => {
+      expect(target).toEqual({ tabId: 4, sessionId: "child-session" });
+      childCalls.push(method);
+      if (method === "DOM.describeNode") {
+        return { node: { backendNodeId: 77, nodeName: "INPUT", attributes: [] } };
+      }
+      if (method === "DOM.scrollIntoViewIfNeeded" || method === "DOM.focus") return {};
+      if (method === "DOM.resolveNode") return { object: { objectId: "child-input" } };
+      if (method === "Input.insertText") {
+        expect(params).toEqual({ text: "中文输入" });
+        return {};
+      }
+      if (method === "Runtime.callFunctionOn") {
+        const fn = String((params as { functionDeclaration?: string }).functionDeclaration ?? "");
+        if (fn.includes("bsk-editable-state")) {
+          return {
+            result: {
+              value: {
+                supported: true,
+                connected: true,
+                focused: true,
+                value: "中文输入",
+                kind: "input",
+              },
+            },
+          };
+        }
+        if (fn.includes("bsk-editable-selection")) {
+          return { result: { value: { selected: true } } };
+        }
+        return { result: { type: "undefined" } };
+      }
+      throw new Error(`unexpected child call ${method}`);
+    }) as CdpRunner["sendToTarget"];
+
+    const res = await handleFill(
+      sm,
+      { session_id: "aa11", ref: "@e1", value: "中文输入" },
+      { cdp: fake.cdp, tabsApi: fake.tabsApi },
+    );
+
+    if ("code" in res) throw new Error(JSON.stringify(res));
+    expect(res.value_length).toBe(4);
+    expect(childCalls).toContain("DOM.focus");
+    expect(childCalls).toContain("Input.insertText");
+    expect(childCalls.filter((method) => method === "Runtime.callFunctionOn")).toHaveLength(4);
+    expect(fake.sent.some((call) => call.method === "Input.insertText")).toBe(false);
+  });
+
+  it("returns input_not_applied when a controlled input resets the value", async () => {
+    const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
+    const ctx = await sm.start("aa11");
+    ctx.refStore.set("e1", 8, { tabId: 4 });
+    const fake = makeFakeCdp({
+      "DOM.describeNode": () => ({ node: { backendNodeId: 8, nodeName: "INPUT" } }),
+      "DOM.scrollIntoViewIfNeeded": () => ({}),
+      "DOM.focus": () => ({}),
+      "DOM.resolveNode": () => ({ object: { objectId: "controlled" } }),
+      "Input.insertText": () => ({}),
+      "Runtime.callFunctionOn": fillRuntimeReadback(""),
+    });
+
+    const res = await handleFill(
+      sm,
+      { session_id: "aa11", ref: "@e1", value: "rejected" },
+      { cdp: fake.cdp, tabsApi: fake.tabsApi },
+    );
+
+    expect(res).toMatchObject({
+      code: "cdp_failed",
+      data: { reason: "input_not_applied", expected_value_length: 8, actual_value_length: 0 },
+    });
+  });
+
+  it("stops before Input.insertText when abort fires after preparing replacement", async () => {
     const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
     const ctx = await sm.start("aa11");
     ctx.refStore.set("e1", 555, { tabId: 4 });
@@ -689,8 +964,25 @@ describe("handleFill", () => {
       "DOM.scrollIntoViewIfNeeded": () => ({}),
       "DOM.focus": () => ({}),
       "DOM.resolveNode": () => ({ object: { objectId: "obj-1" } }),
-      "Runtime.callFunctionOn": () => {
-        abort.abort();
+      "Runtime.callFunctionOn": (params) => {
+        const fn = String((params as { functionDeclaration?: string }).functionDeclaration ?? "");
+        if (fn.includes("bsk-editable-state")) {
+          return {
+            result: {
+              value: {
+                supported: true,
+                connected: true,
+                focused: true,
+                value: "",
+                kind: "input",
+              },
+            },
+          };
+        }
+        if (fn.includes("bsk-editable-selection")) {
+          abort.abort();
+          return { result: { value: { selected: true } } };
+        }
         return { result: { type: "undefined" } };
       },
       "Input.insertText": () => ({}),
@@ -733,6 +1025,204 @@ describe("handleFill", () => {
   });
 });
 
+describe("handleType", () => {
+  it("requires an explicit target mode instead of silently using focus", async () => {
+    const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
+    await sm.start("aa11");
+    const fake = makeFakeCdp({});
+
+    const missing = await handleType(
+      sm,
+      { session_id: "aa11", text: "x" },
+      { cdp: fake.cdp, tabsApi: fake.tabsApi },
+    );
+    expect(missing).toMatchObject({
+      code: "invalid_params",
+      message: expect.stringContaining("focused=true"),
+    });
+
+    const mixed = await handleType(
+      sm,
+      { session_id: "aa11", ref: "@e1", focused: true, text: "x" },
+      { cdp: fake.cdp, tabsApi: fake.tabsApi },
+    );
+    expect(mixed).toMatchObject({
+      code: "invalid_params",
+      message: expect.stringContaining("not both"),
+    });
+    expect(fake.sent).toHaveLength(0);
+  });
+
+  it("routes committed IME text through the explicit OOPIF target without DOM readback", async () => {
+    const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
+    const ctx = await sm.start("aa11");
+    ctx.refStore.set("e1", 77, {
+      tabId: 4,
+      frameId: "child-frame",
+      cdpSessionId: "child-session",
+    });
+    const fake = makeFakeCdp({ "DOM.scrollIntoViewIfNeeded": () => ({}) });
+    fake.cdp.getFrameGraph = vi.fn(async () => ({
+      rootFrameId: "main",
+      frames: [
+        { frameId: "main", target: { tabId: 4 } },
+        {
+          frameId: "child-frame",
+          parentFrameId: "main",
+          ownerBackendNodeId: 12,
+          target: { tabId: 4, sessionId: "child-session" },
+        },
+      ],
+    }));
+    const childCalls: Array<{ method: string; params?: object }> = [];
+    fake.cdp.sendToTarget = vi.fn(async (_target, method, params) => {
+      childCalls.push({ method, params });
+      if (method === "DOM.describeNode") {
+        return { node: { backendNodeId: 77, nodeName: "TEXTAREA", attributes: [] } };
+      }
+      return {};
+    }) as CdpRunner["sendToTarget"];
+
+    const result = await handleType(
+      sm,
+      { session_id: "aa11", ref: "@e1", text: "测试" },
+      { cdp: fake.cdp, tabsApi: fake.tabsApi },
+    );
+
+    if ("code" in result) throw new Error(JSON.stringify(result));
+    expect(result).toMatchObject({
+      tab_id: 4,
+      used_ref: "e1",
+      text_length: 2,
+      verification: "dispatched",
+    });
+    expect(childCalls.map((call) => call.method)).toEqual([
+      "DOM.describeNode",
+      "DOM.scrollIntoViewIfNeeded",
+      "DOM.focus",
+      "Input.dispatchKeyEvent",
+      "Input.insertText",
+      "Input.dispatchKeyEvent",
+    ]);
+    expect(fake.sent.some((call) => call.method.startsWith("Input."))).toBe(false);
+    expect(childCalls.some((call) => call.method === "Runtime.callFunctionOn")).toBe(false);
+  });
+
+  it("uses the focused OOPIF only when focused mode resolves a text-entry target", async () => {
+    const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
+    await sm.start("aa11");
+    const fake = makeFakeCdp({
+      "Runtime.evaluate": () => ({
+        result: { value: { hasFocus: true, activeTag: "iframe", textEntry: false } },
+      }),
+    });
+    fake.cdp.getFrameGraph = vi.fn(async () => ({
+      rootFrameId: "main",
+      frames: [
+        { frameId: "main", target: { tabId: 4 } },
+        {
+          frameId: "child-frame",
+          parentFrameId: "main",
+          target: { tabId: 4, sessionId: "child-session" },
+        },
+      ],
+    }));
+    const childCalls: string[] = [];
+    fake.cdp.sendToTarget = vi.fn(async (_target, method) => {
+      childCalls.push(method);
+      if (method === "Runtime.evaluate") {
+        return {
+          result: { value: { hasFocus: true, activeTag: "textarea", textEntry: true } },
+        };
+      }
+      return {};
+    }) as CdpRunner["sendToTarget"];
+
+    const result = await handleType(
+      sm,
+      { session_id: "aa11", focused: true, text: "A" },
+      { cdp: fake.cdp, tabsApi: fake.tabsApi },
+    );
+
+    if ("code" in result) throw new Error(JSON.stringify(result));
+    expect(result).toMatchObject({ text_length: 1, verification: "dispatched" });
+    expect(childCalls).toEqual([
+      "Runtime.evaluate",
+      "Input.dispatchKeyEvent",
+      "Input.dispatchKeyEvent",
+      "Input.dispatchKeyEvent",
+    ]);
+    expect(fake.sent.some((call) => call.method.startsWith("Input."))).toBe(false);
+  });
+
+  it("refuses focused mode when focus is not owned by a text-entry target", async () => {
+    const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
+    await sm.start("aa11");
+    const fake = makeFakeCdp({
+      "Runtime.evaluate": () => ({
+        result: { value: { hasFocus: true, activeTag: "button", textEntry: false } },
+      }),
+    });
+
+    const result = await handleType(
+      sm,
+      { session_id: "aa11", focused: true, text: "x" },
+      { cdp: fake.cdp, tabsApi: fake.tabsApi },
+    );
+
+    expect(result).toMatchObject({
+      code: "cdp_failed",
+      data: { reason: "focus_target_unresolved", candidate_count: 0 },
+    });
+    expect(fake.sent.some((call) => call.method.startsWith("Input."))).toBe(false);
+  });
+
+  it("rejects explicit non-text input controls", async () => {
+    const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
+    const ctx = await sm.start("aa11");
+    ctx.refStore.set("e1", 77, { tabId: 4 });
+    const fake = makeFakeCdp({
+      "DOM.describeNode": () => ({
+        node: { backendNodeId: 77, nodeName: "INPUT", attributes: ["type", "checkbox"] },
+      }),
+    });
+
+    const result = await handleType(
+      sm,
+      { session_id: "aa11", ref: "@e1", text: "x" },
+      { cdp: fake.cdp, tabsApi: fake.tabsApi },
+    );
+
+    expect(result).toMatchObject({
+      code: "invalid_params",
+      data: { reason: "target_not_fillable" },
+    });
+    expect(fake.sent.some((call) => call.method.startsWith("Input."))).toBe(false);
+  });
+
+  it("rejects a Surface ref instead of treating it as a text target", async () => {
+    const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
+    const ctx = await sm.start("aa11");
+    ctx.refStore.set("e1", 77, {
+      tabId: 4,
+      kind: "surface",
+      capabilities: ["screenshot"],
+    });
+    const fake = makeFakeCdp({});
+
+    const result = await handleType(
+      sm,
+      { session_id: "aa11", ref: "@e1", text: "x" },
+      { cdp: fake.cdp, tabsApi: fake.tabsApi },
+    );
+
+    expect(result).toMatchObject({
+      code: "permission_denied",
+      data: { reason: "ref_capability_denied", required_capability: "interact" },
+    });
+  });
+});
+
 describe("parseKeySpec", () => {
   it("splits compound expressions into modifiers + base key", () => {
     expect(parseKeySpec("Ctrl+A")).toEqual({ key: "A", modifiers: ["ctrl"] });
@@ -871,6 +1361,125 @@ describe("handlePress", () => {
     expect(res.key).toBe("Enter");
   });
 
+  it("routes ref-targeted key events through the OOPIF session", async () => {
+    const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
+    const ctx = await sm.start("aa11");
+    ctx.refStore.set("e1", 555, {
+      tabId: 4,
+      frameId: "child-frame",
+      cdpSessionId: "child-session",
+    });
+    const fake = makeFakeCdp({ "DOM.scrollIntoViewIfNeeded": () => ({}) });
+    fake.cdp.getFrameGraph = vi.fn(async () => ({
+      rootFrameId: "main",
+      frames: [
+        { frameId: "main", target: { tabId: 4 } },
+        {
+          frameId: "child-frame",
+          parentFrameId: "main",
+          ownerBackendNodeId: 12,
+          target: { tabId: 4, sessionId: "child-session" },
+        },
+      ],
+    }));
+    const childCalls: string[] = [];
+    fake.cdp.sendToTarget = vi.fn(async (_target, method) => {
+      childCalls.push(method);
+      return {};
+    }) as CdpRunner["sendToTarget"];
+
+    const res = await handlePress(
+      sm,
+      { session_id: "aa11", key: "Enter", ref: "@e1" },
+      { cdp: fake.cdp, tabsApi: fake.tabsApi },
+    );
+
+    expectPressOk(res);
+    expect(childCalls).toEqual([
+      "DOM.scrollIntoViewIfNeeded",
+      "DOM.focus",
+      "Input.dispatchKeyEvent",
+      "Input.dispatchKeyEvent",
+      "Input.dispatchKeyEvent",
+    ]);
+    expect(fake.sent.some((call) => call.method === "Input.dispatchKeyEvent")).toBe(false);
+  });
+
+  it("routes an untargeted press to the uniquely focused OOPIF session", async () => {
+    const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
+    await sm.start("aa11");
+    const fake = makeFakeCdp({
+      "Runtime.evaluate": () => ({
+        result: { value: { hasFocus: true, activeTag: "iframe" } },
+      }),
+    });
+    fake.cdp.getFrameGraph = vi.fn(async () => ({
+      rootFrameId: "main",
+      frames: [
+        { frameId: "main", target: { tabId: 4 } },
+        {
+          frameId: "child-frame",
+          parentFrameId: "main",
+          target: { tabId: 4, sessionId: "child-session" },
+        },
+      ],
+    }));
+    const childCalls: string[] = [];
+    fake.cdp.sendToTarget = vi.fn(async (_target, method) => {
+      childCalls.push(method);
+      if (method === "Runtime.evaluate") {
+        return { result: { value: { hasFocus: true, activeTag: "textarea" } } };
+      }
+      return {};
+    }) as CdpRunner["sendToTarget"];
+
+    const res = await handlePress(
+      sm,
+      { session_id: "aa11", key: "a" },
+      { cdp: fake.cdp, tabsApi: fake.tabsApi },
+    );
+
+    expectPressOk(res);
+    expect(childCalls).toEqual([
+      "Runtime.evaluate",
+      "Input.dispatchKeyEvent",
+      "Input.dispatchKeyEvent",
+      "Input.dispatchKeyEvent",
+    ]);
+  });
+
+  it("refuses an untargeted press when focus ownership is ambiguous", async () => {
+    const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
+    await sm.start("aa11");
+    const fake = makeFakeCdp({
+      "Runtime.evaluate": () => ({
+        result: { value: { hasFocus: true, activeTag: "input" } },
+      }),
+    });
+    fake.cdp.getFrameGraph = vi.fn(async () => ({
+      rootFrameId: "main",
+      frames: [
+        { frameId: "main", target: { tabId: 4 } },
+        { frameId: "child", target: { tabId: 4, sessionId: "child-session" } },
+      ],
+    }));
+    fake.cdp.sendToTarget = vi.fn(async () => ({
+      result: { value: { hasFocus: true, activeTag: "textarea" } },
+    })) as CdpRunner["sendToTarget"];
+
+    const res = await handlePress(
+      sm,
+      { session_id: "aa11", key: "Enter" },
+      { cdp: fake.cdp, tabsApi: fake.tabsApi },
+    );
+
+    expect(res).toMatchObject({
+      code: "cdp_failed",
+      data: { reason: "focus_target_unresolved", candidate_count: 2 },
+    });
+    expect(fake.sent.some((call) => call.method === "Input.dispatchKeyEvent")).toBe(false);
+  });
+
   it("stops before focus when abort fires after press scroll", async () => {
     const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
     const ctx = await sm.start("aa11");

apps/extension/src/tools/__tests__/mouse-input.test.ts

@@ -0,0 +1,119 @@
+import { describe, expect, it, vi } from "vitest";
+import { dispatchMouseClick } from "../mouse-input";
+import type { CdpRunner } from "../shared";
+
+function fakeCdp(onSend?: (method: string, params: Record<string, unknown>) => void): {
+  cdp: CdpRunner;
+  events: Record<string, unknown>[];
+} {
+  const events: Record<string, unknown>[] = [];
+  return {
+    events,
+    cdp: {
+      send: vi.fn(async (_tabId, method, params) => {
+        const event = params as Record<string, unknown>;
+        events.push(event);
+        onSend?.(method, event);
+        return {} as never;
+      }),
+    },
+  };
+}
+
+describe("dispatchMouseClick", () => {
+  it("emits one move, press, and release sequence", async () => {
+    const { cdp, events } = fakeCdp();
+
+    expect(
+      await dispatchMouseClick(
+        cdp,
+        4,
+        { x: 12, y: 34 },
+        {
+          button: "left",
+          clickCount: 1,
+          modifiers: 0,
+          moveSettleMs: 0,
+        },
+      ),
+    ).toBe("completed");
+
+    expect(events).toEqual([
+      {
+        type: "mouseMoved",
+        x: 12,
+        y: 34,
+        modifiers: 0,
+      },
+      {
+        type: "mousePressed",
+        x: 12,
+        y: 34,
+        button: "left",
+        clickCount: 1,
+        modifiers: 0,
+      },
+      {
+        type: "mouseReleased",
+        x: 12,
+        y: 34,
+        button: "left",
+        clickCount: 1,
+        modifiers: 0,
+      },
+    ]);
+  });
+
+  it("releases a pressed button when cancellation arrives", async () => {
+    const controller = new AbortController();
+    const { cdp, events } = fakeCdp((_method, event) => {
+      if (event.type === "mousePressed") controller.abort();
+    });
+
+    expect(
+      await dispatchMouseClick(
+        cdp,
+        4,
+        { x: 1, y: 2 },
+        {
+          button: "left",
+          clickCount: 1,
+          modifiers: 0,
+          signal: controller.signal,
+          moveSettleMs: 0,
+        },
+      ),
+    ).toBe("cancelled");
+    expect(events.at(-1)).toMatchObject({ type: "mouseReleased" });
+  });
+
+  it("settles pointer-move hit testing before pressing", async () => {
+    vi.useFakeTimers();
+    try {
+      const { cdp, events } = fakeCdp();
+      const click = dispatchMouseClick(
+        cdp,
+        4,
+        { x: 1, y: 2 },
+        {
+          button: "left",
+          clickCount: 1,
+          modifiers: 0,
+          moveSettleMs: 32,
+        },
+      );
+
+      await vi.advanceTimersByTimeAsync(31);
+      expect(events.map((event) => event.type)).toEqual(["mouseMoved"]);
+      await vi.advanceTimersByTimeAsync(1);
+      await click;
+      expect(events.map((event) => event.type)).toEqual([
+        "mouseMoved",
+        "mousePressed",
+        "mouseReleased",
+      ]);
+    } finally {
+      vi.useRealTimers();
+    }
+  });
+});

apps/extension/src/tools/__tests__/navigation-recovery.test.ts

@@ -1,266 +0,0 @@
-import { describe, expect, it, vi } from "vitest";
-import { SessionManager } from "@/session-manager/manager";
-import type { BrowserNavigationApi } from "../browser-navigation";
-import { handleNavigate, handleReload } from "../navigation";
-import type { CdpRunner } from "../shared";
-
-const denied = "Cannot access a chrome-extension:// URL of different extension";
-const url = "https://example.test/recovered";
-
-function event<T extends (...args: never[]) => void>() {
-  const listeners = new Set<T>();
-  return {
-    listeners,
-    addListener: (listener: T) => listeners.add(listener),
-    removeListener: (listener: T) => listeners.delete(listener),
-    fire: (...args: Parameters<T>) => {
-      for (const listener of [...listeners]) listener(...args);
-    },
-  };
-}
-
-async function fixture() {
-  const manager = new SessionManager({
-    agentWindow: {
-      create: vi.fn(async () => 100),
-      remove: vi.fn(async () => {}),
-      ensureActiveTab: vi.fn(async () => 4),
-    },
-  });
-  await manager.start("test");
-  type Listener = Parameters<BrowserNavigationApi["onCommitted"]["addListener"]>[0];
-  const events = {
-    onCommitted: event<Listener>(),
-    onDOMContentLoaded: event<Listener>(),
-    onCompleted: event<Listener>(),
-    onErrorOccurred: event<Listener>(),
-  };
-  const cdpEvents = event<Parameters<NonNullable<CdpRunner["onEvent"]>>[0]>();
-  const state = { blocked: true, destinationBlocked: false, emitIdle: true };
-  const send = vi.fn(async (_tabId: number, method: string) => {
-    if (state.blocked) throw new Error(denied);
-    if (method === "Page.getFrameTree") {
-      return { frameTree: { frame: { id: "main", loaderId: "new-loader" } } };
-    }
-    if (method === "Runtime.evaluate") return { result: { value: "complete" } };
-    if (method === "Page.setLifecycleEventsEnabled" && state.emitIdle) {
-      cdpEvents.fire({ tabId: 4 }, "Page.lifecycleEvent", {
-        frameId: "main",
-        loaderId: "new-loader",
-        name: "networkIdle",
-      });
-    }
-    if (method === "Page.navigate") return { frameId: "main", loaderId: "new-loader" };
-    return {};
-  });
-  const cdp: CdpRunner = {
-    send: send as CdpRunner["send"],
-    onEvent: (listener) => {
-      cdpEvents.addListener(listener);
-      return { dispose: () => cdpEvents.removeListener(listener) };
-    },
-  };
-  const main = { tabId: 4, frameId: 0, documentId: "new-document" };
-  const navigate = async () => {
-    state.blocked = state.destinationBlocked;
-    events.onCommitted.fire(main);
-    events.onDOMContentLoaded.fire(main);
-    events.onCompleted.fire(main);
-  };
-  const browserNavigation = { ...events, update: vi.fn(navigate), reload: vi.fn(navigate) };
-  const tab = { id: 4, windowId: 100, active: true, url } as chrome.tabs.Tab;
-  const tabsApi = { get: vi.fn(async () => tab), query: vi.fn(async () => [tab]) };
-  const deps = { cdp, tabsApi, browserNavigation, defaultTimeoutMs: 1000 };
-  const expectCleanedUp = () => {
-    for (const entry of Object.values(events)) expect(entry.listeners.size).toBe(0);
-    expect(cdpEvents.listeners.size).toBe(0);
-  };
-  return { manager, deps, state, send, events, main, cdpEvents, expectCleanedUp };
-}
-
-describe("navigation after Chrome denies extension-frame access", () => {
-  it.each([
-    "load",
-    "commit",
-    "domcontentloaded",
-    "networkidle",
-  ] as const)("recovers the same tab with the requested %s checkpoint", async (waitUntil) => {
-    const f = await fixture();
-    const result = await handleNavigate(
-      f.manager,
-      {
-        session_id: "test",
-        url,
-        wait_until: waitUntil,
-      },
-      f.deps,
-    );
-
-    expect(result).toMatchObject({ tab_id: 4, final_url: url, reached: waitUntil });
-    expect(f.deps.browserNavigation.update).toHaveBeenCalledExactlyOnceWith(4, { url });
-    expect(f.send.mock.calls.some(([, method]) => method === "Page.navigate")).toBe(false);
-    expect(f.send).toHaveBeenCalledWith(4, "Page.enable", {});
-    f.expectCleanedUp();
-  });
-
-  it("recovers reload through tabs.reload and preserves hard reload", async () => {
-    const f = await fixture();
-    const result = await handleReload(f.manager, { session_id: "test", hard: true }, f.deps);
-    expect(result).toMatchObject({ tab_id: 4, reached: "load" });
-    expect(f.deps.browserNavigation.reload).toHaveBeenCalledExactlyOnceWith(4, {
-      bypassCache: true,
-    });
-    expect(f.deps.browserNavigation.update).not.toHaveBeenCalled();
-    f.expectCleanedUp();
-  });
-
-  it("reports the restriction when reload leaves the conflicting frame present", async () => {
-    const f = await fixture();
-    f.state.destinationBlocked = true;
-    const result = await handleReload(f.manager, { session_id: "test" }, f.deps);
-    expect(result).toMatchObject({
-      code: "cdp_failed",
-      data: { reason: "cdp_extension_access_denied" },
-    });
-    expect(f.deps.browserNavigation.reload).toHaveBeenCalledOnce();
-    f.expectCleanedUp();
-  });
-
-  it("does not fall back for unrelated debugger failures", async () => {
-    const f = await fixture();
-    f.send.mockRejectedValue(new Error("Another debugger is already attached"));
-    const result = await handleNavigate(f.manager, { session_id: "test", url }, f.deps);
-    expect(result).toMatchObject({
-      code: "cdp_failed",
-      message: "Another debugger is already attached",
-    });
-    expect(f.deps.browserNavigation.update).not.toHaveBeenCalled();
-    f.expectCleanedUp();
-  });
-
-  it("does not replay navigation when a dispatched CDP command fails", async () => {
-    const f = await fixture();
-    f.state.blocked = false;
-    const send = f.send.getMockImplementation()!;
-    f.send.mockImplementation(async (tabId, method) => {
-      if (method === "Page.navigate") throw new Error(denied);
-      return send(tabId, method);
-    });
-    const result = await handleNavigate(f.manager, { session_id: "test", url }, f.deps);
-    expect(result).toMatchObject({
-      code: "cdp_failed",
-      data: { reason: "cdp_extension_access_denied" },
-    });
-    expect(f.deps.browserNavigation.update).not.toHaveBeenCalled();
-    f.expectCleanedUp();
-  });
-
-  it("keeps the Agent Window guard in front of the recovery", async () => {
-    const f = await fixture();
-    f.deps.tabsApi.get.mockResolvedValue({
-      id: 4,
-      windowId: 200,
-      active: true,
-      url,
-    } as chrome.tabs.Tab);
-    const result = await handleNavigate(f.manager, { session_id: "test", tab_id: 4, url }, f.deps);
-    expect(result).toMatchObject({ code: "permission_denied" });
-    expect(f.send).not.toHaveBeenCalled();
-    expect(f.deps.browserNavigation.update).not.toHaveBeenCalled();
-  });
-
-  it("ignores old-document, subframe, and other-tab completion events", async () => {
-    const f = await fixture();
-    f.deps.browserNavigation.update.mockImplementation(async () => {});
-    let settled = false;
-    const work = handleNavigate(f.manager, { session_id: "test", url }, f.deps).then((result) => {
-      settled = true;
-      return result;
-    });
-    await vi.waitFor(() => expect(f.deps.browserNavigation.update).toHaveBeenCalledOnce());
-    f.events.onCompleted.fire(f.main);
-    f.events.onCommitted.fire({ ...f.main, tabId: 9 });
-    f.events.onCommitted.fire({ ...f.main, frameId: 1 });
-    f.events.onCompleted.fire(f.main);
-    await Promise.resolve();
-    expect(settled).toBe(false);
-    f.events.onCommitted.fire(f.main);
-    f.events.onCompleted.fire({ ...f.main, documentId: "old-document" });
-    await Promise.resolve();
-    expect(settled).toBe(false);
-    f.state.blocked = false;
-    f.events.onCompleted.fire(f.main);
-    expect(await work).toMatchObject({ reached: "load" });
-    f.expectCleanedUp();
-  });
-
-  it("does not infer network idle from a browser load event", async () => {
-    const f = await fixture();
-    f.state.emitIdle = false;
-    const result = await handleNavigate(
-      f.manager,
-      {
-        session_id: "test",
-        url,
-        wait_until: "networkidle",
-        timeout_ms: 10,
-      },
-      f.deps,
-    );
-    expect(result).toMatchObject({ reached: "timeout" });
-    f.expectCleanedUp();
-  });
-
-  it("cleans up browser listeners on timeout", async () => {
-    const f = await fixture();
-    f.deps.browserNavigation.update.mockImplementation(async () => {});
-    const result = await handleNavigate(
-      f.manager,
-      { session_id: "test", url, timeout_ms: 10 },
-      f.deps,
-    );
-    expect(result).toMatchObject({ reached: "timeout" });
-    f.expectCleanedUp();
-  });
-
-  it("does not navigate an already-cancelled request", async () => {
-    const f = await fixture();
-    const result = await handleNavigate(
-      f.manager,
-      { session_id: "test", url },
-      {
-        ...f.deps,
-        signal: AbortSignal.abort(),
-      },
-    );
-    expect(result).toMatchObject({ code: "cancelled" });
-    expect(f.deps.browserNavigation.update).not.toHaveBeenCalled();
-    f.expectCleanedUp();
-  });
-
-  it("cleans up browser listeners when cancelled after dispatch", async () => {
-    const f = await fixture();
-    const ac = new AbortController();
-    f.deps.browserNavigation.update.mockImplementation(async () => {
-      ac.abort();
-    });
-    const result = await handleNavigate(
-      f.manager,
-      { session_id: "test", url },
-      { ...f.deps, signal: ac.signal },
-    );
-    expect(result).toMatchObject({ code: "cancelled" });
-    expect(f.deps.browserNavigation.update).toHaveBeenCalledOnce();
-    f.expectCleanedUp();
-  });
-
-  it("reports browser navigation errors and releases listeners", async () => {
-    const f = await fixture();
-    f.deps.browserNavigation.update.mockImplementation(async () => {
-      f.events.onErrorOccurred.fire({ ...f.main, error: "net::ERR_NAME_NOT_RESOLVED" });
-    });
-    const result = await handleNavigate(f.manager, { session_id: "test", url }, f.deps);
-    expect(result).toMatchObject({ code: "cdp_failed", message: "net::ERR_NAME_NOT_RESOLVED" });
-    f.expectCleanedUp();
-  });
-});

apps/extension/src/tools/__tests__/navigation.test.ts

@@ -21,7 +21,7 @@ function fakeAgentWindow(ids: number[]) {
       return id;
     }),
     remove: vi.fn(async () => {}),
-    ensureActiveTab: vi.fn(async () => 1),
+    ensureActiveTab: vi.fn(async () => {}),
   };
 }
 

apps/extension/src/tools/__tests__/network.test.ts

@@ -12,7 +12,7 @@ function fakeAgentWindow(ids: number[]) {
       return id;
     }),
     remove: vi.fn(async () => {}),
-    ensureActiveTab: vi.fn(async () => 1),
+    ensureActiveTab: vi.fn(async () => {}),
   };
 }
 

apps/extension/src/tools/__tests__/observation.test.ts

@@ -28,7 +28,7 @@ function fakeAgentWindow(ids: number[]) {
       return id;
     }),
     remove: vi.fn(async () => {}),
-    ensureActiveTab: vi.fn(async () => 1),
+    ensureActiveTab: vi.fn(async () => {}),
   };
 }
 
@@ -67,6 +67,9 @@ function makeFakeCdp(handlers: Record<string, (params?: object) => unknown>) {
     if (!handler && method === "Page.getLayoutMetrics") {
       return { cssLayoutViewport: { clientWidth: 1280, clientHeight: 720 } };
     }
+    if (!handler && method === "Page.getFrameTree") {
+      return { frameTree: { frame: { id: "main", loaderId: "loader-1", url: "https://test/" } } };
+    }
     if (!handler) throw new Error(`unexpected CDP call ${method}`);
     return handler(params);
   });
@@ -300,6 +303,82 @@ describe("handleScreenshot", () => {
     expect(clip).toMatchObject({ x: 10, y: 20, width: 100, height: 40 });
   });
 
+  it("captures a bounded visible crop for a visual surface without scrolling it", async () => {
+    const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
+    const ctx = await sm.start("aa11");
+    ctx.refStore.set("e5", 999, {
+      tabId: 7,
+      kind: "surface",
+      capabilities: ["screenshot"],
+    });
+    const { cdp, sent } = makeFakeCdp({
+      "Page.getLayoutMetrics": () => ({
+        cssLayoutViewport: { clientWidth: 4096, clientHeight: 4096 },
+      }),
+      "DOM.getContentQuads": () => ({ quads: [[0, 0, 4096, 0, 4096, 4096, 0, 4096]] }),
+      "Page.captureScreenshot": () => ({ data: TINY_PNG }),
+    });
+
+    const res = await handleScreenshot(
+      sm,
+      { session_id: "aa11", ref: "@e5", tab_id: 7 },
+      makeScreenshotDeps({
+        cdp,
+        get: vi.fn(async () => ({ id: 7, windowId: 100, active: false }) as chrome.tabs.Tab),
+        query: vi.fn(),
+        captureVisibleTab: vi.fn(),
+      }),
+    );
+
+    if ("code" in res) throw new Error(`unexpected error: ${JSON.stringify(res)}`);
+    expect(sent.some((call) => call.method === "DOM.scrollIntoViewIfNeeded")).toBe(false);
+    const clip = sent.find((call) => call.method === "Page.captureScreenshot")?.params as {
+      clip?: { width: number; height: number; scale: number };
+    };
+    expect(clip.clip).toMatchObject({ width: 4096, height: 4096 });
+    expect(clip.clip?.scale).toBeCloseTo(Math.sqrt(4_000_000 / (4096 * 4096)));
+    expect(res.capture).toMatchObject({
+      surface_ref: "@e5",
+      coordinate_space: "capture-image-pixel",
+    });
+    expect(ctx.surfaceCaptures.size()).toBe(1);
+  });
+
+  it("crops a visual surface screenshot to its observation-time visible region", async () => {
+    const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
+    const ctx = await sm.start("aa11");
+    ctx.refStore.set("e5", 999, {
+      tabId: 7,
+      kind: "surface",
+      visibleRect: { x: 25, y: 30, w: 50, h: 40 },
+    });
+    const { cdp, sent } = makeFakeCdp({
+      "Page.getLayoutMetrics": () => ({
+        cssLayoutViewport: { clientWidth: 800, clientHeight: 600 },
+      }),
+      "DOM.getContentQuads": () => ({ quads: [[0, 0, 200, 0, 200, 100, 0, 100]] }),
+      "Page.captureScreenshot": () => ({ data: TINY_PNG }),
+    });
+
+    const res = await handleScreenshot(
+      sm,
+      { session_id: "aa11", ref: "@e5", tab_id: 7 },
+      makeScreenshotDeps({
+        cdp,
+        get: vi.fn(async () => ({ id: 7, windowId: 100, active: false }) as chrome.tabs.Tab),
+        query: vi.fn(),
+        captureVisibleTab: vi.fn(),
+      }),
+    );
+
+    if ("code" in res) throw new Error(`unexpected error: ${JSON.stringify(res)}`);
+    const clip = sent.find((call) => call.method === "Page.captureScreenshot")?.params as {
+      clip?: { x: number; y: number; width: number; height: number };
+    };
+    expect(clip.clip).toMatchObject({ x: 25, y: 30, width: 50, height: 40 });
+    expect(res.capture?.id).toMatch(/^sc_/);
+  });
+
   it("returns not_found for unknown ref", async () => {
     const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
     await sm.start("aa11");
@@ -1904,41 +1983,36 @@ describe("buildVomScene", () => {
         backendDOMNodeId: 20,
       },
     ];
-    const scene = buildVomScene(
-      axNodes,
-      {
-        viewport: { width: 1000, height: 800 },
-        iframeNodes: new Map(),
-        excludedBackendNodeIds: new Set(),
-        nodes: [
-          {
-            backendNodeId: 10,
-            parentBackendNodeId: null,
-            tag: "body",
-            attrs: {},
-            rect: { x: 0, y: 0, w: 1000, h: 800 },
-            paintOrder: 0,
-            position: "static",
-            pointerEvents: "auto",
-          },
-          {
-            backendNodeId: 20,
-            parentBackendNodeId: 10,
-            tag: "button",
-            attrs: {},
-            rect: { x: 20, y: 20, w: 120, h: 40 },
-            paintOrder: 1,
-            position: "static",
-            pointerEvents: "auto",
-          },
-        ],
-      },
-      {
-        surfaceProbes: [
-          { triggerBackendNodeId: 20, triggerAction: "hover", subItems: ["Shoes", "Bags"] },
-        ],
-      },
-    );
+    const scene = buildVomScene(axNodes, {
+      viewport: { width: 1000, height: 800 },
+      iframeNodes: new Map(),
+      excludedBackendNodeIds: new Set(),
+      surfaceProbes: [
+        { triggerBackendNodeId: 20, triggerAction: "hover", subItems: ["Shoes", "Bags"] },
+      ],
+      nodes: [
+        {
+          backendNodeId: 10,
+          parentBackendNodeId: null,
+          tag: "body",
+          attrs: {},
+          rect: { x: 0, y: 0, w: 1000, h: 800 },
+          paintOrder: 0,
+          position: "static",
+          pointerEvents: "auto",
+        },
+        {
+          backendNodeId: 20,
+          parentBackendNodeId: 10,
+          tag: "button",
+          attrs: {},
+          rect: { x: 20, y: 20, w: 120, h: 40 },
+          paintOrder: 1,
+          position: "static",
+          pointerEvents: "auto",
+        },
+      ],
+    });
 
     expect(scene.surfaces).toEqual([
       { triggerId: 20, triggerAction: "hover", subItems: ["Shoes", "Bags"] },
@@ -1962,51 +2036,46 @@ describe("buildVomScene", () => {
         backendDOMNodeId: 21,
       },
     ];
-    const scene = buildVomScene(
-      axNodes,
-      {
-        viewport: { width: 1000, height: 800 },
-        iframeNodes: new Map(),
-        excludedBackendNodeIds: new Set(),
-        nodes: [
-          {
-            backendNodeId: 10,
-            parentBackendNodeId: null,
-            tag: "body",
-            attrs: {},
-            rect: { x: 0, y: 0, w: 1000, h: 800 },
-            paintOrder: 0,
-            position: "static",
-            pointerEvents: "auto",
-          },
-          {
-            backendNodeId: 20,
-            parentBackendNodeId: 10,
-            tag: "div",
-            attrs: { class: "tg-avatar" },
-            rect: { x: 900, y: 10, w: 30, h: 30 },
-            paintOrder: 1,
-            position: "static",
-            pointerEvents: "auto",
-          },
-          {
-            backendNodeId: 21,
-            parentBackendNodeId: 20,
-            tag: "div",
-            attrs: { class: "tg-avatar__inner" },
-            rect: { x: 902, y: 12, w: 26, h: 26 },
-            paintOrder: 2,
-            position: "static",
-            pointerEvents: "auto",
-          },
-        ],
-      },
-      {
-        surfaceProbes: [
-          { triggerBackendNodeId: 20, triggerAction: "hover", subItems: ["My profile"] },
-        ],
-      },
-    );
+    const scene = buildVomScene(axNodes, {
+      viewport: { width: 1000, height: 800 },
+      iframeNodes: new Map(),
+      excludedBackendNodeIds: new Set(),
+      surfaceProbes: [
+        { triggerBackendNodeId: 20, triggerAction: "hover", subItems: ["My profile"] },
+      ],
+      nodes: [
+        {
+          backendNodeId: 10,
+          parentBackendNodeId: null,
+          tag: "body",
+          attrs: {},
+          rect: { x: 0, y: 0, w: 1000, h: 800 },
+          paintOrder: 0,
+          position: "static",
+          pointerEvents: "auto",
+        },
+        {
+          backendNodeId: 20,
+          parentBackendNodeId: 10,
+          tag: "div",
+          attrs: { class: "tg-avatar" },
+          rect: { x: 900, y: 10, w: 30, h: 30 },
+          paintOrder: 1,
+          position: "static",
+          pointerEvents: "auto",
+        },
+        {
+          backendNodeId: 21,
+          parentBackendNodeId: 20,
+          tag: "div",
+          attrs: { class: "tg-avatar__inner" },
+          rect: { x: 902, y: 12, w: 26, h: 26 },
+          paintOrder: 2,
+          position: "static",
+          pointerEvents: "auto",
+        },
+      ],
+    });
 
     expect(renderVom(scene).text).toContain('@e1 button "image" [hover first: My profile]');
   });
@@ -2027,53 +2096,48 @@ describe("buildVomScene", () => {
         backendDOMNodeId: 21,
       },
     ];
-    const scene = buildVomScene(
-      axNodes,
-      {
-        viewport: { width: 1000, height: 800 },
-        iframeNodes: new Map(),
-        excludedBackendNodeIds: new Set(),
-        nodes: [
-          {
-            backendNodeId: 10,
-            parentBackendNodeId: null,
-            tag: "body",
-            attrs: {},
-            rect: { x: 0, y: 0, w: 1000, h: 800 },
-            paintOrder: 0,
-            position: "static",
-            pointerEvents: "auto",
-          },
-          {
-            backendNodeId: 20,
-            parentBackendNodeId: 10,
-            tag: "div",
-            attrs: { class: "tg-avatar" },
-            rect: { x: 900, y: 10, w: 30, h: 30 },
-            paintOrder: 1,
-            position: "static",
-            pointerEvents: "auto",
-          },
-          {
-            backendNodeId: 21,
-            parentBackendNodeId: 20,
-            tag: "div",
-            attrs: { class: "tg-avatar__inner" },
-            rect: { x: 902, y: 12, w: 26, h: 26 },
-            paintOrder: 2,
-            position: "static",
-            pointerEvents: "auto",
-          },
-        ],
-      },
-      {
-        surfaceProbes: [
-          { triggerBackendNodeId: 20, triggerAction: "hover", subItems: ["My profile"] },
-          { triggerBackendNodeId: 20, triggerAction: "hover", subItems: ["Sign out"] },
-          { triggerBackendNodeId: 21, triggerAction: "hover", subItems: ["Settings"] },
-        ],
-      },
-    );
+    const scene = buildVomScene(axNodes, {
+      viewport: { width: 1000, height: 800 },
+      iframeNodes: new Map(),
+      excludedBackendNodeIds: new Set(),
+      surfaceProbes: [
+        { triggerBackendNodeId: 20, triggerAction: "hover", subItems: ["My profile"] },
+        { triggerBackendNodeId: 20, triggerAction: "hover", subItems: ["Sign out"] },
+        { triggerBackendNodeId: 21, triggerAction: "hover", subItems: ["Settings"] },
+      ],
+      nodes: [
+        {
+          backendNodeId: 10,
+          parentBackendNodeId: null,
+          tag: "body",
+          attrs: {},
+          rect: { x: 0, y: 0, w: 1000, h: 800 },
+          paintOrder: 0,
+          position: "static",
+          pointerEvents: "auto",
+        },
+        {
+          backendNodeId: 20,
+          parentBackendNodeId: 10,
+          tag: "div",
+          attrs: { class: "tg-avatar" },
+          rect: { x: 900, y: 10, w: 30, h: 30 },
+          paintOrder: 1,
+          position: "static",
+          pointerEvents: "auto",
+        },
+        {
+          backendNodeId: 21,
+          parentBackendNodeId: 20,
+          tag: "div",
+          attrs: { class: "tg-avatar__inner" },
+          rect: { x: 902, y: 12, w: 26, h: 26 },
+          paintOrder: 2,
+          position: "static",
+          pointerEvents: "auto",
+        },
+      ],
+    });
 
     expect(scene.surfaces).toEqual([
       { triggerId: 21, triggerAction: "hover", subItems: ["My profile"] },
@@ -2096,59 +2160,54 @@ describe("buildVomScene", () => {
         backendDOMNodeId: 21,
       },
     ];
-    const scene = buildVomScene(
-      axNodes,
-      {
-        viewport: { width: 1000, height: 800 },
-        iframeNodes: new Map(),
-        excludedBackendNodeIds: new Set(),
-        nodes: [
-          {
-            backendNodeId: 10,
-            parentBackendNodeId: null,
-            tag: "body",
-            attrs: {},
-            rect: { x: 0, y: 0, w: 1000, h: 800 },
-            paintOrder: 0,
-            position: "static",
-            pointerEvents: "auto",
-            cursor: "auto",
-          },
-          {
-            backendNodeId: 20,
-            parentBackendNodeId: 10,
-            tag: "div",
-            attrs: { class: "tg-avatar" },
-            rect: { x: 900, y: 10, w: 30, h: 30 },
-            paintOrder: 1,
-            position: "static",
-            pointerEvents: "auto",
-            cursor: "pointer",
-          },
-          {
-            backendNodeId: 21,
-            parentBackendNodeId: 20,
-            tag: "div",
-            attrs: { class: "tg-avatar__inner" },
-            rect: { x: 902, y: 12, w: 26, h: 26 },
-            paintOrder: 2,
-            position: "static",
-            pointerEvents: "auto",
-            cursor: "pointer",
-          },
-        ],
-      },
-      {
-        surfaceProbes: [
-          {
-            triggerBackendNodeId: 20,
-            triggerPoint: { x: 915, y: 25 },
-            triggerAction: "hover",
-            subItems: ["My profile", "Sign out"],
-          },
-        ],
-      },
-    );
+    const scene = buildVomScene(axNodes, {
+      viewport: { width: 1000, height: 800 },
+      iframeNodes: new Map(),
+      excludedBackendNodeIds: new Set(),
+      surfaceProbes: [
+        {
+          triggerBackendNodeId: 20,
+          triggerPoint: { x: 915, y: 25 },
+          triggerAction: "hover",
+          subItems: ["My profile", "Sign out"],
+        },
+      ],
+      nodes: [
+        {
+          backendNodeId: 10,
+          parentBackendNodeId: null,
+          tag: "body",
+          attrs: {},
+          rect: { x: 0, y: 0, w: 1000, h: 800 },
+          paintOrder: 0,
+          position: "static",
+          pointerEvents: "auto",
+          cursor: "auto",
+        },
+        {
+          backendNodeId: 20,
+          parentBackendNodeId: 10,
+          tag: "div",
+          attrs: { class: "tg-avatar" },
+          rect: { x: 900, y: 10, w: 30, h: 30 },
+          paintOrder: 1,
+          position: "static",
+          pointerEvents: "auto",
+          cursor: "pointer",
+        },
+        {
+          backendNodeId: 21,
+          parentBackendNodeId: 20,
+          tag: "div",
+          attrs: { class: "tg-avatar__inner" },
+          rect: { x: 902, y: 12, w: 26, h: 26 },
+          paintOrder: 2,
+          position: "static",
+          pointerEvents: "auto",
+          cursor: "pointer",
+        },
+      ],
+    });
 
     expect(scene.surfaces).toEqual([
       { triggerId: 20, triggerAction: "hover", subItems: ["My profile", "Sign out"] },
@@ -2174,46 +2233,41 @@ describe("buildVomScene", () => {
         backendDOMNodeId: 42,
       },
     ];
-    const scene = buildVomScene(
-      axNodes,
-      {
-        viewport: { width: 1000, height: 800 },
-        iframeNodes: new Map(),
-        excludedBackendNodeIds: new Set(),
-        nodes: [
-          {
-            backendNodeId: 10,
-            parentBackendNodeId: null,
-            tag: "body",
-            attrs: {},
-            rect: { x: 0, y: 0, w: 1000, h: 800 },
-            paintOrder: 0,
-            position: "static",
-            pointerEvents: "auto",
-          },
-          {
-            backendNodeId: 42,
-            parentBackendNodeId: 10,
-            tag: "a",
-            attrs: {},
-            rect: { x: 860, y: 120, w: 40, h: 30 },
-            paintOrder: 1,
-            position: "static",
-            pointerEvents: "auto",
-          },
-        ],
-      },
-      {
-        surfaceProbes: [
-          {
-            triggerBackendNodeId: 999,
-            triggerPoint: { x: 900, y: 20 },
-            triggerAction: "hover",
-            subItems: ["My profile"],
-          },
-        ],
-      },
-    );
+    const scene = buildVomScene(axNodes, {
+      viewport: { width: 1000, height: 800 },
+      iframeNodes: new Map(),
+      excludedBackendNodeIds: new Set(),
+      surfaceProbes: [
+        {
+          triggerBackendNodeId: 999,
+          triggerPoint: { x: 900, y: 20 },
+          triggerAction: "hover",
+          subItems: ["My profile"],
+        },
+      ],
+      nodes: [
+        {
+          backendNodeId: 10,
+          parentBackendNodeId: null,
+          tag: "body",
+          attrs: {},
+          rect: { x: 0, y: 0, w: 1000, h: 800 },
+          paintOrder: 0,
+          position: "static",
+          pointerEvents: "auto",
+        },
+        {
+          backendNodeId: 42,
+          parentBackendNodeId: 10,
+          tag: "a",
+          attrs: {},
+          rect: { x: 860, y: 120, w: 40, h: 30 },
+          paintOrder: 1,
+          position: "static",
+          pointerEvents: "auto",
+        },
+      ],
+    });
 
     expect(scene.surfaces).toBeUndefined();
     expect(renderVom(scene).text).not.toContain("[hover first:");
@@ -2631,106 +2685,7 @@ describe("handleSnapshot", () => {
     );
   });
 
-  it("hovers the page only after the accessibility tree has been captured", async () => {
-    // Hovering can open menus and reflow the page. Probing between the DOM and
-    // AX captures would leave the two halves of one observation describing the
-    // page on either side of that change.
-    const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
-    await sm.start("aa11");
-    const root: CdpAxNode = {
-      nodeId: "1",
-      role: { type: "role", value: "RootWebArea" },
-      name: { type: "computedString", value: "Example" },
-      backendDOMNodeId: 100,
-      childIds: ["2"],
-    };
-    const button: CdpAxNode = {
-      nodeId: "2",
-      parentId: "1",
-      role: { type: "role", value: "button" },
-      name: { type: "computedString", value: "Products" },
-      backendDOMNodeId: 200,
-    };
-    const strings = ["body", "button", "position", "static", "pointer-events", "auto", "cursor"];
-    const i = (s: string) => strings.indexOf(s);
-    const methodOrder: string[] = [];
-    const send = vi.fn(async (_tabId: number, method: string, params?: object) => {
-      methodOrder.push(method);
-      if (method === "Accessibility.enable") return {};
-      if (method === "Accessibility.getFullAXTree") return { nodes: [root, button] };
-      if (method === "Page.getLayoutMetrics") {
-        return { cssLayoutViewport: { clientWidth: 1000, clientHeight: 800, pageX: 0, pageY: 0 } };
-      }
-      if (method === "DOMSnapshot.enable") return {};
-      if (method === "DOMSnapshot.captureSnapshot") {
-        return {
-          strings,
-          documents: [
-            {
-              nodes: {
-                parentIndex: [-1, 0],
-                nodeName: [i("body"), i("button")],
-                backendNodeId: [100, 200],
-                attributes: [[], []],
-              },
-              layout: {
-                nodeIndex: [0, 1],
-                styles: [
-                  [i("static"), i("auto"), i("auto")],
-                  [i("static"), i("auto"), i("auto")],
-                ],
-                bounds: [
-                  [0, 0, 1000, 800],
-                  [20, 20, 120, 40],
-                ],
-                paintOrders: [0, 1],
-              },
-            },
-          ],
-        };
-      }
-      if (method === "Input.dispatchMouseEvent") return {};
-      if (method === "Runtime.evaluate") {
-        const expression = (params as { expression?: string } | undefined)?.expression ?? "";
-        // Report the button as a `:hover` rule target so a real hover fires.
-        if (expression.includes("document.styleSheets")) {
-          return { result: { value: [{ x: 80, y: 40 }] } };
-        }
-        return { result: { value: [] } };
-      }
-      throw new Error(`unexpected CDP method ${method}`);
-    });
-
-    const res = await handleObserve(
-      sm,
-      { session_id: "aa11", probe_hover: true },
-      {
-        cdp: {
-          send: send as unknown as <T = unknown>(
-            tabId: number,
-            method: string,
-            params?: object,
-          ) => Promise<T>,
-          trackSessionTab: vi.fn(),
-        },
-        tabsApi: {
-          get: vi.fn(
-            async (tabId: number) =>
-              ({ id: tabId, windowId: 100, active: true }) as chrome.tabs.Tab,
-          ),
-          query: vi.fn(async () => [{ id: 4, windowId: 100, active: true } as chrome.tabs.Tab]),
-        },
-      },
-    );
-
-    if ("code" in res) throw new Error(`unexpected error: ${JSON.stringify(res)}`);
-    const firstHover = methodOrder.indexOf("Input.dispatchMouseEvent");
-    expect(firstHover).toBeGreaterThan(-1);
-    expect(methodOrder.lastIndexOf("DOMSnapshot.captureSnapshot")).toBeLessThan(firstHover);
-    expect(methodOrder.lastIndexOf("Accessibility.getFullAXTree")).toBeLessThan(firstHover);
-  });
-
-  it("runs conditional surface probing only when observe opts in", async () => {
+  it("enables conditional surface probing for semantic observe", async () => {
     const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
     await sm.start("aa11");
     const root: CdpAxNode = {
@@ -2786,6 +2741,83 @@ describe("handleSnapshot", () => {
       if (method === "Runtime.evaluate") return { result: { value: [] } };
       throw new Error(`unexpected CDP method ${method}`);
     });
+    const res = await handleObserve(
+      sm,
+      { session_id: "aa11", debug_surfaces: true },
+      {
+        cdp: {
+          send: send as unknown as <T = unknown>(
+            tabId: number,
+            method: string,
+            params?: object,
+          ) => Promise<T>,
+          trackSessionTab: vi.fn(),
+        },
+        tabsApi: {
+          get: vi.fn(
+            async (tabId: number) =>
+              ({ id: tabId, windowId: 100, active: true }) as chrome.tabs.Tab,
+          ),
+          query: vi.fn(async () => [{ id: 4, windowId: 100, active: true } as chrome.tabs.Tab]),
+        },
+      },
+    );
+
+    if ("code" in res) throw new Error(`unexpected error: ${JSON.stringify(res)}`);
+    expect(res.debug).toEqual({ surface_probes: [] });
+    expect(send).toHaveBeenCalledWith(
+      4,
+      "Runtime.evaluate",
+      expect.objectContaining({ returnByValue: true }),
+    );
+  });
+
+  it("discovers rendered canvases only in observe and stores a screenshot-only ref", async () => {
+    const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
+    const ctx = await sm.start("aa11");
+    const root: CdpAxNode = {
+      nodeId: "1",
+      role: { type: "role", value: "RootWebArea" },
+      name: { type: "computedString", value: "Canvas app" },
+      backendDOMNodeId: 100,
+    };
+    const strings = ["body", "canvas", "static", "auto"];
+    const i = (value: string) => strings.indexOf(value);
+    const send = vi.fn(async (_tabId: number, method: string) => {
+      if (method === "Accessibility.enable" || method === "DOMSnapshot.enable") return {};
+      if (method === "Accessibility.getFullAXTree") return { nodes: [root] };
+      if (method === "Page.getLayoutMetrics") {
+        return { cssLayoutViewport: { clientWidth: 1000, clientHeight: 800, pageX: 0, pageY: 0 } };
+      }
+      if (method === "DOMSnapshot.captureSnapshot") {
+        return {
+          strings,
+          documents: [
+            {
+              nodes: {
+                parentIndex: [-1, 0],
+                nodeName: [i("body"), i("canvas")],
+                backendNodeId: [100, 200],
+                attributes: [[], []],
+              },
+              layout: {
+                nodeIndex: [0, 1],
+                styles: [
+                  [i("static"), i("auto"), i("auto")],
+                  [i("static"), i("auto"), i("auto")],
+                ],
+                bounds: [
+                  [0, 0, 1000, 800],
+                  [20, 30, 800, 500],
+                ],
+                paintOrders: [0, 1],
+              },
+            },
+          ],
+        };
+      }
+      throw new Error(`unexpected CDP method ${method}`);
+    });
     const deps = {
       cdp: {
         send: send as unknown as <T = unknown>(
@@ -2801,36 +2833,22 @@ describe("handleSnapshot", () => {
         ),
         query: vi.fn(async () => [{ id: 4, windowId: 100, active: true } as chrome.tabs.Tab]),
       },
+      conditionalSurfaceProbe: false,
     };
 
-    const withoutProbe = await handleObserve(
-      sm,
-      { session_id: "aa11", debug_surfaces: true },
-      deps,
-    );
-    if ("code" in withoutProbe) {
-      throw new Error(`unexpected error: ${JSON.stringify(withoutProbe)}`);
-    }
-    expect(withoutProbe.hover_probe).toBeUndefined();
-    expect(send).not.toHaveBeenCalledWith(
-      4,
-      "Runtime.evaluate",
-      expect.objectContaining({ returnByValue: true }),
-    );
+    const observed = await handleObserve(sm, { session_id: "aa11" }, deps);
+    if ("code" in observed) throw new Error(`unexpected error: ${JSON.stringify(observed)}`);
+    expect(observed.text).toContain('@e1 surface "canvas visual surface"');
+    expect(ctx.refStore.resolveEntry("e1")).toMatchObject({
+      backendNodeId: 200,
+      kind: "surface",
+      capabilities: ["screenshot"],
+    });
 
-    const withProbe = await handleObserve(
-      sm,
-      { session_id: "aa11", debug_surfaces: true, probe_hover: true },
-      deps,
-    );
-    if ("code" in withProbe) throw new Error(`unexpected error: ${JSON.stringify(withProbe)}`);
-    expect(withProbe.debug).toEqual({ surface_probes: [] });
-    expect(withProbe.hover_probe).toEqual({ performed: true, revealed_content: false });
-    expect(send).toHaveBeenCalledWith(
-      4,
-      "Runtime.evaluate",
-      expect.objectContaining({ returnByValue: true }),
-    );
+    const snapshot = await handleSnapshot(sm, { session_id: "aa11" }, deps);
+    if ("code" in snapshot) throw new Error(`unexpected error: ${JSON.stringify(snapshot)}`);
+    expect(snapshot.text).not.toContain("visual surface");
+    expect(snapshot.ref_count).toBe(0);
   });
 
   it("allows passive snapshots of explicit user-window tabs", async () => {

apps/extension/src/tools/__tests__/session.test.ts

@@ -1,26 +1,8 @@
 import { describe, expect, it, vi } from "vitest";
 import { SessionManager } from "@/session-manager/manager";
 import { handleSessionStop } from "../session";
-import type { ChromeTabsApi } from "../shared";
 import { type AgentOverlayResetApi, type ChromeWindowsApi, type TabMutationApi } from "../tabs";
 
-/** Build a read-only query api over a FakeState. */
-function makeQuery(state: FakeState): ChromeTabsApi {
-  return {
-    get: vi.fn(async (id) => {
-      const t = state.tabs.get(id);
-      if (!t) throw new Error(`tab ${id} not found`);
-      return t;
-    }),
-    query: vi.fn(async (q: chrome.tabs.QueryInfo) => {
-      const w = q.windowId;
-      return Array.from(state.tabs.values()).filter(
-        (t) => typeof w !== "number" || t.windowId === w,
-      );
-    }),
-  };
-}
-
 function fakeAgentWindow(ids: number[]) {
   let i = 0;
   const create = vi.fn(async () => {
@@ -29,7 +11,7 @@ function fakeAgentWindow(ids: number[]) {
     return id;
   });
   const remove = vi.fn(async () => {});
-  const ensureActiveTab = vi.fn(async () => 0);
+  const ensureActiveTab = vi.fn(async () => {});
   return { create, remove, ensureActiveTab };
 }
 
@@ -48,9 +30,7 @@ function makeApis(
 } {
   const tabs: TabMutationApi = {
     create: vi.fn(),
-    remove: vi.fn(async (id: number) => {
-      state.tabs.delete(id);
-    }),
+    remove: vi.fn(async () => {}),
     update: vi.fn(async (_id, _p) => undefined),
     get: vi.fn(async (id) => {
       const t = state.tabs.get(id);
@@ -257,15 +237,11 @@ describe("handleSessionStop with auto-return", () => {
       moves: [],
     };
     const { tabs, windows } = makeApis(state, { moveThrowsFor: new Set([1]) });
-    const cdp = {
-      detachSession: vi.fn(async () => {}),
-      releaseSessionTab: vi.fn(async () => {}),
-    };
 
     const res = await handleSessionStop(
       sm,
       { session_id: "aa11" },
-      { cdp, tabManagement: { tabs, windows } },
+      { tabManagement: { tabs, windows } },
     );
     if ("code" in res) throw new Error(`unexpected error: ${JSON.stringify(res)}`);
     expect(res.return_failures?.map((f) => f.tab_id)).toEqual([1]);
@@ -273,17 +249,26 @@ describe("handleSessionStop with auto-return", () => {
     expect(sm.has("aa11")).toBe(true);
     expect(ctx.borrowedTabs.has(1)).toBe(true);
     expect(ctx.borrowedTabs.has(2)).toBe(false);
-    expect(cdp.releaseSessionTab).toHaveBeenCalledExactlyOnceWith("aa11", 2);
-    expect(cdp.detachSession).not.toHaveBeenCalled();
     expect(aw.remove).not.toHaveBeenCalled();
   });
 
-  it("clears the RefStore before window teardown", async () => {
+  it("clears refs and Surface captures before window teardown", async () => {
     const aw = fakeAgentWindow([100]);
     const sm = new SessionManager({ agentWindow: aw });
     const ctx = await sm.start("aa11");
     // Insert a fake ref so we can verify clear() ran.
     ctx.refStore.set("e1", 123, { tabId: 7 });
+    ctx.surfaceCaptures.create({
+      sessionId: "aa11",
+      tabId: 7,
+      navigationIdentity: "navigation",
+      surface: { ref: "e1", backendNodeId: 123, observationGeneration: 0 },
+      topViewportRect: { x: 0, y: 0, w: 100, h: 50 },
+      imageWidth: 100,
+      imageHeight: 50,
+      viewportSignature: "viewport",
+      frameProjectionSignature: "frame",
+    });
     const state: FakeState = {
       tabs: new Map(),
       windowsClosed: new Set(),
@@ -292,276 +277,6 @@ describe("handleSessionStop with auto-return", () => {
     const { tabs, windows } = makeApis(state);
     await handleSessionStop(sm, { session_id: "aa11" }, { tabManagement: { tabs, windows } });
     expect(ctx.refStore.isEmpty()).toBe(true);
-  });
-
-  it.each([
-    "before teardown",
-    "during its return",
-    "after its return failed",
-    "before its turn in the return loop",
-  ])("finishes stopping when a borrowed tab closes %s", async (timing) => {
-    const aw = fakeAgentWindow([100]);
-    const sm = new SessionManager({ agentWindow: aw });
-    const ctx = await sm.start("aa11");
-    ctx.borrowedTabs.set(1, { tabId: 1, originalWindowId: 200, originalIndex: 0 });
-    ctx.borrowedTabs.set(2, { tabId: 2, originalWindowId: 200, originalIndex: 1 });
-    const state: FakeState = {
-      tabs: new Map([
-        [1, { id: 1, windowId: 100 } as chrome.tabs.Tab],
-        [2, { id: 2, windowId: 100 } as chrome.tabs.Tab],
-      ]),
-      windowsClosed: new Set(),
-      moves: [],
-    };
-    const { tabs, windows } = makeApis(state);
-    const cdp = { detachSession: vi.fn(async () => {}) };
-    const closeTab = (tabId: number) => {
-      state.tabs.delete(tabId);
-      sm.forgetClosedTab(tabId);
-    };
-    const baseMove = tabs.move;
-    tabs.move = vi.fn(async (tabId, props) => {
-      if (tabId === 1 && timing === "during its return") closeTab(1);
-      if (tabId === 1 && timing === "after its return failed") {
-        throw new Error("simulated move failure");
-      }
-      if (tabId === 2 && timing === "after its return failed") closeTab(1);
-      if (tabId === 1 && timing === "before its turn in the return loop") closeTab(2);
-      if (!state.tabs.has(tabId)) throw new Error(`tab ${tabId} not found`);
-      return baseMove(tabId, props);
-    });
-    if (timing === "before teardown") closeTab(1);
-
-    const res = await handleSessionStop(
-      sm,
-      { session_id: "aa11" },
-      {
-        cdp,
-        tabManagement: {
-          tabs,
-          windows,
-          agentOverlayReset: { resetAgentOverlays: vi.fn(async () => {}) },
-        },
-        tabsQuery: makeQuery(state),
-      },
-    );
-
-    if ("code" in res) throw new Error(`unexpected error: ${JSON.stringify(res)}`);
-    const closedTabId = timing === "before its turn in the return loop" ? 2 : 1;
-    expect(res.returned_tab_ids).toEqual([closedTabId === 1 ? 2 : 1]);
-    expect(res.return_failures).toBeUndefined();
-    expect(ctx.borrowedTabs.size).toBe(0);
-    expect(sm.has("aa11")).toBe(false);
-    expect(cdp.detachSession).toHaveBeenCalledWith("aa11");
-    expect(aw.remove).toHaveBeenCalledWith(100);
-    if (timing === "before teardown" || timing === "before its turn in the return loop") {
-      expect(tabs.move).not.toHaveBeenCalledWith(closedTabId, expect.anything());
-    }
-  });
-});
-
-describe("handleSessionStop window release (issue #57)", () => {
-  const agentWindowId = 100;
-
-  it("releases the window when user-created tabs remain", async () => {
-    const aw = fakeAgentWindow([agentWindowId]);
-    // ensureActiveTab returns the home tab id (10). We override the fake to
-    // return a known id.
-    aw.ensureActiveTab = vi.fn(async () => 10);
-    const sm = new SessionManager({ agentWindow: aw });
-    const ctx = await sm.start("aa11");
-    ctx.agentCreatedTabs.add(11);
-    ctx.agentCreatedTabs.add(12);
-    // user tab 99 (created via Chrome UI, NOT in agentCreatedTabs)
-    const state: FakeState = {
-      tabs: new Map([
-        [10, { id: 10, windowId: agentWindowId } as chrome.tabs.Tab],
-        [11, { id: 11, windowId: agentWindowId } as chrome.tabs.Tab],
-        [12, { id: 12, windowId: agentWindowId } as chrome.tabs.Tab],
-        [99, { id: 99, windowId: agentWindowId } as chrome.tabs.Tab],
-      ]),
-      windowsClosed: new Set(),
-      moves: [],
-    };
-    const { tabs, windows } = makeApis(state);
-    const query = makeQuery(state);
-
-    const res = await handleSessionStop(
-      sm,
-      { session_id: "aa11" },
-      { tabManagement: { tabs, windows }, tabsQuery: query },
-    );
-    if ("code" in res) throw new Error(`unexpected error: ${JSON.stringify(res)}`);
-    expect(res.window_released).toBe(true);
-    // agent tabs + home removed, user tab 99 kept
-    expect((tabs.remove as ReturnType<typeof vi.fn>).mock.calls.map((c) => c[0]).sort()).toEqual([
-      10, 11, 12,
-    ]);
-    expect(state.tabs.has(99)).toBe(true);
-    // window released (dropOnly), not closed
-    expect(aw.remove).not.toHaveBeenCalled();
-    expect(sm.has("aa11")).toBe(false);
-  });
-
-  it("closes the window (not release) when no user tabs remain", async () => {
-    const aw = fakeAgentWindow([agentWindowId]);
-    aw.ensureActiveTab = vi.fn(async () => 10);
-    const sm = new SessionManager({ agentWindow: aw });
-    const ctx = await sm.start("aa11");
-    ctx.agentCreatedTabs.add(11);
-    const state: FakeState = {
-      tabs: new Map([
-        [10, { id: 10, windowId: agentWindowId } as chrome.tabs.Tab],
-        [11, { id: 11, windowId: agentWindowId } as chrome.tabs.Tab],
-      ]),
-      windowsClosed: new Set(),
-      moves: [],
-    };
-    const { tabs, windows } = makeApis(state);
-    const query = makeQuery(state);
-
-    const res = await handleSessionStop(
-      sm,
-      { session_id: "aa11" },
-      { tabManagement: { tabs, windows }, tabsQuery: query },
-    );
-    if ("code" in res) throw new Error(`unexpected error: ${JSON.stringify(res)}`);
-    expect(res.window_released).toBeFalsy();
-    expect(aw.remove).toHaveBeenCalledWith(agentWindowId);
-    expect(sm.has("aa11")).toBe(false);
-  });
-
-  it("is non-fatal when an agent tab is already gone", async () => {
-    const aw = fakeAgentWindow([agentWindowId]);
-    aw.ensureActiveTab = vi.fn(async () => 10);
-    const sm = new SessionManager({ agentWindow: aw });
-    const ctx = await sm.start("aa11");
-    ctx.agentCreatedTabs.add(11); // no longer in state → remove throws
-    const state: FakeState = {
-      tabs: new Map([[10, { id: 10, windowId: agentWindowId } as chrome.tabs.Tab]]),
-      windowsClosed: new Set(),
-      moves: [],
-    };
-    const { tabs, windows } = makeApis(state);
-    (tabs.remove as ReturnType<typeof vi.fn>).mockImplementation(async (id: number) => {
-      if (id === 11) throw new Error("tab 11 already closed");
-      state.tabs.delete(id); // home (10) removed normally
-    });
-    const query = makeQuery(state);
-
-    const res = await handleSessionStop(
-      sm,
-      { session_id: "aa11" },
-      { tabManagement: { tabs, windows }, tabsQuery: query },
-    );
-    if ("code" in res) throw new Error(`unexpected error: ${JSON.stringify(res)}`);
-    // user tab 10? no — 10 is home, removed. window would be empty → close.
-    expect(res.window_released).toBeFalsy();
-    expect(aw.remove).toHaveBeenCalled();
-  });
-
-  it("degrades to close-window when tabsQuery is missing", async () => {
-    const aw = fakeAgentWindow([agentWindowId]);
-    aw.ensureActiveTab = vi.fn(async () => 10);
-    const sm = new SessionManager({ agentWindow: aw });
-    const ctx = await sm.start("aa11");
-    ctx.agentCreatedTabs.add(11);
-    const state: FakeState = {
-      tabs: new Map([
-        [10, { id: 10, windowId: agentWindowId } as chrome.tabs.Tab],
-        [11, { id: 11, windowId: agentWindowId } as chrome.tabs.Tab],
-      ]),
-      windowsClosed: new Set(),
-      moves: [],
-    };
-    const { tabs, windows } = makeApis(state);
-
-    const res = await handleSessionStop(
-      sm,
-      { session_id: "aa11" },
-      { tabManagement: { tabs, windows } }, // no tabsQuery
-    );
-    if ("code" in res) throw new Error(`unexpected error: ${JSON.stringify(res)}`);
-    // No query available → conservatively close the window.
-    expect(res.window_released).toBeFalsy();
-    expect(aw.remove).toHaveBeenCalledWith(agentWindowId);
-  });
-
-  it("degrades to close-window when tabsQuery.query throws", async () => {
-    const aw = fakeAgentWindow([agentWindowId]);
-    aw.ensureActiveTab = vi.fn(async () => 10);
-    const sm = new SessionManager({ agentWindow: aw });
-    const ctx = await sm.start("aa11");
-    ctx.agentCreatedTabs.add(11);
-    const state: FakeState = {
-      tabs: new Map([[11, { id: 11, windowId: agentWindowId } as chrome.tabs.Tab]]),
-      windowsClosed: new Set(),
-      moves: [],
-    };
-    const { tabs, windows } = makeApis(state);
-    const query = makeQuery(state);
-    (query.query as ReturnType<typeof vi.fn>).mockImplementation(async () => {
-      throw new Error("query failed");
-    });
-
-    const res = await handleSessionStop(
-      sm,
-      { session_id: "aa11" },
-      { tabManagement: { tabs, windows }, tabsQuery: query },
-    );
-    if ("code" in res) throw new Error(`unexpected error: ${JSON.stringify(res)}`);
-    expect(res.window_released).toBeFalsy();
-    expect(aw.remove).toHaveBeenCalled();
-  });
-
-  it("tracks agentCreatedTabs across tab_create / tab_close", async () => {
-    const aw = fakeAgentWindow([agentWindowId]);
-    aw.ensureActiveTab = vi.fn(async () => 10);
-    const sm = new SessionManager({ agentWindow: aw });
-    const ctx = await sm.start("aa11");
-    // Simulate handleTabCreate adding, then handleTabClose removing.
-    ctx.agentCreatedTabs.add(11);
-    expect(ctx.agentCreatedTabs.has(11)).toBe(true);
-    ctx.agentCreatedTabs.delete(11); // as handleTabClose does on success
-    expect(ctx.agentCreatedTabs.has(11)).toBe(false);
-    // Unknown id delete is a safe no-op (user-closed tab).
-    expect(() => ctx.agentCreatedTabs.delete(999)).not.toThrow();
-  });
-
-  it("closes the window (not release) when an agent tab fails to close", async () => {
-    // Regression: a leaked agent tab (still in agentCreatedTabs, but
-    // still present because Step 4 remove() threw) must NOT be mistaken
-    // for a user tab. Otherwise the window would be released (dropOnly)
-    // and the agent tab would leak (issue #57 regression).
-    const aw = fakeAgentWindow([agentWindowId]);
-    aw.ensureActiveTab = vi.fn(async () => 10);
-    const sm = new SessionManager({ agentWindow: aw });
-    const ctx = await sm.start("aa11");
-    ctx.agentCreatedTabs.add(11); // Step 4 remove() will throw for this tab
-    const state: FakeState = {
-      tabs: new Map([
-        [10, { id: 10, windowId: agentWindowId } as chrome.tabs.Tab],
-        [11, { id: 11, windowId: agentWindowId } as chrome.tabs.Tab],
-      ]),
-      windowsClosed: new Set(),
-      moves: [],
-    };
-    const { tabs, windows } = makeApis(state);
-    (tabs.remove as ReturnType<typeof vi.fn>).mockImplementation(async (id: number) => {
-      if (id === 11) throw new Error("tab 11 failed to close");
-      state.tabs.delete(id); // home (10) removed normally; 11 lingers
-    });
-    const query = makeQuery(state);
-
-    const res = await handleSessionStop(
-      sm,
-      { session_id: "aa11" },
-      { tabManagement: { tabs, windows }, tabsQuery: query },
-    );
-    if ("code" in res) throw new Error(`unexpected error: ${JSON.stringify(res)}`);
-    // The leaked agent tab must NOT keep the window open.
-    expect(res.window_released).toBeFalsy();
-    expect(aw.remove).toHaveBeenCalledWith(agentWindowId);
-    expect(sm.has("aa11")).toBe(false);
+    expect(ctx.surfaceCaptures.size()).toBe(0);
   });
 });

apps/extension/src/tools/__tests__/shared.test.ts

@@ -12,7 +12,7 @@ function fakeAgentWindow() {
   return {
     create: vi.fn(async () => 100),
     remove: vi.fn(async () => {}),
-    ensureActiveTab: vi.fn(async () => 1),
+    ensureActiveTab: vi.fn(async () => {}),
   };
 }
 

apps/extension/src/tools/__tests__/snapshot-ref.test.ts

@@ -11,7 +11,7 @@ function fakeAgentWindow(ids: number[]) {
       return id;
     },
     remove: async () => {},
-    ensureActiveTab: async () => 1,
+    ensureActiveTab: async () => {},
   };
 }
 
@@ -24,10 +24,16 @@ describe("lookupSnapshotRef", () => {
     expect(lookupSnapshotRef(ctx, "@e3", 4)).toEqual({
       backendNodeId: 1234,
       refKey: "e3",
+      kind: "dom",
+      capabilities: ["interact", "screenshot"],
+      generation: 0,
     });
     expect(lookupSnapshotRef(ctx, "e3", 4)).toEqual({
       backendNodeId: 1234,
       refKey: "e3",
+      kind: "dom",
+      capabilities: ["interact", "screenshot"],
+      generation: 0,
     });
   });
 
@@ -53,6 +59,8 @@ describe("resolveSnapshotRef", () => {
       tabId: 4,
       frameId: "child-frame",
       cdpSessionId: "child-session",
+      kind: "dom",
+      capabilities: ["interact", "screenshot"],
     });
 
     const expected = {
@@ -60,6 +68,9 @@ describe("resolveSnapshotRef", () => {
       refKey: "e3",
       frameId: "child-frame",
       cdpSessionId: "child-session",
+      kind: "dom" as const,
+      capabilities: ["interact", "screenshot"] as const,
+      generation: 0,
     };
     expect(lookupSnapshotRef(ctx, "@e3", 4)).toEqual(expected);
     expect(resolveSnapshotRef(ctx, "@e3", 4)).toEqual(expected);
@@ -105,6 +116,29 @@ describe("resolveSnapshotRef", () => {
     expect(resolveSnapshotRef(ctx, "@e3", 4)).toEqual({
       backendNodeId: 1234,
       refKey: "e3",
+      kind: "dom",
+      capabilities: ["interact", "screenshot"],
+      generation: 0,
+    });
+  });
+
+  it("rejects an operation outside the ref's declared capabilities", async () => {
+    const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
+    const ctx = await sm.start("aa11");
+    ctx.refStore.set("e3", 1234, {
+      tabId: 4,
+      kind: "surface",
+      capabilities: ["screenshot"],
+    });
+
+    expect(resolveSnapshotRef(ctx, "@e3", 4, "interact")).toMatchObject({
+      code: "permission_denied",
+      message: "ref @e3 does not support interact",
+      data: {
+        reason: "ref_capability_denied",
+        required_capability: "interact",
+        kind: "surface",
+      },
     });
   });
 });

apps/extension/src/tools/__tests__/surface-coordinate.test.ts

@@ -0,0 +1,52 @@
+import { describe, expect, it } from "vitest";
+import {
+  mapImagePointToViewport,
+  pointInRegion,
+  sameRect,
+  surfaceVisibleRect,
+} from "../surface-coordinate";
+
+describe("surface coordinate mapping", () => {
+  it("maps capture image pixels through the actual screenshot dimensions", () => {
+    expect(mapImagePointToViewport({ x: 100, y: 50, w: 400, h: 200 }, 800, 400, 200, 100)).toEqual({
+      x: 200,
+      y: 100,
+    });
+  });
+
+  it("rejects invalid image coordinates and dimensions", () => {
+    const rect = { x: 0, y: 0, w: 100, h: 50 };
+    expect(mapImagePointToViewport(rect, 100, 50, -1, 0)).toBeNull();
+    expect(mapImagePointToViewport(rect, 100, 50, 100, 0)).toBeNull();
+    expect(mapImagePointToViewport(rect, 100, 50, 0, 50)).toBeNull();
+    expect(mapImagePointToViewport(rect, 100, 50, Number.NaN, 0)).toBeNull();
+    expect(mapImagePointToViewport(rect, 100, 50, Number.POSITIVE_INFINITY, 0)).toBeNull();
+    expect(mapImagePointToViewport(rect, 0, 50, 0, 0)).toBeNull();
+  });
+
+  it("intersects live and observed visible regions and compares them strictly", () => {
+    const clipped = surfaceVisibleRect(
+      { x: 0, y: 0, width: 200, height: 100 },
+      { x: 25, y: 30, w: 50, h: 40 },
+    );
+    expect(clipped).toEqual({ x: 25, y: 30, w: 50, h: 40 });
+    expect(sameRect(clipped as NonNullable<typeof clipped>, { x: 25, y: 30, w: 50, h: 40 })).toBe(
+      true,
+    );
+    expect(sameRect(clipped as NonNullable<typeof clipped>, { x: 26, y: 30, w: 50, h: 40 })).toBe(
+      false,
+    );
+  });
+
+  it("checks the projected visible region rather than only its bounding box", () => {
+    const triangle = [
+      [
+        { x: 0, y: 0 },
+        { x: 100, y: 0 },
+        { x: 0, y: 100 },
+      ],
+    ];
+    expect(pointInRegion({ x: 20, y: 20 }, triangle)).toBe(true);
+    expect(pointInRegion({ x: 90, y: 90 }, triangle)).toBe(false);
+  });
+});

apps/extension/src/tools/__tests__/surface-target.test.ts

@@ -0,0 +1,306 @@
+import { describe, expect, it, vi } from "vitest";
+import { SessionManager } from "@/session-manager/manager";
+import type { CdpRunner } from "../shared";
+import {
+  captureSurfaceEnvironment,
+  resolveSurfacePointerTarget,
+  SURFACE_POINTER_MOVE_SETTLE_MS,
+} from "../surface-target";
+
+function fakeAgentWindow() {
+  return {
+    create: vi.fn(async () => 100),
+    remove: vi.fn(async () => {}),
+    ensureActiveTab: vi.fn(async () => {}),
+  };
+}
+
+function fakeBrowser() {
+  let loaderId = "loader-1";
+  let pageY = 0;
+  let quad = [10, 20, 110, 20, 110, 60, 10, 60];
+  const sent: Array<{ method: string; params?: object }> = [];
+  const cdp: CdpRunner = {
+    send: vi.fn(async (_tabId: number, method: string, params?: object) => {
+      sent.push({ method, params });
+      if (method === "Page.getFrameTree") {
+        return {
+          frameTree: { frame: { id: "main", loaderId, url: "https://fixture.test/" } },
+        } as never;
+      }
+      if (method === "Page.getLayoutMetrics") {
+        return {
+          cssLayoutViewport: { clientWidth: 800, clientHeight: 600, pageX: 0, pageY },
+          cssVisualViewport: {
+            clientWidth: 800,
+            clientHeight: 600,
+            pageX: 0,
+            pageY,
+            scale: 1,
+          },
+        } as never;
+      }
+      if (method === "DOM.getContentQuads") return { quads: [quad] } as never;
+      throw new Error(`unexpected CDP call ${method}`);
+    }) as CdpRunner["send"],
+    trackSessionTab: vi.fn(),
+  };
+  return {
+    cdp,
+    sent,
+    setLoaderId: (value: string) => {
+      loaderId = value;
+    },
+    setPageY: (value: number) => {
+      pageY = value;
+    },
+    setQuad: (value: number[]) => {
+      quad = value;
+    },
+  };
+}
+
+async function setupCapture() {
+  const manager = new SessionManager({ agentWindow: fakeAgentWindow() });
+  const context = await manager.start("aa11");
+  context.refStore.set("e3", 99, {
+    tabId: 4,
+    kind: "surface",
+    visibleRect: { x: 10, y: 20, w: 100, h: 40 },
+  });
+  const browser = fakeBrowser();
+  const environment = await captureSurfaceEnvironment(browser.cdp, 4, undefined);
+  if ("code" in environment) throw new Error(environment.message);
+  const entry = context.refStore.resolveEntry("e3");
+  if (!entry) throw new Error("missing ref");
+  const capture = context.surfaceCaptures.create({
+    sessionId: "aa11",
+    tabId: 4,
+    navigationIdentity: environment.navigationIdentity,
+    surface: {
+      ref: "e3",
+      backendNodeId: 99,
+      observationGeneration: entry.generation,
+    },
+    topViewportRect: { x: 10, y: 20, w: 100, h: 40 },
+    imageWidth: 200,
+    imageHeight: 80,
+    viewportSignature: environment.viewportSignature,
+    frameProjectionSignature: environment.frameProjectionSignature,
+  });
+  return { manager, context, browser, capture };
+}
+
+function pointParams(captureId: string) {
+  return {
+    ref: "@e3",
+    captureId,
+    imageX: 50,
+    imageY: 20,
+  };
+}
+
+describe("Surface screenshot-bound pointer target", () => {
+  it("maps one fresh capture coordinate without executing an action", async () => {
+    const { context, browser, capture } = await setupCapture();
+    const result = await resolveSurfacePointerTarget(
+      context,
+      { tabId: 4 },
+      pointParams(capture.id),
+      { cdp: browser.cdp },
+    );
+
+    if ("code" in result) throw new Error(JSON.stringify(result));
+    expect(result).toMatchObject({
+      usedRef: "e3",
+      point: { x: 35, y: 30 },
+      moveSettleMs: SURFACE_POINTER_MOVE_SETTLE_MS,
+      capture: { id: capture.id, imageX: 50, imageY: 20 },
+    });
+    expect(browser.sent.some((call) => call.method === "Input.dispatchMouseEvent")).toBe(false);
+
+    const repeated = await resolveSurfacePointerTarget(
+      context,
+      { tabId: 4 },
+      pointParams(capture.id),
+      { cdp: browser.cdp },
+    );
+    expect(repeated).toMatchObject({
+      code: "permission_denied",
+      data: { reason: "surface_capture_consumed" },
+    });
+  });
+
+  it("rejects a new observation generation before sending input", async () => {
+    const { context, browser, capture } = await setupCapture();
+    context.refStore.replace([
+      [
+        "e3",
+        {
+          backendNodeId: 99,
+          tabId: 4,
+          kind: "surface" as const,
+          visibleRect: { x: 10, y: 20, w: 100, h: 40 },
+        },
+      ],
+    ]);
+
+    const result = await resolveSurfacePointerTarget(
+      context,
+      { tabId: 4 },
+      pointParams(capture.id),
+      { cdp: browser.cdp },
+    );
+    expect(result).toMatchObject({
+      code: "permission_denied",
+      data: { reason: "surface_capture_stale" },
+    });
+    expect(browser.sent.some((call) => call.method === "Input.dispatchMouseEvent")).toBe(false);
+  });
+
+  it("rejects navigation, scroll, and visible geometry changes", async () => {
+    for (const mutate of [
+      (browser: ReturnType<typeof fakeBrowser>) => browser.setLoaderId("loader-2"),
+      (browser: ReturnType<typeof fakeBrowser>) => browser.setPageY(20),
+      (browser: ReturnType<typeof fakeBrowser>) =>
+        browser.setQuad([11, 20, 111, 20, 111, 60, 11, 60]),
+    ]) {
+      const { context, browser, capture } = await setupCapture();
+      mutate(browser);
+      const result = await resolveSurfacePointerTarget(
+        context,
+        { tabId: 4 },
+        pointParams(capture.id),
+        { cdp: browser.cdp },
+      );
+      expect(result).toMatchObject({
+        code: "permission_denied",
+        data: { reason: "surface_capture_stale" },
+      });
+      expect(browser.sent.some((call) => call.method === "Input.dispatchMouseEvent")).toBe(false);
+    }
+  });
+
+  it("consumes the capture when coordinate validation fails", async () => {
+    const invalid = await setupCapture();
+    const outside = await resolveSurfacePointerTarget(
+      invalid.context,
+      { tabId: 4 },
+      { ...pointParams(invalid.capture.id), imageX: 200 },
+      { cdp: invalid.browser.cdp },
+    );
+    expect(outside).toMatchObject({
+      code: "invalid_params",
+      data: { reason: "surface_coordinate_invalid" },
+    });
+    expect(
+      await resolveSurfacePointerTarget(
+        invalid.context,
+        { tabId: 4 },
+        pointParams(invalid.capture.id),
+        { cdp: invalid.browser.cdp },
+      ),
+    ).toMatchObject({ data: { reason: "surface_capture_consumed" } });
+  });
+
+  it("maps image coordinates through an OOPIF frame projection", async () => {
+    const manager = new SessionManager({ agentWindow: fakeAgentWindow() });
+    const context = await manager.start("aa11");
+    context.refStore.set("e3", 99, {
+      tabId: 4,
+      frameId: "child",
+      cdpSessionId: "child-session",
+      kind: "surface",
+      visibleRect: { x: 120, y: 110, w: 100, h: 40 },
+    });
+    const graph = {
+      rootFrameId: "main",
+      frames: [
+        { frameId: "main", target: { tabId: 4 } },
+        {
+          frameId: "child",
+          parentFrameId: "main",
+          ownerBackendNodeId: 77,
+          target: { tabId: 4, sessionId: "child-session" },
+        },
+      ],
+    };
+    const sent: Array<{ method: string; params?: object }> = [];
+    const cdp: CdpRunner = {
+      send: vi.fn(async (_tabId: number, method: string, params?: object) => {
+        sent.push({ method, params });
+        if (method === "Page.getFrameTree") {
+          return {
+            frameTree: {
+              frame: { id: "main", loaderId: "loader-1", url: "https://fixture.test/" },
+            },
+          } as never;
+        }
+        if (method === "Page.getLayoutMetrics") {
+          return {
+            cssLayoutViewport: { clientWidth: 800, clientHeight: 600, pageX: 0, pageY: 0 },
+            cssVisualViewport: {
+              clientWidth: 800,
+              clientHeight: 600,
+              pageX: 0,
+              pageY: 0,
+              scale: 1,
+            },
+          } as never;
+        }
+        if (method === "DOM.getBoxModel") {
+          return { model: { content: [100, 100, 300, 100, 300, 200, 100, 200] } } as never;
+        }
+        if (method === "Input.dispatchMouseEvent") return {} as never;
+        throw new Error(`unexpected root CDP call ${method}`);
+      }) as CdpRunner["send"],
+      sendToTarget: vi.fn(async (_target, method) => {
+        if (method === "DOM.getContentQuads") {
+          return { quads: [[20, 10, 120, 10, 120, 50, 20, 50]] } as never;
+        }
+        if (method === "Page.getLayoutMetrics") {
+          return { cssLayoutViewport: { clientWidth: 200, clientHeight: 100 } } as never;
+        }
+        throw new Error(`unexpected child CDP call ${method}`);
+      }) as CdpRunner["sendToTarget"],
+      getFrameGraph: vi.fn(async () => graph),
+      trackSessionTab: vi.fn(),
+    };
+    const environment = await captureSurfaceEnvironment(cdp, 4, "child");
+    if ("code" in environment) throw new Error(environment.message);
+    const entry = context.refStore.resolveEntry("e3");
+    if (!entry) throw new Error("missing ref");
+    const capture = context.surfaceCaptures.create({
+      sessionId: "aa11",
+      tabId: 4,
+      navigationIdentity: environment.navigationIdentity,
+      surface: {
+        ref: "e3",
+        frameId: "child",
+        backendNodeId: 99,
+        observationGeneration: entry.generation,
+      },
+      topViewportRect: { x: 120, y: 110, w: 100, h: 40 },
+      imageWidth: 200,
+      imageHeight: 80,
+      viewportSignature: environment.viewportSignature,
+      frameProjectionSignature: environment.frameProjectionSignature,
+    });
+
+    const result = await resolveSurfacePointerTarget(
+      context,
+      { tabId: 4 },
+      {
+        ref: "@e3",
+        captureId: capture.id,
+        imageX: 100,
+        imageY: 40,
+      },
+      { cdp },
+    );
+
+    if ("code" in result) throw new Error(JSON.stringify(result));
+    expect(result).toMatchObject({ point: { x: 170, y: 130 } });
+    expect(sent.some((call) => call.method === "Input.dispatchMouseEvent")).toBe(false);
+  });
+});

apps/extension/src/tools/__tests__/tabs.test.ts

@@ -1,7 +1,5 @@
 import { describe, expect, it, vi } from "vitest";
-import { type CdpDebuggerApi, ChromiumCdp } from "@/browser-driver/chromium-cdp";
 import { SessionManager } from "@/session-manager/manager";
-import { handleHover } from "../interaction";
 import {
   type AgentOverlayResetApi,
   type ChromeWindowsApi,
@@ -25,7 +23,7 @@ function fakeAgentWindow(ids: number[]) {
       return id;
     }),
     remove: vi.fn(async () => {}),
-    ensureActiveTab: vi.fn(async () => 1),
+    ensureActiveTab: vi.fn(async () => {}),
   };
 }
 
@@ -223,7 +221,6 @@ describe("handleTabCreate", () => {
       active: true,
     });
     expect(res).toMatchObject({ tab_id: 10, window_id: 100 });
-    expect(sm.get("aa11")?.agentCreatedTabs.has(10)).toBe(true);
   });
 
   it("forwards explicit url + index + active=false", async () => {
@@ -259,41 +256,6 @@ describe("handleTabCreate", () => {
     const res = await handleTabCreate(sm, { session_id: "aa11", index: -1 }, { tabs: api });
     expect(res).toMatchObject({ code: "invalid_params" });
   });
-
-  it("does not claim any tab when chrome.tabs.create fails", async () => {
-    const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
-    await sm.start("aa11");
-    const state: FakeTabState = { tabs: new Map(), nextTabId: 10, windowsClosed: new Set() };
-    const { api, spies } = makeTabMutationApi(state);
-    spies.create.mockRejectedValueOnce(new Error("chrome.tabs.create failed"));
-
-    const res = await handleTabCreate(sm, { session_id: "aa11" }, { tabs: api });
-    expect(res).toMatchObject({ code: "protocol_error" });
-
-    expect(sm.get("aa11")?.agentCreatedTabs).toEqual(new Set([1]));
-  });
-
-  it("claims the returned id even when a user tab appears during tab creation", async () => {
-    const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
-    const ctx = await sm.start("aa11");
-    const state: FakeTabState = { tabs: new Map(), nextTabId: 10, windowsClosed: new Set() };
-    const { api, spies } = makeTabMutationApi(state);
-    let resolveCreate!: (tab: chrome.tabs.Tab) => void;
-    const createResult = new Promise<chrome.tabs.Tab>((resolve) => {
-      resolveCreate = resolve;
-    });
-    spies.create.mockImplementationOnce(() => createResult);
-
-    const pending = handleTabCreate(sm, { session_id: "aa11" }, { tabs: api });
-    // A user can create an unrelated same-window tab while Chrome is still
-    // resolving the agent request. There is no pending slot for it to steal.
-    expect(ctx.agentCreatedTabs.has(99)).toBe(false);
-    resolveCreate({ id: 10, windowId: 100, active: true } as chrome.tabs.Tab);
-
-    await expect(pending).resolves.toMatchObject({ tab_id: 10 });
-    expect(ctx.agentCreatedTabs.has(10)).toBe(true);
-    expect(ctx.agentCreatedTabs.has(99)).toBe(false);
-  });
 });
 
 describe("handleTabClose", () => {
@@ -549,148 +511,6 @@ describe("handleTabBorrow", () => {
 });
 
 describe("handleTabReturn", () => {
-  it.each([
-    "not_found",
-    "cdp_failed",
-  ])("detaches after selector resolution fails with %s", async (code) => {
-    const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
-    const ctx = await sm.start("aa11");
-    ctx.borrowedTabs.set(7, { tabId: 7, originalWindowId: 200, originalIndex: 4 });
-    const state: FakeTabState = {
-      tabs: new Map([[7, { id: 7, windowId: 100 } as chrome.tabs.Tab]]),
-      nextTabId: 50,
-      windowsClosed: new Set(),
-    };
-    const { api } = makeTabMutationApi(state);
-    const { api: windows } = makeWindowsApi(state);
-    const debuggerApi = {
-      attach: vi.fn(async () => {}),
-      detach: vi.fn(async () => {}),
-      sendCommand: vi.fn(async (_target: unknown, method: string) => {
-        if (method === "DOM.getDocument") return { root: { nodeId: 1 } };
-        if (method === "DOM.querySelector") {
-          if (code === "cdp_failed") throw new Error("selector query failed");
-          return { nodeId: 0 };
-        }
-        return {};
-      }),
-      onEvent: { addListener: vi.fn(), removeListener: vi.fn() },
-      onDetach: { addListener: vi.fn(), removeListener: vi.fn() },
-    };
-    const cdp = new ChromiumCdp(debuggerApi as unknown as CdpDebuggerApi);
-
-    const hover = await handleHover(
-      sm,
-      { session_id: "aa11", tab_id: 7, selector: "#missing" },
-      {
-        cdp,
-        tabsApi: { get: api.get, query: vi.fn(async () => []) },
-      },
-    );
-    expect(hover).toMatchObject({ code });
-    expect(cdp.isAttached(7)).toBe(true);
-
-    const returned = await handleTabReturn(
-      sm,
-      { session_id: "aa11", tab_id: 7 },
-      {
-        tabs: api,
-        windows,
-        cdp,
-        agentOverlayReset: { resetAgentOverlays: vi.fn(async () => {}) },
-      },
-    );
-    expect(returned).toMatchObject({ returned_to_window_id: 200 });
-    expect(ctx.borrowedTabs.has(7)).toBe(false);
-    expect(cdp.isAttached(7)).toBe(false);
-    expect(debuggerApi.detach).toHaveBeenCalledExactlyOnceWith({ tabId: 7 });
-    cdp.dispose();
-  });
-
-  it("honours cancellation while preparing a validated return", async () => {
-    const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
-    const ctx = await sm.start("aa11");
-    ctx.borrowedTabs.set(7, { tabId: 7, originalWindowId: 200, originalIndex: 4 });
-    const state: FakeTabState = {
-      tabs: new Map([[7, { id: 7, windowId: 100 } as chrome.tabs.Tab]]),
-      nextTabId: 50,
-      windowsClosed: new Set(),
-    };
-    const { api, spies } = makeTabMutationApi(state);
-    const { api: windows } = makeWindowsApi(state);
-    const controller = new AbortController();
-    const beforeReturn = vi.fn(async () => {
-      controller.abort();
-    });
-    const result = await handleTabReturn(
-      sm,
-      { session_id: "aa11", tab_id: 7 },
-      {
-        tabs: api,
-        windows,
-        signal: controller.signal,
-        beforeReturn,
-      },
-    );
-
-    expect(result).toMatchObject({ code: "cancelled" });
-    expect(beforeReturn).toHaveBeenCalledExactlyOnceWith("aa11", 7);
-    expect(spies.move).not.toHaveBeenCalled();
-    expect(ctx.borrowedTabs.has(7)).toBe(true);
-  });
-
-  it.each([
-    "original",
-    "fallback",
-    "cleanup_failed",
-    "failed",
-    "cancelled",
-  ])("releases CDP only after a successful return (%s)", async (mode) => {
-    const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
-    const ctx = await sm.start("aa11");
-    ctx.borrowedTabs.set(7, { tabId: 7, originalWindowId: 200, originalIndex: 4 });
-    const state: FakeTabState = {
-      tabs: new Map([[7, { id: 7, windowId: 100 } as chrome.tabs.Tab]]),
-      nextTabId: 50,
-      windowsClosed: new Set(),
-    };
-    const { api, spies } = makeTabMutationApi(state);
-    const { api: windowsApi } = makeWindowsApi(state, { lastFocused: 500 });
-    const success = mode !== "failed" && mode !== "cancelled";
-    if (mode === "fallback") spies.move.mockRejectedValueOnce(new Error("original move failed"));
-    if (mode === "failed") spies.move.mockRejectedValue(new Error("move failed"));
-    let windowAtRelease: number | undefined;
-    const cdp = {
-      releaseSessionTab: vi.fn(async () => {
-        windowAtRelease = state.tabs.get(7)?.windowId;
-        if (mode === "cleanup_failed") throw new Error("debugger release failed");
-      }),
-    };
-    const controller = new AbortController();
-    if (mode === "cancelled") controller.abort();
-
-    const result = await handleTabReturn(
-      sm,
-      { session_id: "aa11", tab_id: 7 },
-      {
-        tabs: api,
-        windows: windowsApi,
-        cdp,
-        signal: controller.signal,
-        agentOverlayReset: { resetAgentOverlays: vi.fn(async () => {}) },
-      },
-    );
-
-    expect("code" in result).toBe(!success);
-    expect(ctx.borrowedTabs.has(7)).toBe(!success);
-    expect(cdp.releaseSessionTab).toHaveBeenCalledTimes(success ? 1 : 0);
-    if (success) {
-      expect(cdp.releaseSessionTab).toHaveBeenCalledWith("aa11", 7);
-      expect(windowAtRelease).toBe(mode === "fallback" ? 500 : 200);
-      expect(spies.move).toHaveBeenCalledTimes(mode === "fallback" ? 2 : 1);
-    }
-  });
-
   it("returns the tab to its original window/index and clears borrowedTabs", async () => {
     const sm = new SessionManager({ agentWindow: fakeAgentWindow([100]) });
     const ctx = await sm.start("aa11");

apps/extension/src/tools/__tests__/text-input.test.ts

@@ -0,0 +1,128 @@
+import { describe, expect, it, vi } from "vitest";
+import type { CdpRunner } from "../shared";
+import { dispatchTextInput } from "../text-input";
+
+function makeCdp(onSend?: (method: string, params?: object) => void) {
+  const sent: Array<{ method: string; params?: object }> = [];
+  const send = vi.fn(async <T>(_tabId: number, method: string, params?: object): Promise<T> => {
+    sent.push({ method, params });
+    onSend?.(method, params);
+    return {} as T;
+  });
+  return { cdp: { send } as CdpRunner, sent };
+}
+
+describe("dispatchTextInput", () => {
+  it("types ASCII through complete keydown, char, and keyup sequences", async () => {
+    const fake = makeCdp();
+
+    const result = await dispatchTextInput(fake.cdp, 4, "Ab");
+
+    expect(result).toBeNull();
+    expect(
+      fake.sent.map((call) => `${call.method}:${(call.params as { type?: string }).type}`),
+    ).toEqual([
+      "Input.dispatchKeyEvent:rawKeyDown",
+      "Input.dispatchKeyEvent:char",
+      "Input.dispatchKeyEvent:keyUp",
+      "Input.dispatchKeyEvent:rawKeyDown",
+      "Input.dispatchKeyEvent:char",
+      "Input.dispatchKeyEvent:keyUp",
+    ]);
+    expect(fake.sent[0].params).toMatchObject({ key: "A", code: "KeyA", modifiers: 8 });
+    expect(fake.sent[1].params).toMatchObject({ text: "A" });
+  });
+
+  it("types non-ASCII through the committed IME insertion path", async () => {
+    const fake = makeCdp();
+
+    const result = await dispatchTextInput(fake.cdp, 4, "测试");
+
+    expect(result).toBeNull();
+    expect(fake.sent).toEqual([
+      {
+        method: "Input.dispatchKeyEvent",
+        params: {
+          type: "rawKeyDown",
+          key: "Process",
+          code: "",
+          windowsVirtualKeyCode: 229,
+          nativeVirtualKeyCode: 229,
+          modifiers: 0,
+        },
+      },
+      { method: "Input.insertText", params: { text: "测试" } },
+      {
+        method: "Input.dispatchKeyEvent",
+        params: {
+          type: "keyUp",
+          key: "Process",
+          code: "",
+          windowsVirtualKeyCode: 229,
+          nativeVirtualKeyCode: 229,
+          modifiers: 0,
+        },
+      },
+    ]);
+  });
+
+  it("preserves text order across keyboard and IME segments", async () => {
+    const fake = makeCdp();
+
+    const result = await dispatchTextInput(fake.cdp, 4, "A测试1");
+
+    expect(result).toBeNull();
+    const meaningful = fake.sent.flatMap((call) => {
+      if (call.method === "Input.insertText") {
+        return [`ime:${(call.params as { text: string }).text}`];
+      }
+      if (
+        call.method === "Input.dispatchKeyEvent" &&
+        (call.params as { type?: string }).type === "char"
+      ) {
+        return [`key:${(call.params as { text: string }).text}`];
+      }
+      return [];
+    });
+    expect(meaningful).toEqual(["key:A", "ime:测试", "key:1"]);
+  });
+
+  it("reports cancellation observed after an accepted IME insertion", async () => {
+    const abort = new AbortController();
+    const fake = makeCdp((method) => {
+      if (method === "Input.insertText") abort.abort();
+    });
+
+    const result = await dispatchTextInput(fake.cdp, 4, "测试", abort.signal);
+
+    expect(result).toMatchObject({ code: "cancelled" });
+    expect(fake.sent.map((call) => call.method)).toEqual([
+      "Input.dispatchKeyEvent",
+      "Input.insertText",
+      "Input.dispatchKeyEvent",
+    ]);
+    expect(fake.sent.at(-1)?.params).toMatchObject({ type: "keyUp", key: "Process" });
+  });
+
+  it("releases the IME process key without committing when cancellation follows keydown", async () => {
+    const abort = new AbortController();
+    const fake = makeCdp((method, params) => {
+      if (
+        method === "Input.dispatchKeyEvent" &&
+        (params as { type?: string }).type === "rawKeyDown"
+      ) {
+        abort.abort();
+      }
+    });
+
+    const result = await dispatchTextInput(fake.cdp, 4, "测试", abort.signal);
+
+    expect(result).toMatchObject({ code: "cancelled" });
+    expect(fake.sent.map((call) => call.method)).toEqual([
+      "Input.dispatchKeyEvent",
+      "Input.dispatchKeyEvent",
+    ]);
+    expect(fake.sent[0].params).toMatchObject({ type: "rawKeyDown", key: "Process" });
+    expect(fake.sent[1].params).toMatchObject({ type: "keyUp", key: "Process" });
+  });
+});

apps/extension/src/tools/__tests__/waits.test.ts

@@ -12,7 +12,7 @@ function fakeAgentWindow(ids: number[]) {
       return id;
     }),
     remove: vi.fn(async () => {}),
-    ensureActiveTab: vi.fn(async () => 1),
+    ensureActiveTab: vi.fn(async () => {}),
   };
 }
 

apps/extension/src/tools/__tests__/window.test.ts

@@ -11,7 +11,7 @@ function fakeAgentWindow(ids: number[]) {
     return id;
   });
   const remove = vi.fn(async () => {});
-  const ensureActiveTab = vi.fn(async () => 1);
+  const ensureActiveTab = vi.fn(async () => {});
   return { create, remove, ensureActiveTab };
 }
 

apps/extension/src/tools/borrow-confirmation.ts

@@ -24,7 +24,7 @@ export interface BorrowResponseMessage {
   allowed: boolean;
 }
 
-export const CONFIRMATION_TIMEOUT_MS = 60_000;
+export const CONFIRMATION_TIMEOUT_MS = 5000;
 const EXIT_ANIMATION_MS = 150;
 /** Matches BorrowConfirmationOverlay progress ring/bar transition (duration-1000). */
 export const PROGRESS_TRANSITION_MS = 1000;
@@ -480,13 +480,6 @@ export async function requestBorrowConfirmation(
       settle(false);
     }, BACKGROUND_TIMEOUT_MS);
 
-    // Bring the user window to the front so the overlay is visible even
-    // when the Agent Window has stolen focus. This is fire-and-forget: if
-    // it fails the overlay / OS notification still give the user a path.
-    void windowsApi.update(notificationAnchor.windowId, { focused: true }).catch((err) => {
-      console.debug("[bsk borrow] proactive focus of user window failed", err);
-    });
-
     // Surface the OS notification *before* messaging any candidate so the
     // user has a parallel signal even if every content script is missing.
     // The Allow / Deny buttons let the user authorize explicitly when the

apps/extension/src/tools/browser-navigation.ts

@@ -1,110 +0,0 @@
-import type { RpcError, WaitUntil } from "@/transport/types";
-import { cdpError } from "./errors";
-
-interface NavigationEvent {
-  tabId: number;
-  frameId: number;
-  documentId?: string;
-  error?: string;
-}
-
-interface NavigationEvents {
-  addListener(listener: (details: NavigationEvent) => void): void;
-  removeListener(listener: (details: NavigationEvent) => void): void;
-}
-
-/** Normal browser navigation remains available when chrome.debugger is denied. */
-export interface BrowserNavigationApi {
-  update(tabId: number, props: { url: string }): Promise<unknown>;
-  reload(tabId: number, props: { bypassCache: boolean }): Promise<void>;
-  onCommitted: NavigationEvents;
-  onDOMContentLoaded: NavigationEvents;
-  onCompleted: NavigationEvents;
-  onErrorOccurred: NavigationEvents;
-}
-
-export const chromeBrowserNavigationApi: BrowserNavigationApi = {
-  update: (tabId, props) => chrome.tabs.update(tabId, props),
-  reload: (tabId, props) => chrome.tabs.reload(tabId, props),
-  get onCommitted() {
-    return chrome.webNavigation.onCommitted;
-  },
-  get onDOMContentLoaded() {
-    return chrome.webNavigation.onDOMContentLoaded;
-  },
-  get onCompleted() {
-    return chrome.webNavigation.onCompleted;
-  },
-  get onErrorOccurred() {
-    return chrome.webNavigation.onErrorOccurred;
-  },
-};
-
-export type NavigationOutcome =
-  | { reached: "match" | "timeout" | "cancelled"; lastLifecycle?: string }
-  | { reached: "failed"; error: RpcError };
-
-/** Subscribe before dispatch and ignore the departing document's load events. */
-export async function navigateWithBrowserApi(
-  api: BrowserNavigationApi,
-  tabId: number,
-  action: () => Promise<unknown>,
-  waitUntil: Exclude<WaitUntil, "networkidle">,
-  timeoutMs: number,
-  signal?: AbortSignal,
-): Promise<NavigationOutcome> {
-  if (signal?.aborted) return { reached: "cancelled" };
-  if (timeoutMs <= 0) return { reached: "timeout" };
-  let committed = false;
-  let documentId: string | undefined;
-  let lastLifecycle: string | undefined;
-  let finish!: (outcome: NavigationOutcome) => void;
-  const outcome = new Promise<NavigationOutcome>((resolve) => {
-    finish = resolve;
-  });
-  const isMainFrame = (details: NavigationEvent) =>
-    details.tabId === tabId && details.frameId === 0;
-  const onCommitted = (details: NavigationEvent) => {
-    if (!isMainFrame(details)) return;
-    committed = true;
-    documentId = details.documentId;
-    lastLifecycle = "commit";
-    if (waitUntil === "commit") finish({ reached: "match", lastLifecycle });
-  };
-  const onDOMContentLoaded = (details: NavigationEvent) => {
-    if (!isMainFrame(details) || !committed || (documentId && details.documentId !== documentId))
-      return;
-    lastLifecycle = "DOMContentLoaded";
-    if (waitUntil === "domcontentloaded") finish({ reached: "match", lastLifecycle });
-  };
-  const onCompleted = (details: NavigationEvent) => {
-    if (!isMainFrame(details) || !committed || (documentId && details.documentId !== documentId))
-      return;
-    finish({ reached: "match", lastLifecycle: "load" });
-  };
-  const onError = (details: NavigationEvent) => {
-    if (!isMainFrame(details)) return;
-    finish({ reached: "failed", error: cdpError(details.error ?? "browser navigation failed") });
-  };
-  const onAbort = () => finish({ reached: "cancelled", lastLifecycle });
-  const subscriptions = [
-    [api.onCommitted, onCommitted],
-    [api.onDOMContentLoaded, onDOMContentLoaded],
-    [api.onCompleted, onCompleted],
-    [api.onErrorOccurred, onError],
-  ] as const;
-  for (const [event, listener] of subscriptions) event.addListener(listener);
-  signal?.addEventListener("abort", onAbort, { once: true });
-  const timer = setTimeout(() => finish({ reached: "timeout", lastLifecycle }), timeoutMs);
-  try {
-    if (signal?.aborted) return { reached: "cancelled" };
-    await action();
-    return await outcome;
-  } catch (error) {
-    return { reached: "failed", error: cdpError(error) };
-  } finally {
-    clearTimeout(timer);
-    signal?.removeEventListener("abort", onAbort);
-    for (const [event, listener] of subscriptions) event.removeListener(listener);
-  }
-}

apps/extension/src/tools/dispatcher.ts

@@ -4,7 +4,6 @@ import type { Transport } from "@/transport/transport";
 import type {
   ClickParams,
   ConsoleParams,
-  DownloadParams,
   EmulateParams,
   EvaluateParams,
   FillParams,
@@ -29,17 +28,22 @@ import type {
   ScreenshotParams,
   SelectParams,
   SnapshotParams,
-  UploadParams,
+  TypeParams,
   WaitForNavigationParams,
 } from "@/transport/types";
 import { isRequestFrame } from "@/transport/types";
 import { handleConsole } from "./console";
-import { handleDownload } from "./download";
 import { type EmulateCdpRunner, handleEmulate } from "./emulate";
-import { classifyCdpError } from "./errors";
 import { handleEvaluate } from "./evaluate";
 import { handleRequestHelp } from "./human-loop";
-import { handleClick, handleFill, handleHover, handlePress, handleSelect } from "./interaction";
+import {
+  handleClick,
+  handleFill,
+  handleHover,
+  handlePress,
+  handleSelect,
+  handleType,
+} from "./interaction";
 import {
   handleNavigate,
   handleNavigateBack,
@@ -70,7 +74,6 @@ import {
 import { chromeTabsApi, lookupSession, resolveTargetTab } from "./shared";
 import {
   type BorrowConfirmationApprover,
-  chromeTabMutationApi,
   handleTabBorrow,
   handleTabClose,
   handleTabCreate,
@@ -84,7 +87,6 @@ import {
   type TabReturnParams,
   type TabSelectParams,
 } from "./tabs";
-import { handleUpload } from "./upload";
 import { handleWaitForNavigation } from "./waits";
 import { handleWindowResize, type WindowResizeParams } from "./window";
 
@@ -120,8 +122,6 @@ export interface DispatcherDeps {
   onSessionsChanged?: () => void;
   /** Invoked before a tool that dispatches page input or mutates browser state is forwarded. */
   onBrowserControlResumed?: (sessionId: string) => void;
-  /** Invoked after a tab is explicitly claimed so its overlay can be refreshed immediately. */
-  onAgentTabClaimed?: (tabId: number, windowId: number) => void;
   /** User approval for `tool.tab_borrow` (overlay in content script). */
   approveBorrow?: BorrowConfirmationApprover;
   /** i18n notification copy for `tool.request_help` (resolved per-call). */
@@ -151,7 +151,6 @@ export class ToolDispatcher {
   private readonly recording?: RecordRuntimeDeps;
   private readonly onSessionsChanged?: () => void;
   private readonly onBrowserControlResumed?: (sessionId: string) => void;
-  private readonly onAgentTabClaimed?: (tabId: number, windowId: number) => void;
   private readonly approveBorrow?: BorrowConfirmationApprover;
   private readonly helpNotificationCopy?: () => { title: string; body: string };
   private subscription: { dispose(): void } | null = null;
@@ -172,7 +171,6 @@ export class ToolDispatcher {
     this.recording = deps.recording;
     this.onSessionsChanged = deps.onSessionsChanged;
     this.onBrowserControlResumed = deps.onBrowserControlResumed;
-    this.onAgentTabClaimed = deps.onAgentTabClaimed;
     this.approveBorrow = deps.approveBorrow;
     this.helpNotificationCopy = deps.helpNotificationCopy;
   }
@@ -240,7 +238,7 @@ export class ToolDispatcher {
       if (sessionId) this.onBrowserControlResumed?.(sessionId);
       const result = await this.invoke(req, ac.signal);
       if (isRpcError(result)) {
-        body = { id: req.id, error: classifyCdpError(result) };
+        body = { id: req.id, error: result };
       } else {
         body = { id: req.id, result };
         if (req.method === "tool.session_start") {
@@ -308,25 +306,13 @@ export class ToolDispatcher {
         await this.releaseHoverLatch((req.params as SessionStopParams).session_id);
         return handleSessionStop(this.sessions, req.params as SessionStopParams, {
           cdp: this.cdp,
-          // Must be wired in production: the agent-tab cleanup and the
-          // window-release decision (issue #57) read these deps directly
-          // and silently no-op when they are absent.
-          tabManagement: { tabs: chromeTabMutationApi },
-          tabsQuery: chromeTabsApi,
           signal,
         });
       }
       case "tool.tab_list":
         return handleTabList(this.sessions, req.params as TabListParams, chromeTabsApi, signal);
-      case "tool.tab_create": {
-        const result = await handleTabCreate(this.sessions, req.params as TabCreateParams, {
-          signal,
-        });
-        if (!isRpcError(result)) {
-          this.onAgentTabClaimed?.(result.tab_id, result.window_id);
-        }
-        return result;
-      }
+      case "tool.tab_create":
+        return handleTabCreate(this.sessions, req.params as TabCreateParams, { signal });
       case "tool.tab_close":
         return this.withHoverReleaseForRequest(
           req.params as TabCloseParams,
@@ -335,22 +321,13 @@ export class ToolDispatcher {
         );
       case "tool.tab_select":
         return handleTabSelect(this.sessions, req.params as TabSelectParams, { signal });
-      case "tool.tab_borrow": {
-        const result = await handleTabBorrow(this.sessions, req.params as TabBorrowParams, {
+      case "tool.tab_borrow":
+        return handleTabBorrow(this.sessions, req.params as TabBorrowParams, {
           signal,
           approveBorrow: this.approveBorrow,
         });
-        if (!isRpcError(result)) {
-          this.onAgentTabClaimed?.(result.tab_id, result.agent_window_id);
-        }
-        return result;
-      }
       case "tool.tab_return":
-        return handleTabReturn(this.sessions, req.params as TabReturnParams, {
-          signal,
-          cdp: this.cdp,
-          beforeReturn: (sessionId, tabId) => this.releaseHoverLatch(sessionId, tabId),
-        });
+        return handleTabReturn(this.sessions, req.params as TabReturnParams, { signal });
       case "tool.window_resize":
         return handleWindowResize(
           this.sessions,
@@ -414,11 +391,7 @@ export class ToolDispatcher {
                 ? {
                     cdp: this.cdp,
                     tabsApi: chromeTabsCaptureApi,
-                    // Active hover probing is opt-in. A held hover latch still
-                    // suppresses it, because probing would move the cursor off
-                    // the element the caller is deliberately holding.
-                    conditionalSurfaceProbe:
-                      params.probe_hover === true && !this.hasHoverLatchForScope(hoverScope),
+                    conditionalSurfaceProbe: !this.hasHoverLatchForScope(hoverScope),
                     hoverProbeBypassOverlay: bypassOverlay,
                   }
                 : undefined,
@@ -526,6 +499,17 @@ export class ToolDispatcher {
             ),
           signal,
         );
+      case "tool.type":
+        return this.withHoverReleaseForRequest(
+          req.params as TypeParams,
+          () =>
+            handleType(
+              this.sessions,
+              req.params as TypeParams,
+              this.cdp ? { cdp: this.cdp, tabsApi: chromeTabsApi, signal } : undefined,
+            ),
+          signal,
+        );
       case "tool.press":
         return this.withHoverReleaseForRequest(
           req.params as PressParams,
@@ -548,40 +532,6 @@ export class ToolDispatcher {
             ),
           signal,
         );
-      case "tool.upload":
-        return this.withHoverReleaseForRequest(
-          req.params as UploadParams,
-          () =>
-            this.cdp
-              ? handleUpload(this.sessions, req.params as UploadParams, {
-                  cdp: this.cdp,
-                  tabsApi: chromeTabsApi,
-                  signal,
-                  bypassOverlay,
-                })
-              : Promise.resolve({
-                  code: "unsupported",
-                  message: "upload requires CDP",
-                } satisfies RpcError),
-          signal,
-        );
-      case "tool.download":
-        return this.withHoverReleaseForRequest(
-          req.params as DownloadParams,
-          () =>
-            this.cdp
-              ? handleDownload(this.sessions, req.params as DownloadParams, {
-                  cdp: this.cdp,
-                  tabsApi: chromeTabsApi,
-                  signal,
-                  bypassOverlay,
-                })
-              : Promise.resolve({
-                  code: "unsupported",
-                  message: "download requires CDP",
-                } satisfies RpcError),
-          signal,
-        );
       case "tool.evaluate":
         return handleEvaluate(
           this.sessions,
@@ -786,10 +736,9 @@ function sessionIdForBrowserControlMethod(req: RequestFrame): string | null {
     case "tool.click":
     case "tool.hover":
     case "tool.fill":
+    case "tool.type":
     case "tool.press":
     case "tool.select":
-    case "tool.upload":
-    case "tool.download":
     case "tool.evaluate":
     case "tool.observe":
     case "tool.request_help":

apps/extension/src/tools/download-capture.ts

@@ -1,364 +0,0 @@
-// Order-independent coordinator for one browser download. CDP supplies the
-// exact target/frame intent while chrome.downloads supplies the download id
-// and filename routing hook; neither event is assumed to arrive first.
-
-import type { CdpTarget } from "@/browser-driver/frame-graph";
-import type { ClickResult, RpcError, TransferEffectState } from "@/transport/types";
-import { transferError } from "./errors";
-import { type CdpRunner, isRpcError } from "./shared";
-
-const CORRELATION_GRACE_MS = 750;
-const UNIQUE_SETTLE_MS = 50;
-const SIZE_POLL_MS = 250;
-
-type DeterminingFilenameListener = (
-  item: chrome.downloads.DownloadItem,
-  suggest: (suggestion?: chrome.downloads.DownloadFilenameSuggestion) => void,
-) => void | true;
-
-interface ListenerEvent<T> {
-  addListener(listener: T): void;
-  removeListener(listener: T): void;
-}
-
-export interface DownloadsApi {
-  onCreated: ListenerEvent<(item: chrome.downloads.DownloadItem) => void>;
-  onChanged: ListenerEvent<(delta: chrome.downloads.DownloadDelta) => void>;
-  onDeterminingFilename: ListenerEvent<DeterminingFilenameListener>;
-  search(query: chrome.downloads.DownloadQuery): Promise<chrome.downloads.DownloadItem[]>;
-  cancel(downloadId: number): Promise<void>;
-  removeFile(downloadId: number): Promise<void>;
-}
-
-export const chromeDownloadsApi: DownloadsApi = {
-  get onCreated() {
-    return chrome.downloads.onCreated;
-  },
-  get onChanged() {
-    return chrome.downloads.onChanged;
-  },
-  get onDeterminingFilename() {
-    return chrome.downloads.onDeterminingFilename;
-  },
-  search: (query) => chrome.downloads.search(query),
-  cancel: (id) => chrome.downloads.cancel(id),
-  removeFile: (id) => chrome.downloads.removeFile(id),
-};
-
-export interface DownloadCaptureOptions {
-  cdp: CdpRunner;
-  target: CdpTarget;
-  expectedFrameId?: string;
-  downloads: DownloadsApi;
-  browserRelativeDir: string;
-  maxByteSize?: number;
-  timeoutMs: number;
-  signal?: AbortSignal;
-  trigger(): Promise<ClickResult | RpcError>;
-}
-
-export interface DownloadCaptureResult {
-  click: ClickResult;
-  item: chrome.downloads.DownloadItem;
-}
-
-interface DownloadIntent {
-  url: string;
-  suggestedFilename: string;
-  frameId?: string;
-}
-
-interface DownloadCandidate {
-  item: chrome.downloads.DownloadItem;
-  suggest: (suggestion?: chrome.downloads.DownloadFilenameSuggestion) => void;
-  suggested: boolean;
-  graceTimer: ReturnType<typeof setTimeout>;
-}
-
-function safeBasename(filename: string): string {
-  const basename = filename.split(/[\\/]/).pop()?.trim();
-  return basename && basename !== "." && basename !== ".." ? basename : "download";
-}
-
-function sameTarget(source: { tabId?: number; sessionId?: string }, target: CdpTarget): boolean {
-  return source.tabId === target.tabId && source.sessionId === target.sessionId;
-}
-
-function matchesIntent(item: chrome.downloads.DownloadItem, intent: DownloadIntent): boolean {
-  const urlMatches = item.url === intent.url || item.finalUrl === intent.url;
-  return urlMatches && safeBasename(item.filename) === safeBasename(intent.suggestedFilename);
-}
-
-function knownSize(item: chrome.downloads.DownloadItem): number | undefined {
-  if (item.fileSize >= 0) return item.fileSize;
-  if (item.totalBytes >= 0) return item.totalBytes;
-  return undefined;
-}
-
-function captureError(
-  message: string,
-  effectState: TransferEffectState,
-  phase: string,
-  cleanupFailed = false,
-): RpcError {
-  return transferError("cdp_failed", "download_capture_failed", message, {
-    effectState,
-    phase,
-    ...(cleanupFailed ? { cleanupState: "failed" } : {}),
-  });
-}
-
-async function cleanupClaimedDownload(downloads: DownloadsApi, downloadId: number): Promise<void> {
-  const lookup = async () => (await downloads.search({ id: downloadId }))[0];
-  const item = await lookup();
-  if (!item || item.state === "interrupted") return;
-  if (item.state === "complete") {
-    await downloads.removeFile(downloadId);
-    return;
-  }
-
-  try {
-    await downloads.cancel(downloadId);
-  } catch (cancelError) {
-    // Completion can win the race after the lookup but before cancellation.
-    // Reconcile against Chrome's authoritative state before declaring cleanup
-    // failed so every terminal state has one explicit cleanup path.
-    const reconciled = await lookup();
-    if (!reconciled || reconciled.state === "interrupted") return;
-    if (reconciled.state === "complete") {
-      await downloads.removeFile(downloadId);
-      return;
-    }
-    throw cancelError;
-  }
-}
-
-export async function captureBrowserDownload(
-  options: DownloadCaptureOptions,
-): Promise<DownloadCaptureResult | RpcError> {
-  let click: ClickResult | undefined;
-  let intent: DownloadIntent | undefined;
-  let capturedId: number | undefined;
-  let settled = false;
-  let succeeded = false;
-  let failureResult: RpcError | undefined;
-  let uniquenessTimer: ReturnType<typeof setTimeout> | undefined;
-  let operationTimer: ReturnType<typeof setTimeout> | undefined;
-  let sizePoll: ReturnType<typeof setInterval> | undefined;
-  const candidates = new Map<number, DownloadCandidate>();
-  const createdItems = new Map<number, chrome.downloads.DownloadItem>();
-
-  let resolveCompletion!: (item: chrome.downloads.DownloadItem) => void;
-  let rejectCompletion!: (error: Error) => void;
-  const completion = new Promise<chrome.downloads.DownloadItem>((resolve, reject) => {
-    resolveCompletion = resolve;
-    rejectCompletion = reject;
-  });
-  const fail = (error: Error) => {
-    if (settled) return;
-    settled = true;
-    rejectCompletion(error);
-  };
-  const complete = (item: chrome.downloads.DownloadItem) => {
-    if (settled) return;
-    const size = knownSize(item);
-    if (size !== undefined && options.maxByteSize !== undefined && size > options.maxByteSize) {
-      fail(new Error(`download exceeds transfer limit ${options.maxByteSize}`));
-      return;
-    }
-    settled = true;
-    resolveCompletion(item);
-  };
-  const suggestDefault = (candidate: DownloadCandidate) => {
-    if (candidate.suggested) return;
-    candidate.suggested = true;
-    candidate.suggest();
-  };
-  const matchingCandidates = (): DownloadCandidate[] => {
-    const currentIntent = intent;
-    return currentIntent
-      ? [...candidates.values()].filter(
-          (candidate) => !candidate.suggested && matchesIntent(candidate.item, currentIntent),
-        )
-      : [];
-  };
-
-  const claimUnique = () => {
-    uniquenessTimer = undefined;
-    if (settled || capturedId !== undefined || !intent) return;
-    const matches = matchingCandidates();
-    if (matches.length !== 1) {
-      if (matches.length > 1) {
-        for (const candidate of matches) suggestDefault(candidate);
-        fail(new Error("download attribution is ambiguous"));
-      }
-      return;
-    }
-    const candidate = matches[0];
-    candidate.suggested = true;
-    clearTimeout(candidate.graceTimer);
-    capturedId = candidate.item.id;
-    candidate.suggest({
-      filename: `${options.browserRelativeDir}/${safeBasename(intent.suggestedFilename)}`,
-      conflictAction: "overwrite",
-    });
-    const size = knownSize(candidate.item);
-    if (size !== undefined && options.maxByteSize !== undefined && size > options.maxByteSize) {
-      fail(new Error(`download exceeds transfer limit ${options.maxByteSize}`));
-      return;
-    }
-    const created = createdItems.get(candidate.item.id);
-    if (created?.state === "interrupted") {
-      fail(new Error(created.error ?? "download interrupted"));
-    } else if (created?.state === "complete") {
-      complete(created);
-    }
-  };
-  const reconcile = () => {
-    if (settled || capturedId !== undefined || !intent) return;
-    const matches = matchingCandidates();
-    if (matches.length > 1) {
-      for (const candidate of matches) suggestDefault(candidate);
-      fail(new Error("download attribution is ambiguous"));
-      return;
-    }
-    if (matches.length === 1 && !uniquenessTimer) {
-      uniquenessTimer = setTimeout(claimUnique, UNIQUE_SETTLE_MS);
-    }
-  };
-
-  const determiningListener: DeterminingFilenameListener = (item, suggest) => {
-    const candidate: DownloadCandidate = {
-      item,
-      suggest,
-      suggested: false,
-      graceTimer: setTimeout(() => {
-        suggestDefault(candidate);
-        candidates.delete(item.id);
-        if (intent && matchesIntent(item, intent) && capturedId === undefined) {
-          fail(new Error("download correlation grace elapsed before unique attribution"));
-        }
-      }, CORRELATION_GRACE_MS),
-    };
-    candidates.set(item.id, candidate);
-    reconcile();
-    return true;
-  };
-  const createdListener = (item: chrome.downloads.DownloadItem) => {
-    createdItems.set(item.id, item);
-    if (capturedId !== item.id) return;
-    if (item.state === "interrupted") {
-      fail(new Error(item.error ?? "download interrupted"));
-    } else if (item.state === "complete") {
-      complete(item);
-    }
-  };
-  const changedListener = async (delta: chrome.downloads.DownloadDelta) => {
-    if (capturedId === undefined || delta.id !== capturedId || settled) return;
-    if (delta.state?.current === "interrupted" || delta.error?.current) {
-      fail(new Error(delta.error?.current ?? "download interrupted"));
-      return;
-    }
-    if (delta.state?.current === "complete") {
-      try {
-        const [item] = await options.downloads.search({ id: delta.id });
-        if (item) complete(item);
-        else fail(new Error("completed download disappeared"));
-      } catch (err) {
-        fail(err instanceof Error ? err : new Error(String(err)));
-      }
-    }
-  };
-  const onAbort = () => fail(new DOMException("aborted", "AbortError"));
-  const cdpSubscription = options.cdp.onEvent?.((source, method, raw) => {
-    if (method !== "Page.downloadWillBegin" || !sameTarget(source, options.target)) return;
-    const event = raw as { url?: unknown; suggestedFilename?: unknown; frameId?: unknown };
-    if (typeof event.url !== "string" || typeof event.suggestedFilename !== "string") return;
-    if (options.expectedFrameId && event.frameId !== options.expectedFrameId) {
-      fail(new Error("download originated from a different frame"));
-      return;
-    }
-    if (intent) {
-      fail(new Error("download trigger produced more than one browser download intent"));
-      return;
-    }
-    intent = {
-      url: event.url,
-      suggestedFilename: event.suggestedFilename,
-      ...(typeof event.frameId === "string" ? { frameId: event.frameId } : {}),
-    };
-    reconcile();
-  });
-  if (!cdpSubscription) {
-    return captureError("CDP download intent subscription unavailable", "none", "arm");
-  }
-
-  options.downloads.onDeterminingFilename.addListener(determiningListener);
-  options.downloads.onCreated.addListener(createdListener);
-  options.downloads.onChanged.addListener(changedListener);
-  options.signal?.addEventListener("abort", onAbort, { once: true });
-  operationTimer = setTimeout(
-    () => fail(new Error("download did not complete before timeout")),
-    options.timeoutMs,
-  );
-  sizePoll = setInterval(() => {
-    if (capturedId === undefined || settled || options.maxByteSize === undefined) return;
-    void options.downloads
-      .search({ id: capturedId })
-      .then(([item]) => {
-        if (!item || settled) return;
-        if (item.bytesReceived > (options.maxByteSize as number)) {
-          fail(new Error(`download exceeds transfer limit ${options.maxByteSize}`));
-        }
-      })
-      .catch((err) => fail(err instanceof Error ? err : new Error(String(err))));
-  }, SIZE_POLL_MS);
-
-  try {
-    const triggered = await options.trigger();
-    if (isRpcError(triggered)) {
-      void completion.catch(() => undefined);
-      const effect: TransferEffectState =
-        capturedId !== undefined ? "committed" : intent ? "unknown" : "none";
-      failureResult = {
-        ...triggered,
-        data: { ...triggered.data, effect_state: effect, phase: "trigger" },
-      };
-      return failureResult;
-    }
-    click = triggered;
-    const item = await completion;
-    succeeded = true;
-    return { click, item };
-  } catch (err) {
-    const effect: TransferEffectState =
-      capturedId !== undefined ? "committed" : click ? "unknown" : "none";
-    failureResult = captureError(
-      err instanceof Error ? err.message : String(err),
-      effect,
-      capturedId !== undefined ? "download" : "attribution",
-    );
-    return failureResult;
-  } finally {
-    settled = true;
-    if (operationTimer) clearTimeout(operationTimer);
-    if (uniquenessTimer) clearTimeout(uniquenessTimer);
-    if (sizePoll) clearInterval(sizePoll);
-    options.signal?.removeEventListener("abort", onAbort);
-    options.downloads.onDeterminingFilename.removeListener(determiningListener);
-    options.downloads.onCreated.removeListener(createdListener);
-    options.downloads.onChanged.removeListener(changedListener);
-    cdpSubscription.dispose();
-    for (const candidate of candidates.values()) {
-      clearTimeout(candidate.graceTimer);
-      if (candidate.item.id !== capturedId) suggestDefault(candidate);
-    }
-    if (!succeeded && capturedId !== undefined) {
-      try {
-        await cleanupClaimedDownload(options.downloads, capturedId);
-      } catch {
-        if (failureResult?.data) failureResult.data.cleanup_state = "failed";
-      }
-    }
-  }
-}

apps/extension/src/tools/download.ts

@@ -1,64 +0,0 @@
-// Download orchestration: validate session/tab ownership, then delegate one
-// browser-global chrome.downloads transaction to download-capture.ts.
-
-import type { SessionManager } from "@/session-manager/manager";
-import type { DownloadParams, DownloadResult, RpcError } from "@/transport/types";
-import { captureBrowserDownload, chromeDownloadsApi, type DownloadsApi } from "./download-capture";
-import { clickResolvedTarget, type InteractionDeps, resolveActionTarget } from "./interaction";
-import { enforceAgentWindow, isRpcError, lookupSession, resolveTargetTab } from "./shared";
-
-let downloadActive = false;
-
-export type { DownloadsApi } from "./download-capture";
-
-export interface DownloadDeps extends InteractionDeps {
-  downloads?: DownloadsApi;
-}
-
-export async function handleDownload(
-  manager: SessionManager,
-  params: DownloadParams,
-  deps: DownloadDeps,
-): Promise<DownloadResult | RpcError> {
-  if (downloadActive) return { code: "invalid_params", message: "another bsk download is active" };
-  downloadActive = true;
-  try {
-    const ctx = lookupSession(manager, params, "download");
-    if (isRpcError(ctx)) return ctx;
-    const target = await resolveTargetTab(manager, ctx, params.tab_id, deps.tabsApi);
-    if (isRpcError(target)) return target;
-    const denied = enforceAgentWindow(ctx, target, "download");
-    if (denied) return denied;
-    if (!params.browser_relative_dir) {
-      return { code: "invalid_params", message: "download requires a daemon capability directory" };
-    }
-    const address = await resolveActionTarget(deps.cdp, ctx, target, params, "download");
-    if (isRpcError(address)) return address;
-
-    const capture = await captureBrowserDownload({
-      cdp: deps.cdp,
-      target: address.cdpTarget,
-      downloads: deps.downloads ?? chromeDownloadsApi,
-      browserRelativeDir: params.browser_relative_dir,
-      maxByteSize: params.max_byte_size,
-      timeoutMs: params.timeout_ms ?? 120_000,
-      signal: deps.signal,
-      expectedFrameId: address.frameId,
-      trigger: () => clickResolvedTarget(ctx, address, {}, deps),
-    });
-    if (isRpcError(capture)) return capture;
-    const { click, item } = capture;
-    return {
-      tab_id: target.tabId,
-      used_ref: click.used_ref,
-      used_selector: click.used_selector,
-      suggested_filename: item.filename.split(/[\\/]/).pop() ?? "download",
-      byte_size: item.fileSize >= 0 ? item.fileSize : item.totalBytes,
-      mime: item.mime || undefined,
-      danger: item.danger,
-      browser_path: item.filename,
-    };
-  } finally {
-    downloadActive = false;
-  }
-}

apps/extension/src/tools/editable-input.ts

@@ -0,0 +1,282 @@
+// Browser-native editing transaction for input, textarea, and contenteditable
+// targets. The CDP target/session is selected by interaction.ts; this module
+// owns the shorter-lived editable identity inside that target.
+
+import type { RpcError } from "@/transport/types";
+import { rpcError } from "./errors";
+import type { CdpRunner } from "./shared";
+
+interface EditableState {
+  supported: boolean;
+  connected: boolean;
+  focused: boolean;
+  value: string;
+  kind: "input" | "textarea" | "contenteditable" | "unsupported";
+}
+
+export interface EditableFillResult {
+  value: string;
+  replacedNode: boolean;
+}
+
+function cancelled(signal: AbortSignal | undefined): RpcError | null {
+  return signal?.aborted ? { code: "cancelled", message: "fill aborted" } : null;
+}
+
+async function inspectEditable(
+  cdp: CdpRunner,
+  tabId: number,
+  objectId: string,
+): Promise<EditableState> {
+  const reply = await cdp.send<{
+    result?: { value?: EditableState };
+  }>(tabId, "Runtime.callFunctionOn", {
+    objectId,
+    functionDeclaration: `function() {
+      /* bsk-editable-state */
+      const doc = this.ownerDocument;
+      const focused = !!doc && doc.activeElement === this;
+      if (this instanceof HTMLInputElement) {
+        return { supported: true, connected: this.isConnected, focused,
+          value: this.value, kind: 'input' };
+      }
+      if (this instanceof HTMLTextAreaElement) {
+        return { supported: true, connected: this.isConnected, focused,
+          value: this.value, kind: 'textarea' };
+      }
+      if (this.isContentEditable) {
+        return { supported: true, connected: this.isConnected, focused,
+          value: this.innerText ?? this.textContent ?? '', kind: 'contenteditable' };
+      }
+      return { supported: false, connected: this.isConnected === true, focused,
+        value: '', kind: 'unsupported' };
+    }`,
+    returnByValue: true,
+  });
+  const state = reply.result?.value;
+  if (!state || typeof state.value !== "string") {
+    throw new Error("editable state probe returned no value");
+  }
+  return state;
+}
+
+/**
+ * Return the active editable from the anchor's own document. Using the
+ * anchor's execution context also works for same-process nested frames, where
+ * a target-level Runtime.evaluate would otherwise see the top document.
+ */
+async function focusedEditableFromAnchor(
+  cdp: CdpRunner,
+  tabId: number,
+  anchorObjectId: string,
+): Promise<string | null> {
+  const reply = await cdp.send<{
+    result?: { objectId?: string; subtype?: string };
+  }>(tabId, "Runtime.callFunctionOn", {
+    objectId: anchorObjectId,
+    functionDeclaration: `function() {
+      /* bsk-editable-live */
+      const active = this.ownerDocument?.activeElement;
+      if (!active) return null;
+      const editable = active instanceof HTMLInputElement ||
+        active instanceof HTMLTextAreaElement || active.isContentEditable;
+      return editable ? active : null;
+    }`,
+    returnByValue: false,
+  });
+  if (reply.result?.subtype === "null") return null;
+  return typeof reply.result?.objectId === "string" ? reply.result.objectId : null;
+}
+
+async function resolveLiveEditable(
+  cdp: CdpRunner,
+  tabId: number,
+  anchorObjectId: string,
+): Promise<{ objectId: string; state: EditableState; replaced: boolean } | RpcError> {
+  const anchor = await inspectEditable(cdp, tabId, anchorObjectId);
+  if (anchor.supported && anchor.connected && anchor.focused) {
+    return { objectId: anchorObjectId, state: anchor, replaced: false };
+  }
+
+  const focusedObjectId = await focusedEditableFromAnchor(cdp, tabId, anchorObjectId);
+  if (!focusedObjectId) {
+    return rpcError(
+      "cdp_failed",
+      "input_not_applied",
+      "the target editable lost focus or detached before input",
+      { phase: "resolve_live_editable" },
+    );
+  }
+  const focused = await inspectEditable(cdp, tabId, focusedObjectId);
+  if (!focused.supported || !focused.connected || !focused.focused) {
+    return rpcError(
+      "cdp_failed",
+      "input_not_applied",
+      "the replacement editable was not live and focused",
+      { phase: "resolve_live_editable" },
+    );
+  }
+  return { objectId: focusedObjectId, state: focused, replaced: true };
+}
+
+async function selectEditableContents(
+  cdp: CdpRunner,
+  tabId: number,
+  objectId: string,
+): Promise<void> {
+  const reply = await cdp.send<{
+    result?: { value?: { selected?: boolean } };
+  }>(tabId, "Runtime.callFunctionOn", {
+    objectId,
+    functionDeclaration: `function() {
+      /* bsk-editable-selection */
+      if (!this.isConnected || this.ownerDocument?.activeElement !== this) {
+        return { selected: false };
+      }
+      if (this instanceof HTMLInputElement || this instanceof HTMLTextAreaElement) {
+        try {
+          this.setSelectionRange(0, this.value.length);
+        } catch {
+          if (typeof this.select !== 'function') return { selected: false };
+          this.select();
+        }
+        return { selected: true };
+      }
+      if (this.isContentEditable) {
+        const selection = this.ownerDocument.getSelection();
+        if (!selection) return { selected: false };
+        const range = this.ownerDocument.createRange();
+        range.selectNodeContents(this);
+        selection.removeAllRanges();
+        selection.addRange(range);
+        return { selected: true };
+      }
+      return { selected: false };
+    }`,
+    returnByValue: true,
+  });
+  if (reply.result?.value?.selected !== true) {
+    throw new Error("could not select the live editable contents");
+  }
+}
+
+async function waitForEditableSettlement(
+  cdp: CdpRunner,
+  tabId: number,
+  anchorObjectId: string,
+): Promise<void> {
+  await cdp.send(tabId, "Runtime.callFunctionOn", {
+    objectId: anchorObjectId,
+    functionDeclaration: `function() {
+      /* bsk-editable-sync */
+      return new Promise((resolve) => {
+        let settled = false;
+        const finish = () => {
+          if (settled) return;
+          settled = true;
+          clearTimeout(timer);
+          resolve();
+        };
+        const timer = setTimeout(finish, 100);
+        const view = this.ownerDocument?.defaultView;
+        if (!view || typeof view.requestAnimationFrame !== 'function') { finish(); return; }
+        view.requestAnimationFrame(() => view.requestAnimationFrame(finish));
+      });
+    }`,
+    awaitPromise: true,
+    returnByValue: true,
+  });
+}
+
+async function deleteSelection(cdp: CdpRunner, tabId: number): Promise<void> {
+  await cdp.send(tabId, "Input.dispatchKeyEvent", {
+    type: "rawKeyDown",
+    key: "Backspace",
+    code: "Backspace",
+    windowsVirtualKeyCode: 8,
+  });
+  await cdp.send(tabId, "Input.dispatchKeyEvent", {
+    type: "keyUp",
+    key: "Backspace",
+    code: "Backspace",
+    windowsVirtualKeyCode: 8,
+  });
+}
+
+/**
+ * Replace or append text through Chromium's editing pipeline. No value setter
+ * or synthetic DOM events are used: controlled components observe the same
+ * beforeinput/input path as a real browser edit. The live editable is resolved
+ * again after settlement so a framework may remount it without producing a
+ * false failure or a stale-object success.
+ */
+export async function fillEditable(
+  cdp: CdpRunner,
+  tabId: number,
+  anchorObjectId: string,
+  value: string,
+  clearBefore: boolean,
+  signal?: AbortSignal,
+): Promise<EditableFillResult | RpcError> {
+  try {
+    const abortedBefore = cancelled(signal);
+    if (abortedBefore) return abortedBefore;
+
+    const initial = await resolveLiveEditable(cdp, tabId, anchorObjectId);
+    if ("code" in initial) return initial;
+    let replacedNode = initial.replaced;
+
+    if (clearBefore) await selectEditableContents(cdp, tabId, initial.objectId);
+    const abortedAfterSelection = cancelled(signal);
+    if (abortedAfterSelection) return abortedAfterSelection;
+
+    if (value.length > 0) {
+      await cdp.send(tabId, "Input.insertText", { text: value });
+    } else if (clearBefore) {
+      await deleteSelection(cdp, tabId);
+    }
+
+    const abortedAfterInput = cancelled(signal);
+    if (abortedAfterInput) return abortedAfterInput;
+    await waitForEditableSettlement(cdp, tabId, initial.objectId);
+
+    const abortedAfterSettlement = cancelled(signal);
+    if (abortedAfterSettlement) return abortedAfterSettlement;
+
+    let finalObjectId = initial.objectId;
+    let finalState = await inspectEditable(cdp, tabId, finalObjectId);
+    if (!finalState.connected || !finalState.focused) {
+      const live = await resolveLiveEditable(cdp, tabId, initial.objectId);
+      if ("code" in live) return live;
+      finalObjectId = live.objectId;
+      finalState = live.state;
+      replacedNode ||= live.replaced || finalObjectId !== initial.objectId;
+    }
+
+    const applied = clearBefore
+      ? finalState.value === value
+      : value.length === 0 || finalState.value.includes(value);
+    if (!finalState.supported || !applied) {
+      return rpcError(
+        "cdp_failed",
+        "input_not_applied",
+        "the page did not retain the requested input value",
+        {
+          expected_value_length: value.length,
+          actual_value_length: finalState.value.length,
+          phase: "readback",
+          editable_replaced: replacedNode,
+        },
+      );
+    }
+
+    return { value: finalState.value, replacedNode };
+  } catch (err) {
+    return rpcError(
+      "cdp_failed",
+      "input_not_applied",
+      err instanceof Error ? err.message : String(err),
+      { phase: "edit_transaction" },
+    );
+  }
+}

apps/extension/src/tools/errors.ts

@@ -2,51 +2,10 @@
 // rendering. Extension handlers attach reasons here; human-facing copy
 // lives in bsk-cli `render_error.rs`.
 
-import type {
-  ErrorCode,
-  RpcError,
-  RpcErrorData,
-  RpcErrorReason,
-  TransferCleanupState,
-  TransferEffectState,
-} from "@/transport/types";
+import type { ErrorCode, RpcError, RpcErrorData, RpcErrorReason } from "@/transport/types";
 
 export type { RpcErrorData, RpcErrorReason };
 
-/** Chrome checks the whole tab's frame tree, including other extensions' frames. */
-export function isCdpExtensionAccessDenied(error: unknown): boolean {
-  const message = error instanceof Error ? error.message : String(error);
-  return message.includes("Cannot access a chrome-extension:// URL of different extension");
-}
-
-/** Preserve more specific reasons, especially file-transfer outcome/cleanup details. */
-export function classifyCdpError(error: RpcError): RpcError {
-  if (
-    error.code !== "cdp_failed" ||
-    error.data?.reason ||
-    !isCdpExtensionAccessDenied(error.message)
-  ) {
-    return error;
-  }
-  return {
-    ...error,
-    data: { ...error.data, reason: "cdp_extension_access_denied" },
-  };
-}
-
-export function cdpError(error: unknown): RpcError {
-  return classifyCdpError({
-    code: "cdp_failed",
-    message: error instanceof Error ? error.message : String(error),
-  });
-}
-
-export interface TransferErrorOptions {
-  effectState: TransferEffectState;
-  phase: string;
-  cleanupState?: TransferCleanupState;
-}
-
 export function rpcError(
   code: ErrorCode,
   reason: RpcErrorReason,
@@ -56,16 +15,3 @@ export function rpcError(
   const data: RpcErrorData = { reason, ...extra };
   return { code, message, data };
 }
-
-export function transferError(
-  code: ErrorCode,
-  reason: RpcErrorReason,
-  message: string,
-  options: TransferErrorOptions,
-): RpcError {
-  return rpcError(code, reason, message, {
-    effect_state: options.effectState,
-    phase: options.phase,
-    ...(options.cleanupState ? { cleanup_state: options.cleanupState } : {}),
-  });
-}

apps/extension/src/tools/file-drop-transaction.ts

@@ -1,319 +0,0 @@
-// One native file-drop transaction bound to an explicit page target. CDP drag
-// events are sent to the target that owns the node: root targets use top-level
-// coordinates, while OOPIF targets require coordinates local to their session.
-
-import type { RpcError, TransferEffectState } from "@/transport/types";
-import { transferError } from "./errors";
-import { resolveNodeGeometry } from "./frame-geometry";
-import type { ResolvedActionTarget } from "./interaction";
-import { type CdpRunner, isRpcError, sendToCdpTarget } from "./shared";
-import { BoundedWaitError, waitBounded } from "./transfer-transaction";
-
-const CLEANUP_TIMEOUT_MS = 1_000;
-
-type FileDropPhase = "resolve_target" | "drag_enter" | "drag_over" | "drop" | "cleanup";
-
-interface RuntimeReply {
-  result?: { value?: unknown; objectId?: string };
-  exceptionDetails?: { text?: string; exception?: { description?: string } };
-}
-
-interface DragData {
-  items: [];
-  files: string[];
-  dragOperationsMask: number;
-}
-
-export interface FileDropTransactionOptions {
-  cdp: CdpRunner;
-  actionTarget: ResolvedActionTarget;
-  files: string[];
-  timeoutMs: number;
-  signal?: AbortSignal;
-  withOverlayHidden: <T>(operation: () => Promise<T>) => Promise<T>;
-}
-
-export interface FileDropTransactionResult {
-  x: number;
-  y: number;
-}
-
-function failure(
-  code: RpcError["code"],
-  reason: "upload_mechanism_unsupported" | "file_drop_target_unavailable" | "file_drop_failed",
-  message: string,
-  effectState: TransferEffectState,
-  phase: FileDropPhase,
-): RpcError {
-  const error = transferError(code, reason, message, { effectState, phase });
-  error.data = { ...error.data, mechanism: "drop" };
-  return error;
-}
-
-function enrichFailure(
-  error: RpcError,
-  effectState: TransferEffectState,
-  phase: FileDropPhase,
-): RpcError {
-  return {
-    ...error,
-    data: {
-      ...error.data,
-      mechanism: "drop",
-      effect_state: effectState,
-      phase,
-    },
-  };
-}
-
-function messageOf(error: unknown): string {
-  return error instanceof Error ? error.message : String(error);
-}
-
-function methodUnavailable(error: unknown): boolean {
-  return /(?:method|command).*(?:not found|wasn't found|unsupported)|wasn't found|-32601/i.test(
-    messageOf(error),
-  );
-}
-
-function runtimeError(reply: RuntimeReply, fallback: string): Error | null {
-  if (!reply.exceptionDetails) return null;
-  return new Error(
-    reply.exceptionDetails.exception?.description ?? reply.exceptionDetails.text ?? fallback,
-  );
-}
-
-async function pointBelongsToTarget(
-  options: FileDropTransactionOptions,
-  x: number,
-  y: number,
-  objectGroup: string,
-): Promise<boolean> {
-  const target = options.actionTarget.cdpTarget;
-  const hit = await sendToCdpTarget<{ backendNodeId?: number; frameId?: string }>(
-    options.cdp,
-    target,
-    "DOM.getNodeForLocation",
-    { x: Math.round(x), y: Math.round(y), includeUserAgentShadowDOM: true },
-  );
-  if (typeof hit.backendNodeId !== "number") return false;
-  if (options.actionTarget.frameId && hit.frameId !== options.actionTarget.frameId) return false;
-
-  const [resolvedTarget, resolvedHit] = await Promise.all([
-    sendToCdpTarget<{ object?: { objectId?: string } }>(options.cdp, target, "DOM.resolveNode", {
-      backendNodeId: options.actionTarget.backendNodeId,
-      objectGroup,
-    }),
-    sendToCdpTarget<{ object?: { objectId?: string } }>(options.cdp, target, "DOM.resolveNode", {
-      backendNodeId: hit.backendNodeId,
-      objectGroup,
-    }),
-  ]);
-  const targetObjectId = resolvedTarget.object?.objectId;
-  const hitObjectId = resolvedHit.object?.objectId;
-  if (!targetObjectId || !hitObjectId) return false;
-
-  const reply = await sendToCdpTarget<RuntimeReply>(options.cdp, target, "Runtime.callFunctionOn", {
-    objectId: targetObjectId,
-    functionDeclaration: `function(hit) {
-        for (let node = hit; node; ) {
-          if (node === this) return true;
-          if (this.contains?.(node)) return true;
-          const root = node.getRootNode?.();
-          node = node.parentNode || (root && root !== node ? root.host : null);
-        }
-        return false;
-      }`,
-    arguments: [{ objectId: hitObjectId }],
-    returnByValue: true,
-  });
-  const error = runtimeError(reply, "failed to verify file-drop target");
-  if (error) throw error;
-  return reply.result?.value === true;
-}
-
-export async function uploadThroughFileDrop(
-  options: FileDropTransactionOptions,
-): Promise<FileDropTransactionResult | RpcError> {
-  const deadline = Date.now() + options.timeoutMs;
-  const target = options.actionTarget.cdpTarget;
-  const objectGroup = `bsk-file-drop-${crypto.randomUUID()}`;
-  const dragData: DragData = {
-    items: [],
-    files: options.files,
-    dragOperationsMask: 1,
-  };
-  let dragStarted = false;
-  let dropSent = false;
-  let dragPoint = { x: 0, y: 0 };
-  let cleanupFailed = false;
-  let outcome: FileDropTransactionResult | RpcError = failure(
-    "protocol_error",
-    "file_drop_failed",
-    "file-drop transaction ended without an outcome",
-    "none",
-    "cleanup",
-  );
-
-  try {
-    const geometry = await waitBounded(
-      resolveNodeGeometry(
-        options.cdp,
-        options.actionTarget.tab.tabId,
-        {
-          target,
-          backendNodeId: options.actionTarget.backendNodeId,
-          ...(options.actionTarget.frameId ? { frameId: options.actionTarget.frameId } : {}),
-        },
-        { scrollIntoView: true },
-      ),
-      deadline,
-      options.signal,
-      "resolving file-drop target timed out",
-    );
-    if (isRpcError(geometry)) {
-      outcome = enrichFailure(geometry, "none", "resolve_target");
-      return outcome;
-    }
-    const point = geometry.targetActionPoint;
-    dragPoint = point;
-
-    outcome = await options.withOverlayHidden(async () => {
-      try {
-        try {
-          const belongs = await waitBounded(
-            pointBelongsToTarget(options, point.x, point.y, objectGroup),
-            deadline,
-            options.signal,
-            "verifying file-drop target timed out",
-          );
-          if (!belongs) {
-            return failure(
-              "permission_denied",
-              "file_drop_target_unavailable",
-              "file-drop target is not the topmost element at its action point",
-              "none",
-              "resolve_target",
-            );
-          }
-        } catch (error) {
-          return failure(
-            error instanceof BoundedWaitError ? "timeout" : "cdp_failed",
-            "file_drop_target_unavailable",
-            messageOf(error),
-            "none",
-            "resolve_target",
-          );
-        }
-
-        for (const [type, phase] of [
-          ["dragEnter", "drag_enter"],
-          ["dragOver", "drag_over"],
-        ] as const) {
-          try {
-            if (type === "dragEnter") dragStarted = true;
-            await waitBounded(
-              sendToCdpTarget(options.cdp, target, "Input.dispatchDragEvent", {
-                type,
-                x: point.x,
-                y: point.y,
-                data: dragData,
-              }),
-              deadline,
-              options.signal,
-              `${type} timed out`,
-            );
-          } catch (error) {
-            const unsupported = methodUnavailable(error);
-            if (type === "dragEnter" && unsupported) dragStarted = false;
-            return failure(
-              unsupported
-                ? "unsupported"
-                : error instanceof BoundedWaitError
-                  ? "timeout"
-                  : "cdp_failed",
-              unsupported ? "upload_mechanism_unsupported" : "file_drop_failed",
-              messageOf(error),
-              "none",
-              phase,
-            );
-          }
-        }
-
-        try {
-          dropSent = true;
-          await waitBounded(
-            sendToCdpTarget(options.cdp, target, "Input.dispatchDragEvent", {
-              type: "drop",
-              x: point.x,
-              y: point.y,
-              data: dragData,
-            }),
-            deadline,
-            options.signal,
-            "file drop timed out",
-          );
-          return { x: point.x, y: point.y };
-        } catch (error) {
-          outcome = failure(
-            error instanceof BoundedWaitError ? "timeout" : "cdp_failed",
-            "file_drop_failed",
-            messageOf(error),
-            "unknown",
-            "drop",
-          );
-          return outcome;
-        }
-      } finally {
-        if (dragStarted && !dropSent) {
-          try {
-            await waitBounded(
-              sendToCdpTarget(options.cdp, target, "Input.dispatchDragEvent", {
-                type: "dragCancel",
-                x: dragPoint.x,
-                y: dragPoint.y,
-                data: dragData,
-              }),
-              Date.now() + CLEANUP_TIMEOUT_MS,
-              undefined,
-              "cancelling file drag timed out",
-            );
-          } catch {
-            cleanupFailed = true;
-          }
-        }
-      }
-    });
-    return outcome;
-  } catch (error) {
-    outcome = failure(
-      error instanceof BoundedWaitError ? "timeout" : "cdp_failed",
-      "file_drop_target_unavailable",
-      messageOf(error),
-      "none",
-      "resolve_target",
-    );
-    return outcome;
-  } finally {
-    try {
-      await waitBounded(
-        sendToCdpTarget(options.cdp, target, "Runtime.releaseObjectGroup", { objectGroup }),
-        Date.now() + CLEANUP_TIMEOUT_MS,
-        undefined,
-        "releasing file-drop object group timed out",
-      );
-    } catch {
-      cleanupFailed = true;
-    }
-
-    const effect = isRpcError(outcome)
-      ? (outcome.data?.effect_state as TransferEffectState | undefined)
-      : "committed";
-    if (effect === "unknown" || cleanupFailed) {
-      await options.cdp.detach?.(target.tabId);
-    }
-    if (cleanupFailed && isRpcError(outcome)) {
-      outcome.data = { ...outcome.data, cleanup_state: "failed" };
-    }
-  }
-}

apps/extension/src/tools/file-input-transaction.ts

@@ -1,445 +0,0 @@
-// One upload transaction with an explicit browser-side commit boundary.
-// Chrome interception prevents a native chooser from escaping automation;
-// chooser events and a frame-scoped DOM probe are independent input-location
-// signals, so event delivery is not required for standard file inputs.
-
-import type { ClickResult, RpcError, TransferEffectState } from "@/transport/types";
-import { transferError } from "./errors";
-import type { ResolvedActionTarget } from "./interaction";
-import { type CdpRunner, isRpcError, sendToCdpTarget } from "./shared";
-import { BoundedWaitError, remainingMs, waitBounded } from "./transfer-transaction";
-
-const CLEANUP_TIMEOUT_MS = 1_000;
-const ACTIVATION_GRACE_MS = 1_000;
-const PROBE_INTERVAL_MS = 20;
-
-type UploadPhase =
-  | "arm_interception"
-  | "arm_input_probe"
-  | "trigger"
-  | "resolve_input"
-  | "set_files"
-  | "cleanup";
-
-interface RuntimeReply {
-  result?: { value?: unknown; objectId?: string };
-  exceptionDetails?: { text?: string; exception?: { description?: string } };
-}
-
-interface ChooserEvent {
-  frameId?: string;
-  backendNodeId?: number;
-  mode?: "selectSingle" | "selectMultiple";
-}
-
-export interface FileInputTransactionOptions {
-  cdp: CdpRunner;
-  actionTarget: ResolvedActionTarget;
-  files: string[];
-  timeoutMs: number;
-  signal?: AbortSignal;
-  trigger(): Promise<ClickResult | RpcError>;
-}
-
-export interface FileInputTransactionResult {
-  click: ClickResult;
-  multiple: boolean;
-}
-
-function runtimeError(reply: RuntimeReply, fallback: string): Error | null {
-  if (!reply.exceptionDetails) return null;
-  return new Error(
-    reply.exceptionDetails.exception?.description ?? reply.exceptionDetails.text ?? fallback,
-  );
-}
-
-function transferFailure(
-  code: RpcError["code"],
-  reason: "file_input_probe_failed" | "file_input_not_activated" | "set_file_input_failed",
-  message: string,
-  effectState: TransferEffectState,
-  phase: UploadPhase,
-): RpcError {
-  return transferError(code, reason, message, { effectState, phase });
-}
-
-function enrichFailure(
-  error: RpcError,
-  effectState: TransferEffectState,
-  phase: UploadPhase,
-): RpcError {
-  return {
-    ...error,
-    data: { ...error.data, effect_state: effectState, phase },
-  };
-}
-
-function sameTarget(
-  source: { tabId?: number; sessionId?: string },
-  target: { tabId: number; sessionId?: string },
-): boolean {
-  return source.tabId === target.tabId && source.sessionId === target.sessionId;
-}
-
-async function delay(ms: number, signal?: AbortSignal): Promise<void> {
-  await waitBounded(
-    new Promise((resolve) => setTimeout(resolve, ms)),
-    Date.now() + ms + 1,
-    signal,
-    "upload activation probe timed out",
-  );
-}
-
-async function callOnTrigger<T = RuntimeReply>(
-  options: FileInputTransactionOptions,
-  objectId: string,
-  functionDeclaration: string,
-  args: unknown[],
-  returnByValue: boolean,
-): Promise<T> {
-  return sendToCdpTarget<T>(options.cdp, options.actionTarget.cdpTarget, "Runtime.callFunctionOn", {
-    objectId,
-    functionDeclaration,
-    arguments: args.map((value) => ({ value })),
-    returnByValue,
-    awaitPromise: false,
-  });
-}
-
-export async function uploadThroughActivatedFileInput(
-  options: FileInputTransactionOptions,
-): Promise<FileInputTransactionResult | RpcError> {
-  const deadline = Date.now() + options.timeoutMs;
-  const target = options.actionTarget.cdpTarget;
-  const objectGroup = `bsk-upload-${crypto.randomUUID()}`;
-  const stateKey = `__bskUpload_${crypto.randomUUID().replaceAll("-", "")}`;
-  const chooserEvents: ChooserEvent[] = [];
-  let interceptionArmed = false;
-  let probeArmed = false;
-  let triggerObjectId: string | undefined;
-  let outcome: FileInputTransactionResult | RpcError = transferFailure(
-    "protocol_error",
-    "file_input_probe_failed",
-    "upload transaction ended without an outcome",
-    "none",
-    "cleanup",
-  );
-
-  const chooserSubscription = options.cdp.onEvent?.((source, method, raw) => {
-    if (method !== "Page.fileChooserOpened" || !sameTarget(source, target)) return;
-    const event = raw as ChooserEvent;
-    chooserEvents.push(event);
-  });
-
-  try {
-    try {
-      await waitBounded(
-        sendToCdpTarget(options.cdp, target, "Page.setInterceptFileChooserDialog", {
-          enabled: true,
-        }),
-        deadline,
-        options.signal,
-        "arming native file chooser interception timed out",
-      );
-      interceptionArmed = true;
-    } catch (err) {
-      outcome = transferFailure(
-        err instanceof BoundedWaitError ? "timeout" : "cdp_failed",
-        "file_input_probe_failed",
-        err instanceof Error ? err.message : String(err),
-        "none",
-        "arm_interception",
-      );
-      return outcome;
-    }
-
-    try {
-      const resolved = await waitBounded(
-        sendToCdpTarget<{ object?: { objectId?: string } }>(
-          options.cdp,
-          target,
-          "DOM.resolveNode",
-          {
-            backendNodeId: options.actionTarget.backendNodeId,
-            objectGroup,
-          },
-        ),
-        deadline,
-        options.signal,
-        "resolving upload trigger timed out",
-      );
-      triggerObjectId = resolved.object?.objectId;
-      if (!triggerObjectId) throw new Error("DOM.resolveNode returned no trigger objectId");
-
-      const armed = await waitBounded(
-        callOnTrigger<RuntimeReply>(
-          options,
-          triggerObjectId,
-          `function(key) {
-            const doc = this.ownerDocument;
-            const owner = doc.defaultView;
-            if (!owner) return false;
-            const state = { inputs: [], listener: null };
-            Object.defineProperty(owner, key, { value: state, configurable: true });
-            state.listener = event => {
-              const path = typeof event.composedPath === "function" ? event.composedPath() : [];
-              const candidate = path[0] || event.target;
-              if (candidate && candidate.nodeType === 1 &&
-                  candidate.localName === "input" && candidate.type === "file") {
-                if (!state.inputs.includes(candidate)) state.inputs.push(candidate);
-                event.preventDefault();
-              }
-            };
-            doc.addEventListener("click", state.listener, true);
-            return true;
-          }`,
-          [stateKey],
-          true,
-        ),
-        deadline,
-        options.signal,
-        "arming frame-scoped file input probe timed out",
-      );
-      const armError = runtimeError(armed, "failed to arm file input probe");
-      if (armError || armed.result?.value !== true) {
-        throw armError ?? new Error("file input probe did not arm");
-      }
-      probeArmed = true;
-    } catch (err) {
-      outcome = transferFailure(
-        err instanceof BoundedWaitError ? "timeout" : "cdp_failed",
-        "file_input_probe_failed",
-        err instanceof Error ? err.message : String(err),
-        "none",
-        "arm_input_probe",
-      );
-      return outcome;
-    }
-
-    let click: ClickResult | RpcError;
-    try {
-      click = await waitBounded(
-        options.trigger(),
-        deadline,
-        options.signal,
-        "upload trigger timed out",
-      );
-    } catch (err) {
-      outcome = transferFailure(
-        err instanceof BoundedWaitError ? "timeout" : "cdp_failed",
-        "file_input_probe_failed",
-        err instanceof Error ? err.message : String(err),
-        "none",
-        "trigger",
-      );
-      return outcome;
-    }
-    if (isRpcError(click)) {
-      outcome = enrichFailure(click, "none", "trigger");
-      return outcome;
-    }
-
-    try {
-      const activationDeadline = Math.min(deadline, Date.now() + ACTIVATION_GRACE_MS);
-      let summary: { count: number; multiple: boolean } = { count: 0, multiple: false };
-      while (Date.now() < activationDeadline) {
-        if (chooserEvents.length > 0) break;
-        const reply = await callOnTrigger<RuntimeReply>(
-          options,
-          triggerObjectId,
-          `function(key) {
-            const state = this.ownerDocument.defaultView?.[key];
-            return state
-              ? { count: state.inputs.length, multiple: state.inputs[0]?.multiple === true }
-              : { count: 0, multiple: false };
-          }`,
-          [stateKey],
-          true,
-        );
-        const summaryError = runtimeError(reply, "failed to inspect activated file input");
-        if (summaryError) throw summaryError;
-        const value = reply.result?.value as { count?: unknown; multiple?: unknown } | undefined;
-        summary = {
-          count: typeof value?.count === "number" ? value.count : 0,
-          multiple: value?.multiple === true,
-        };
-        if (summary.count > 0) break;
-        await delay(Math.min(PROBE_INTERVAL_MS, remainingMs(activationDeadline)), options.signal);
-      }
-
-      if (chooserEvents.length > 1) {
-        outcome = transferFailure(
-          "unsupported",
-          "file_input_not_activated",
-          "upload trigger activated more than one file chooser",
-          "none",
-          "resolve_input",
-        );
-        return outcome;
-      }
-
-      const chooser = chooserEvents[0];
-      let backendNodeId: number | undefined;
-      let multiple = summary.multiple;
-      if (chooser) {
-        if (options.actionTarget.frameId && chooser.frameId !== options.actionTarget.frameId) {
-          outcome = transferFailure(
-            "unsupported",
-            "file_input_not_activated",
-            "upload trigger activated a file chooser in a different frame",
-            "none",
-            "resolve_input",
-          );
-          return outcome;
-        }
-        if (typeof chooser.backendNodeId !== "number") {
-          outcome = transferFailure(
-            "unsupported",
-            "file_input_not_activated",
-            "upload trigger invoked a non-input file picker",
-            "none",
-            "resolve_input",
-          );
-          return outcome;
-        }
-        backendNodeId = chooser.backendNodeId;
-        multiple = chooser.mode === "selectMultiple";
-      } else {
-        if (summary.count !== 1) {
-          outcome = transferFailure(
-            "unsupported",
-            "file_input_not_activated",
-            summary.count === 0
-              ? "upload trigger did not activate an input[type=file]"
-              : "upload trigger activated more than one input[type=file]",
-            "none",
-            "resolve_input",
-          );
-          return outcome;
-        }
-        const input = await callOnTrigger<RuntimeReply>(
-          options,
-          triggerObjectId,
-          `function(key) { return this.ownerDocument.defaultView?.[key]?.inputs[0]; }`,
-          [stateKey],
-          false,
-        );
-        const inputError = runtimeError(input, "failed to resolve activated file input object");
-        if (inputError) throw inputError;
-        if (!input.result?.objectId) throw new Error("activated file input returned no objectId");
-        const described = await sendToCdpTarget<{ node?: { backendNodeId?: number } }>(
-          options.cdp,
-          target,
-          "DOM.describeNode",
-          { objectId: input.result.objectId },
-        );
-        backendNodeId = described.node?.backendNodeId;
-        if (typeof backendNodeId !== "number") {
-          throw new Error("DOM.describeNode returned no file input backendNodeId");
-        }
-      }
-
-      if (!multiple && options.files.length !== 1) {
-        outcome = enrichFailure(
-          { code: "invalid_params", message: "file input accepts exactly one file" },
-          "none",
-          "resolve_input",
-        );
-        return outcome;
-      }
-
-      try {
-        await waitBounded(
-          sendToCdpTarget(options.cdp, target, "DOM.setFileInputFiles", {
-            files: options.files,
-            backendNodeId,
-          }),
-          deadline,
-          options.signal,
-          "setting file input files timed out",
-        );
-      } catch (err) {
-        outcome = transferFailure(
-          err instanceof BoundedWaitError ? "timeout" : "cdp_failed",
-          "set_file_input_failed",
-          err instanceof Error ? err.message : String(err),
-          "unknown",
-          "set_files",
-        );
-        return outcome;
-      }
-      outcome = { click, multiple };
-      return outcome;
-    } catch (err) {
-      outcome = transferFailure(
-        err instanceof BoundedWaitError ? "timeout" : "cdp_failed",
-        "file_input_probe_failed",
-        err instanceof Error ? err.message : String(err),
-        "none",
-        "resolve_input",
-      );
-      return outcome;
-    }
-  } finally {
-    chooserSubscription?.dispose();
-    let cleanupFailed = false;
-    if (probeArmed && triggerObjectId) {
-      try {
-        await waitBounded(
-          callOnTrigger(
-            options,
-            triggerObjectId,
-            `function(key) {
-              const owner = this.ownerDocument.defaultView;
-              const state = owner?.[key];
-              if (state?.listener) this.ownerDocument.removeEventListener("click", state.listener, true);
-              if (owner) delete owner[key];
-            }`,
-            [stateKey],
-            true,
-          ),
-          Date.now() + CLEANUP_TIMEOUT_MS,
-          undefined,
-          "cleaning file input probe timed out",
-        );
-      } catch {
-        cleanupFailed = true;
-      }
-    }
-    if (interceptionArmed) {
-      try {
-        await waitBounded(
-          sendToCdpTarget(options.cdp, target, "Page.setInterceptFileChooserDialog", {
-            enabled: false,
-            cancel: true,
-          }),
-          Date.now() + CLEANUP_TIMEOUT_MS,
-          undefined,
-          "disabling file chooser interception timed out",
-        );
-      } catch {
-        cleanupFailed = true;
-      }
-    }
-    try {
-      await waitBounded(
-        sendToCdpTarget(options.cdp, target, "Runtime.releaseObjectGroup", { objectGroup }),
-        Date.now() + CLEANUP_TIMEOUT_MS,
-        undefined,
-        "releasing upload object group timed out",
-      );
-    } catch {
-      cleanupFailed = true;
-    }
-
-    const effect = isRpcError(outcome)
-      ? (outcome.data?.effect_state as TransferEffectState | undefined)
-      : "committed";
-    if (effect === "unknown" || cleanupFailed) {
-      await options.cdp.detach?.(target.tabId);
-    }
-    if (cleanupFailed && isRpcError(outcome)) {
-      outcome.data = { ...outcome.data, cleanup_state: "failed" };
-    }
-  }
-}

apps/extension/src/tools/frame-geometry.ts

@@ -29,10 +29,7 @@ export interface NodeAddress {
 export interface ResolvedNodeGeometry {
   topVisibleRegions: Region;
   topBounds: ViewportRect;
-  /** Point in the top-level tab viewport, used by root-target input events. */
   actionPoint: Point;
-  /** Point in the addressed CDP target's viewport, used by OOPIF-local input events. */
-  targetActionPoint: Point;
 }
 
 function geometryError(message: string): RpcError {
@@ -265,30 +262,11 @@ export async function resolveNodeGeometry(
     if (isRpcError(localRegion)) return localRegion;
 
     let topVisibleRegions: Region;
-    let targetActionPoint: Point | null = null;
     if (address.frameId && graph) {
       const projection = await resolveFrameProjection(cdp, graph, address.frameId);
       if (!projection)
         return geometryError(`could not resolve frame geometry for ${address.frameId}`);
       topVisibleRegions = projectRegionToViewport(localRegion, projection);
-      if (address.target.sessionId) {
-        const localViewport = projection.edges[0]?.sourceViewport ?? projection.topViewport;
-        const localVisibleRegions = localRegion
-          .map((polygon) =>
-            clipPolygon(
-              polygon,
-              rectPolygon({
-                x: 0,
-                y: 0,
-                w: localViewport.width,
-                h: localViewport.height,
-              }),
-            ),
-          )
-          .filter((polygon) => polygon.length >= 3);
-        const localActionRegion = largestRegion(localVisibleRegions);
-        targetActionPoint = localActionRegion ? polygonCentroid(localActionRegion) : null;
-      }
     } else {
       const viewport = await targetViewport(cdp, address.target);
       if (!viewport) return geometryError("could not resolve top viewport geometry");
@@ -305,11 +283,7 @@ export async function resolveNodeGeometry(
     if (!topBounds || !actionPoint) {
       return { code: "permission_denied", message: "element not visible" };
     }
-    if (address.target.sessionId && !targetActionPoint) {
-      return { code: "permission_denied", message: "element not visible in its target" };
-    }
-    targetActionPoint ??= actionPoint;
-    return { topVisibleRegions, topBounds, actionPoint, targetActionPoint };
+    return { topVisibleRegions, topBounds, actionPoint };
   } catch (error) {
     return geometryError(error instanceof Error ? error.message : String(error));
   }

apps/extension/src/tools/interaction.ts

@@ -1,5 +1,5 @@
-// DOM interaction tools — `tool.click`, `tool.fill`, `tool.press`, and
-// `tool.select`.
+// DOM interaction tools — `tool.click`, `tool.fill`, `tool.type`,
+// `tool.press`, and `tool.select`.
 //
 // All interaction tools:
 // 1. Resolve target tab (sandbox: must be inside Agent Window).
@@ -11,7 +11,7 @@
 //    commands.
 
 import { ChromiumCdp } from "@/browser-driver/chromium-cdp";
-import type { CdpTarget } from "@/browser-driver/frame-graph";
+import { type CdpTarget, cdpTargetKey } from "@/browser-driver/frame-graph";
 import type { SessionContext, SessionManager } from "@/session-manager/manager";
 import type {
   ClickParams,
@@ -27,11 +27,15 @@ import type {
   RpcError,
   SelectParams,
   SelectResult,
+  TypeParams,
+  TypeResult,
 } from "@/transport/types";
 import { attachDialogs, markDialogCursor } from "./dialogs";
+import { fillEditable } from "./editable-input";
 import { backendNodeToObject } from "./element-geometry";
 import { rpcError } from "./errors";
 import { resolveNodeGeometry, scrollElementAndFramesIntoView } from "./frame-geometry";
+import { dispatchMouseClick } from "./mouse-input";
 import {
   type CdpRunner,
   type ChromeTabsApi,
@@ -40,10 +44,11 @@ import {
   enforceAgentWindow,
   isRpcError,
   lookupSession,
-  type ResolvedTargetTab,
   resolveTargetTab,
 } from "./shared";
 import { resolveSnapshotRef } from "./snapshot-ref";
+import { resolveSurfacePointerTarget } from "./surface-target";
+import { dispatchTextInput } from "./text-input";
 
 export interface InteractionDeps {
   cdp: CdpRunner;
@@ -57,15 +62,6 @@ export interface InteractionDeps {
   keepOverlayBypassAfterHover?: boolean;
 }
 
-export interface ResolvedActionTarget {
-  tab: ResolvedTargetTab;
-  backendNodeId: number;
-  cdpTarget: CdpTarget;
-  frameId?: string;
-  usedRef?: string;
-  usedSelector?: string;
-}
-
 const DEFAULT_TIMEOUT_MS = 30_000;
 const DEFAULT_HOVER_SETTLE_MS = 200;
 
@@ -138,7 +134,7 @@ async function wait(ms: number, signal?: AbortSignal): Promise<void> {
  * `RpcError` if the caller supplied neither (or both), or if neither
  * lookup matched.
  */
-export async function resolveBackendNode(
+async function resolveBackendNode(
   cdp: CdpRunner,
   ctx: SessionContext,
   target: { tabId: number },
@@ -169,7 +165,7 @@ export async function resolveBackendNode(
     };
   }
   if (hasRef) {
-    const resolved = resolveSnapshotRef(ctx, params.ref as string, target.tabId);
+    const resolved = resolveSnapshotRef(ctx, params.ref as string, target.tabId, "interact");
     if (isRpcError(resolved)) return resolved;
     return {
       backendNodeId: resolved.backendNodeId,
@@ -183,8 +179,6 @@ export async function resolveBackendNode(
   }
   // selector path
   try {
-    // Selector lookup itself attaches CDP, even when no element is found.
-    cdp.trackSessionTab?.(ctx.sessionId, target.tabId);
     const doc = await cdp.send<{ root?: { nodeId?: number } }>(target.tabId, "DOM.getDocument", {
       depth: 0,
     });
@@ -227,26 +221,146 @@ export async function resolveBackendNode(
   }
 }
 
-export async function resolveActionTarget(
+interface InputDispatchContext {
+  target: CdpTarget;
+  cdp: CdpRunner;
+  backendNodeId?: number;
+  frameId?: string;
+  usedRef?: string;
+  usedSelector?: string;
+}
+
+function inputContextForNode(
   cdp: CdpRunner,
-  ctx: SessionContext,
-  target: ResolvedTargetTab,
-  params: { ref?: string; selector?: string },
-  toolName: string,
-): Promise<ResolvedActionTarget | RpcError> {
-  const node = await resolveBackendNode(cdp, ctx, target, params, toolName);
-  return isRpcError(node) ? node : { tab: target, ...node };
+  node: Awaited<ReturnType<typeof resolveBackendNode>>,
+): (InputDispatchContext & { backendNodeId: number }) | RpcError {
+  if (isRpcError(node)) return node;
+  return {
+    target: node.cdpTarget,
+    cdp: cdpRunnerForTarget(cdp, node.cdpTarget),
+    backendNodeId: node.backendNodeId,
+    frameId: node.frameId,
+    usedRef: node.usedRef,
+    usedSelector: node.usedSelector,
+  };
+}
+
+interface FocusProbe {
+  hasFocus: boolean;
+  activeTag: string;
+  textEntry: boolean;
+}
+
+/** Resolve keyboard input to the CDP target that currently owns focus. */
+async function resolveFocusedInputContext(
+  cdp: CdpRunner,
+  tabId: number,
+  mode: "press" | "type",
+): Promise<InputDispatchContext | RpcError> {
+  if (!cdp.getFrameGraph) {
+    return rpcError(
+      "cdp_failed",
+      "focus_target_unresolved",
+      `${mode} cannot determine the focused frame target`,
+    );
+  }
+
+  let targets: CdpTarget[];
+  try {
+    const graph = await cdp.getFrameGraph(tabId);
+    const unique = new Map<string, CdpTarget>();
+    unique.set(cdpTargetKey({ tabId }), { tabId });
+    for (const frame of graph.frames) unique.set(cdpTargetKey(frame.target), frame.target);
+    targets = [...unique.values()];
+  } catch (err) {
+    return rpcError(
+      "cdp_failed",
+      "focus_target_unresolved",
+      err instanceof Error ? err.message : String(err),
+    );
+  }
+
+  const focused: Array<{ target: CdpTarget; probe: FocusProbe }> = [];
+  await Promise.all(
+    targets.map(async (candidate) => {
+      try {
+        const reply = await cdpRunnerForTarget(cdp, candidate).send<{
+          result?: { value?: FocusProbe };
+        }>(tabId, "Runtime.evaluate", {
+          expression: `(() => {
+            /* bsk-focus-target-probe */
+            const active = document.activeElement;
+            const inputType = active instanceof HTMLInputElement ? active.type.toLowerCase() : '';
+            const nonTextInputTypes = new Set([
+              'button', 'checkbox', 'color', 'date', 'datetime-local', 'file',
+              'hidden', 'image', 'month', 'radio', 'range', 'reset', 'submit',
+              'time', 'week'
+            ]);
+            const role = active instanceof Element ? active.getAttribute('role') : null;
+            return {
+              hasFocus: document.hasFocus(),
+              activeTag: active instanceof Element ? active.tagName.toLowerCase() : '',
+              textEntry:
+                (active instanceof HTMLInputElement && !nonTextInputTypes.has(inputType)) ||
+                active instanceof HTMLTextAreaElement ||
+                (active instanceof HTMLElement && active.isContentEditable) ||
+                role === 'textbox' || role === 'searchbox'
+            };
+          })()`,
+          returnByValue: true,
+        });
+        const probe = reply.result?.value;
+        if (probe?.hasFocus) focused.push({ target: candidate, probe });
+      } catch {
+        // A target may navigate or detach while the graph is being probed.
+      }
+    }),
+  );
+
+  // A top document whose active element is an iframe delegates focus to
+  // that frame. Removing those delegating targets leaves the deepest OOPIF
+  // target while same-process frames continue to use the shared root target.
+  const owners = focused.filter(({ probe }) => {
+    if (probe.activeTag === "iframe" || probe.activeTag === "frame") return false;
+    return mode !== "type" || probe.textEntry;
+  });
+  if (owners.length !== 1) {
+    return rpcError(
+      "cdp_failed",
+      "focus_target_unresolved",
+      owners.length === 0
+        ? `${mode} found no unique focused${mode === "type" ? " text-entry" : " frame"} target`
+        : `${mode} found multiple focused frame targets`,
+      { candidate_count: owners.length },
+    );
+  }
+  const target = owners[0].target;
+  return { target, cdp: cdpRunnerForTarget(cdp, target) };
 }
 
 // ---------------------------------------------------------------------------
 // tool.click
 // ---------------------------------------------------------------------------
 
+interface ResolvedClickPointerTarget {
+  point: { x: number; y: number };
+  usedRef?: string;
+  usedSelector?: string;
+  moveSettleMs: number;
+  capture?: {
+    id: string;
+    imageX: number;
+    imageY: number;
+  };
+}
+
 export async function handleClick(
   manager: SessionManager,
   params: ClickParams,
   deps: InteractionDeps = getDefaultDeps(),
 ): Promise<ClickResult | RpcError> {
+  const hasSurfacePointParams =
+    params.capture_id !== undefined || params.image_x !== undefined || params.image_y !== undefined;
   const ctxOrErr = lookupSession(manager, params, "click");
   if (isRpcError(ctxOrErr)) return ctxOrErr;
   const ctx = ctxOrErr;
@@ -256,50 +370,69 @@ export async function handleClick(
   if (isRpcError(target)) return target;
   const denied = enforceAgentWindow(ctx, target, "click");
   if (denied) return denied;
-  const resolved = await resolveActionTarget(deps.cdp, ctx, target, params, "click");
-  if (isRpcError(resolved)) return resolved;
-  return clickResolvedTarget(ctx, resolved, params, deps);
-}
-
-export async function clickResolvedTarget(
-  ctx: SessionContext,
-  resolved: ResolvedActionTarget,
-  params: Pick<ClickParams, "button" | "click_count" | "modifiers">,
-  deps: InteractionDeps,
-): Promise<ClickResult | RpcError> {
-  const { tab: target } = resolved;
   const dialogCursor = markDialogCursor(deps.cdp, target.tabId);
 
+  const button: MouseButton = params.button ?? "left";
+  const clickCount = params.click_count ?? 1;
+  if (!Number.isSafeInteger(clickCount) || clickCount < 1) {
+    return { code: "invalid_params", message: "click_count must be a positive integer" };
+  }
+  const modifiers = modifiersBitfield(params.modifiers);
+
   if (throwIfAborted(deps.signal)) {
     return { code: "cancelled", message: "click aborted" };
   }
 
   deps.cdp.trackSessionTab?.(ctx.sessionId, target.tabId);
-  const geometry = await resolveNodeGeometry(
-    deps.cdp,
-    target.tabId,
-    {
-      target: resolved.cdpTarget,
-      backendNodeId: resolved.backendNodeId,
-      ...(resolved.frameId ? { frameId: resolved.frameId } : {}),
-    },
-    { scrollIntoView: true },
-  );
-  if (isRpcError(geometry)) return geometry;
-  const centre = geometry.actionPoint;
+  const pointerTarget: ResolvedClickPointerTarget | RpcError = hasSurfacePointParams
+    ? await resolveSurfacePointerTarget(
+        ctx,
+        target,
+        {
+          ref: params.ref,
+          selector: params.selector,
+          captureId: params.capture_id,
+          imageX: params.image_x,
+          imageY: params.image_y,
+        },
+        {
+          cdp: deps.cdp,
+          signal: deps.signal,
+        },
+      )
+    : await (async () => {
+        const node = await resolveBackendNode(deps.cdp, ctx, target, params, "click");
+        if (isRpcError(node)) return node;
+        const geometry = await resolveNodeGeometry(
+          deps.cdp,
+          target.tabId,
+          {
+            target: node.cdpTarget,
+            backendNodeId: node.backendNodeId,
+            ...(node.frameId ? { frameId: node.frameId } : {}),
+          },
+          { scrollIntoView: true },
+        );
+        if (isRpcError(geometry)) return geometry;
+        return {
+          point: geometry.actionPoint,
+          usedRef: node.usedRef,
+          usedSelector: node.usedSelector,
+          moveSettleMs: 0,
+        };
+      })();
+  if (isRpcError(pointerTarget)) return pointerTarget;
 
   if (throwIfAborted(deps.signal)) {
     return { code: "cancelled", message: "click aborted" };
   }
 
-  const button: MouseButton = params.button ?? "left";
-  const clickCount = params.click_count ?? 1;
-  if (clickCount < 1) {
-    return { code: "invalid_params", message: "click_count must be greater than zero" };
-  }
-  const modifiers = modifiersBitfield(params.modifiers);
-
-  const overlayBlocking = await checkOverlayAtPoint(deps.cdp, target.tabId, centre.x, centre.y);
+  const overlayBlocking = await checkOverlayAtPoint(
+    deps.cdp,
+    target.tabId,
+    pointerTarget.point.x,
+    pointerTarget.point.y,
+  );
   let automationBypassEnabled = false;
   if (overlayBlocking && deps.bypassOverlay) {
     try {
@@ -311,47 +444,16 @@ export async function clickResolvedTarget(
   }
 
   try {
-    // Move first so hover state activates, then press → release.
-    await deps.cdp.send(target.tabId, "Input.dispatchMouseEvent", {
-      type: "mouseMoved",
-      x: centre.x,
-      y: centre.y,
-      modifiers,
-    });
-    if (throwIfAborted(deps.signal)) {
-      return { code: "cancelled", message: "click aborted" };
-    }
-    await deps.cdp.send(target.tabId, "Input.dispatchMouseEvent", {
-      type: "mousePressed",
-      x: centre.x,
-      y: centre.y,
+    const dispatch = await dispatchMouseClick(deps.cdp, target.tabId, pointerTarget.point, {
       button,
       clickCount,
       modifiers,
+      signal: deps.signal,
+      moveSettleMs: pointerTarget.moveSettleMs,
     });
-    if (throwIfAborted(deps.signal)) {
-      try {
-        await deps.cdp.send(target.tabId, "Input.dispatchMouseEvent", {
-          type: "mouseReleased",
-          x: centre.x,
-          y: centre.y,
-          button,
-          clickCount,
-          modifiers,
-        });
-      } catch (err) {
-        console.debug("[bsk interaction] best-effort mouseReleased after abort failed", err);
-      }
+    if (dispatch === "cancelled") {
       return { code: "cancelled", message: "click aborted" };
     }
-    await deps.cdp.send(target.tabId, "Input.dispatchMouseEvent", {
-      type: "mouseReleased",
-      x: centre.x,
-      y: centre.y,
-      button,
-      clickCount,
-      modifiers,
-    });
   } catch (err) {
     return {
       code: "cdp_failed",
@@ -369,10 +471,17 @@ export async function clickResolvedTarget(
 
   return attachDialogs(deps.cdp, target.tabId, dialogCursor, {
     tab_id: target.tabId,
-    used_ref: resolved.usedRef,
-    used_selector: resolved.usedSelector,
-    x: centre.x,
-    y: centre.y,
+    used_ref: pointerTarget.usedRef,
+    used_selector: pointerTarget.usedSelector,
+    x: pointerTarget.point.x,
+    y: pointerTarget.point.y,
+    ...(pointerTarget.capture
+      ? {
+          capture_id: pointerTarget.capture.id,
+          image_x: pointerTarget.capture.imageX,
+          image_y: pointerTarget.capture.imageY,
+        }
+      : {}),
   });
 }
 
@@ -592,6 +701,47 @@ function isFillable(node: DescribedNode): boolean {
   return false;
 }
 
+/** `type` also supports custom text-entry widgets that expose an ARIA role. */
+function isTypeTarget(node: DescribedNode): boolean {
+  const tag = (node.nodeName ?? "").toUpperCase();
+  const attrs = node.attributes ?? [];
+  const attributes = new Map<string, string>();
+  for (let i = 0; i + 1 < attrs.length; i += 2) {
+    attributes.set(attrs[i].toLowerCase(), attrs[i + 1].toLowerCase());
+  }
+  if (tag === "TEXTAREA") return true;
+  if (tag === "INPUT") {
+    const type = attributes.get("type") ?? "text";
+    return !new Set([
+      "button",
+      "checkbox",
+      "color",
+      "date",
+      "datetime-local",
+      "file",
+      "hidden",
+      "image",
+      "month",
+      "radio",
+      "range",
+      "reset",
+      "submit",
+      "time",
+      "week",
+    ]).has(type);
+  }
+  const contentEditable = attributes.get("contenteditable");
+  if (
+    contentEditable === "" ||
+    contentEditable === "true" ||
+    contentEditable === "plaintext-only"
+  ) {
+    return true;
+  }
+  const role = attributes.get("role");
+  return role === "textbox" || role === "searchbox";
+}
+
 type SelectMutationResult =
   | {
       ok: true;
@@ -609,53 +759,6 @@ type SelectMutationResult =
 // tool.fill
 // ---------------------------------------------------------------------------
 
-type FillFailureReason =
-  | "target_not_fillable"
-  | "fill_value_invalid"
-  | "fill_target_changed"
-  | "fill_focus_lost"
-  | "fill_value_mismatch"
-  | "fill_failed";
-
-type FillPreparation =
-  | { before: string; expected: string }
-  | { reason: FillFailureReason; error: string };
-
-type FillReadiness = "ready" | "background" | "fill_target_changed" | "fill_focus_lost";
-
-function fillError(reason: FillFailureReason, message: string): RpcError {
-  return rpcError(
-    reason === "target_not_fillable" || reason === "fill_value_invalid"
-      ? "invalid_params"
-      : "cdp_failed",
-    reason,
-    message,
-  );
-}
-
-interface FillScriptReply<T> {
-  result?: { value?: T };
-  exceptionDetails?: unknown;
-}
-
-const FILL_EDITABLE_FUNCTION = `function() {
-  const tag = this.tagName.toLowerCase();
-  const supported = tag === 'input'
-    ? ['text', 'search', 'tel', 'url', 'email', 'password', 'number'].includes(this.type)
-    : tag === 'textarea' || this.isContentEditable;
-  return this.isConnected && supported && !this.readOnly && !this.matches(':disabled');
-}`;
-
-// Chrome renders a trailing editable newline with an empty <div><br></div>.
-// innerText includes the padding break; it is not an extra typed character.
-const FILL_VALUE_FUNCTION = `function() {
-  if (!this.isContentEditable) return this.value;
-  const value = this.innerText;
-  const tail = this.lastChild;
-  const padding = tail && tail.nodeName === 'DIV' && tail.childNodes.length === 1 && tail.firstChild.nodeName === 'BR';
-  return padding && value.endsWith('\\n\\n') ? value.slice(0, -1) : value;
-}`;
-
 export async function handleFill(
   manager: SessionManager,
   params: FillParams,
@@ -677,8 +780,9 @@ export async function handleFill(
   const dialogCursor = markDialogCursor(deps.cdp, target.tabId);
 
   const node = await resolveBackendNode(deps.cdp, ctx, target, params, "fill");
-  if (isRpcError(node)) return node;
-  const nodeCdp = cdpRunnerForTarget(deps.cdp, node.cdpTarget);
+  const input = inputContextForNode(deps.cdp, node);
+  if (isRpcError(input)) return input;
+  const nodeCdp = input.cdp;
 
   try {
     deps.cdp.trackSessionTab?.(ctx.sessionId, target.tabId);
@@ -686,7 +790,7 @@ export async function handleFill(
       target.tabId,
       "DOM.describeNode",
       {
-        backendNodeId: node.backendNodeId,
+        backendNodeId: input.backendNodeId,
       },
     );
     if (!described.node || !isFillable(described.node)) {
@@ -699,274 +803,139 @@ export async function handleFill(
     const scrollErr = await scrollElementAndFramesIntoView(
       deps.cdp,
       target.tabId,
-      node.cdpTarget,
-      node.backendNodeId,
-      node.frameId,
+      input.target,
+      input.backendNodeId,
+      input.frameId,
     );
     if (scrollErr) return scrollErr;
     if (throwIfAborted(deps.signal)) {
       return { code: "cancelled", message: "fill aborted" };
     }
+    await nodeCdp.send(target.tabId, "DOM.focus", { backendNodeId: input.backendNodeId });
   } catch (err) {
-    return fillError("fill_failed", err instanceof Error ? err.message : String(err));
+    return {
+      code: "cdp_failed",
+      message: err instanceof Error ? err.message : String(err),
+    };
   }
 
   if (throwIfAborted(deps.signal)) {
     return { code: "cancelled", message: "fill aborted" };
   }
 
-  const objectIdOrErr = await backendNodeToObject(nodeCdp, target.tabId, node.backendNodeId);
+  const objectIdOrErr = await backendNodeToObject(nodeCdp, target.tabId, input.backendNodeId);
   if (isRpcError(objectIdOrErr)) return objectIdOrErr;
   const objectId = objectIdOrErr;
   const clearBefore = params.clear_before ?? true;
+  const fillResult = await fillEditable(
+    nodeCdp,
+    target.tabId,
+    objectId,
+    params.value,
+    clearBefore,
+    deps.signal,
+  );
+  if (isRpcError(fillResult)) return fillResult;
 
-  try {
-    if (throwIfAborted(deps.signal)) {
-      return { code: "cancelled", message: "fill aborted" };
-    }
-    try {
-      await nodeCdp.send(target.tabId, "DOM.focus", { backendNodeId: node.backendNodeId });
-    } catch (error) {
-      // Chrome rejects DOM.focus for disabled controls, including inherited
-      // fieldset state. Classify the live target before treating this as a
-      // browser failure; the normal path needs no extra round trip.
-      const editable = await nodeCdp.send<FillScriptReply<boolean>>(
-        target.tabId,
-        "Runtime.callFunctionOn",
-        { objectId, functionDeclaration: FILL_EDITABLE_FUNCTION, returnByValue: true },
-      );
-      if (throwIfAborted(deps.signal)) {
-        return { code: "cancelled", message: "fill aborted" };
-      }
-      if (!editable.exceptionDetails && editable.result?.value === false) {
-        return fillError(
-          "target_not_fillable",
-          "fill target is not editable or its input type is unsupported",
-        );
-      }
-      throw error;
-    }
-    if (throwIfAborted(deps.signal)) {
-      return { code: "cancelled", message: "fill aborted" };
-    }
-    // Check editability before clearing. Keep the expected result tied to
-    // this object, including the existing value on the append path.
-    const prepared = await nodeCdp.send<FillScriptReply<FillPreparation>>(
-      target.tabId,
-      "Runtime.callFunctionOn",
-      {
-        objectId,
-        functionDeclaration: `function(value, clearBefore) {
-          const tag = this.tagName.toLowerCase();
-          const native = tag === 'input' || tag === 'textarea';
-          if (!(${FILL_EDITABLE_FUNCTION}).call(this)) return { reason: 'target_not_fillable', error: 'fill target is not editable or its input type is unsupported' };
-          if (this.getRootNode().activeElement !== this) return { reason: 'fill_focus_lost', error: 'fill target does not have focus' };
-          const before = clearBefore ? '' : (${FILL_VALUE_FUNCTION}).call(this);
-          let expected = before + value;
-          if (native) {
-            // Use the browser's own value sanitization (e.g. textarea
-            // line endings) without changing the live control.
-            const normalizer = this.ownerDocument.createElement(tag);
-            if (tag === 'input') {
-              normalizer.type = this.type;
-              normalizer.multiple = this.multiple;
-            }
-            // insertText treats line breaks in a single-line input as spaces.
-            if (tag === 'input') expected = expected.replace(/\\r\\n|\\r|\\n/g, ' ');
-            normalizer.value = expected;
-            if (expected !== '' && normalizer.value === '' && this.type === 'number') {
-              return { reason: 'fill_value_invalid', error: 'fill value is not valid for a number input' };
-            }
-            expected = normalizer.value;
-            if (this.type !== 'number' && this.maxLength >= 0 && expected.length > this.maxLength) {
-              return { reason: 'fill_value_invalid', error: 'fill value exceeds the target maxlength' };
-            }
-          } else {
-            expected = expected.replace(/\\r\\n?/g, '\\n');
-          }
-          if (clearBefore) {
-            if (native) {
-              const proto = tag === 'textarea' ? HTMLTextAreaElement.prototype : HTMLInputElement.prototype;
-              Object.getOwnPropertyDescriptor(proto, 'value').set.call(this, '');
-            } else {
-              this.textContent = '';
-            }
-            this.dispatchEvent(new Event('input', { bubbles: true }));
-          }
-          return { before, expected };
-        }`,
-        arguments: [{ value: params.value }, { value: clearBefore }],
-        returnByValue: true,
-      },
-    );
-    if (throwIfAborted(deps.signal)) {
-      return { code: "cancelled", message: "fill aborted" };
-    }
-    if (prepared.exceptionDetails) {
-      return fillError("fill_failed", "fill preparation script failed");
-    }
-    const preparation = prepared.result?.value;
-    if (preparation && "error" in preparation) {
-      return fillError(preparation.reason, preparation.error);
-    }
-    if (typeof preparation?.before !== "string" || typeof preparation?.expected !== "string") {
-      return fillError("fill_failed", "fill preparation returned an unexpected result");
-    }
-    // A separate call runs after the clearing event's microtasks drain.
-    // DOM focus works in background tabs. Use document.hasFocus() only to
-    // detect deferred focus events, never to reject background input.
-    const checkReady = async (): Promise<FillReadiness | RpcError> => {
-      const reply = await nodeCdp.send<FillScriptReply<FillReadiness>>(
-        target.tabId,
-        "Runtime.callFunctionOn",
-        {
-          objectId,
-          functionDeclaration: `function(before) {
-            if (!(${FILL_EDITABLE_FUNCTION}).call(this) ||
-                (${FILL_VALUE_FUNCTION}).call(this) !== before) return 'fill_target_changed';
-            if (this.getRootNode().activeElement !== this) return 'fill_focus_lost';
-            return this.ownerDocument.hasFocus() ? 'ready' : 'background';
-          }`,
-          arguments: [{ value: preparation.before }],
-          returnByValue: true,
-        },
-      );
-      const state = reply.result?.value;
-      if (
-        reply.exceptionDetails ||
-        (state !== "ready" &&
-          state !== "background" &&
-          state !== "fill_target_changed" &&
-          state !== "fill_focus_lost")
-      ) {
-        return fillError(
-          "fill_failed",
-          "fill readiness script failed or returned an unexpected result",
-        );
-      }
-      return state;
-    };
-    let ready = await checkReady();
-    if (throwIfAborted(deps.signal)) {
-      return { code: "cancelled", message: "fill aborted" };
-    }
-    // Restore focus once only if clearing left this same target editable
-    // and unchanged. Recheck after focus handlers have run.
-    if (clearBefore && ready === "fill_focus_lost") {
-      await nodeCdp.send(target.tabId, "DOM.focus", { backendNodeId: node.backendNodeId });
-      if (throwIfAborted(deps.signal)) {
-        return { code: "cancelled", message: "fill aborted" };
-      }
-      ready = await checkReady();
-    }
-    if (throwIfAborted(deps.signal)) {
-      return { code: "cancelled", message: "fill aborted" };
-    }
-    if (typeof ready !== "string") return ready;
-    if (ready !== "ready" && ready !== "background") {
-      return fillError(ready, "fill target changed or lost focus before typing");
-    }
-    if (params.value !== "" && (!clearBefore || ready === "background")) {
-      // Native editing commands also support number/email inputs, whose
-      // selection APIs cannot set a caret, and multiline contenteditables.
-      // Also do this for background replacement: CDP input focuses the
-      // renderer, delivering deferred focus events before it inserts text.
-      // Use a command-only event so this does not invoke an End shortcut.
-      try {
-        await deps.cdp.send(target.tabId, "Input.dispatchKeyEvent", {
-          type: "rawKeyDown",
-          commands: ["moveToEndOfDocument"],
-        });
-      } finally {
-        await deps.cdp.send(target.tabId, "Input.dispatchKeyEvent", { type: "keyUp" });
-      }
-      if (throwIfAborted(deps.signal)) {
-        return { code: "cancelled", message: "fill aborted" };
-      }
-      // Focus and key handlers can change the target or move focus too.
-      ready = await checkReady();
-      if (throwIfAborted(deps.signal)) {
-        return { code: "cancelled", message: "fill aborted" };
-      }
-      if (typeof ready !== "string") return ready;
-      if (ready !== "ready" && ready !== "background") {
-        return fillError(ready, "fill target changed or lost focus while positioning the caret");
-      }
-    }
-    // CDP `Input.insertText` handles IME / multi-byte input out of the
-    // box, much more reliably than per-key `dispatchKeyEvent`.
-    if (params.value !== "") {
-      await deps.cdp.send(target.tabId, "Input.insertText", { text: params.value });
-    }
-    if (throwIfAborted(deps.signal)) {
-      return { code: "cancelled", message: "fill aborted" };
-    }
-    // Keep the existing notifications, then read in a separate call so
-    // nested microtasks queued by the page cannot race the verification.
-    const notified = await nodeCdp.send<FillScriptReply<unknown>>(
-      target.tabId,
-      "Runtime.callFunctionOn",
-      {
-        objectId,
-        functionDeclaration: `function() {
-          this.dispatchEvent(new Event('input', { bubbles: true }));
-          this.dispatchEvent(new Event('change', { bubbles: true }));
-        }`,
-        returnByValue: true,
-      },
-    );
-    if (throwIfAborted(deps.signal)) {
-      return { code: "cancelled", message: "fill aborted" };
-    }
-    if (notified.exceptionDetails) {
-      return fillError("fill_failed", "fill notification script failed");
-    }
-    const verified = await nodeCdp.send<
-      FillScriptReply<{ connected: boolean; matches: boolean; valueLength: number }>
-    >(target.tabId, "Runtime.callFunctionOn", {
-      objectId,
-      functionDeclaration: `function(expected) {
-          const value = (${FILL_VALUE_FUNCTION}).call(this);
-          return { connected: this.isConnected, matches: value === expected, valueLength: value.length };
-        }`,
-      arguments: [{ value: preparation.expected }],
-      returnByValue: true,
-    });
-    if (throwIfAborted(deps.signal)) {
-      return { code: "cancelled", message: "fill aborted" };
-    }
-    const verification = verified.result?.value;
-    if (
-      verified.exceptionDetails ||
-      typeof verification?.connected !== "boolean" ||
-      typeof verification?.matches !== "boolean" ||
-      typeof verification?.valueLength !== "number"
-    ) {
-      return fillError(
-        "fill_failed",
-        "fill verification script failed or returned an unexpected result",
-      );
-    }
-    if (!verification.connected) {
-      return fillError("fill_target_changed", "fill target was removed or replaced during input");
-    }
-    if (!verification.matches) {
-      return fillError(
-        "fill_value_mismatch",
-        "fill could not verify the expected value; observe the page before deciding whether to retry",
-      );
-    }
-    return attachDialogs(deps.cdp, target.tabId, dialogCursor, {
-      tab_id: target.tabId,
-      used_ref: node.usedRef,
-      used_selector: node.usedSelector,
-      value_length: verification.valueLength,
-    });
-  } catch (err) {
-    return fillError("fill_failed", err instanceof Error ? err.message : String(err));
-  } finally {
-    await nodeCdp.send(target.tabId, "Runtime.releaseObject", { objectId }).catch(() => undefined);
+  return attachDialogs(deps.cdp, target.tabId, dialogCursor, {
+    tab_id: target.tabId,
+    used_ref: input.usedRef,
+    used_selector: input.usedSelector,
+    value_length: fillResult.value.length,
+  });
+}
+
+// ---------------------------------------------------------------------------
+// tool.type
+// ---------------------------------------------------------------------------
+
+export async function handleType(
+  manager: SessionManager,
+  params: TypeParams,
+  deps: InteractionDeps = getDefaultDeps(),
+): Promise<TypeResult | RpcError> {
+  if (!params || typeof params.text !== "string") {
+    return { code: "invalid_params", message: "type requires a text string" };
   }
+  if (params.focused !== undefined && typeof params.focused !== "boolean") {
+    return { code: "invalid_params", message: "type focused must be a boolean" };
+  }
+  const hasExplicitTarget =
+    (typeof params.ref === "string" && params.ref.length > 0) ||
+    (typeof params.selector === "string" && params.selector.length > 0);
+  if (params.focused === true && hasExplicitTarget) {
+    return {
+      code: "invalid_params",
+      message: "type: pass a ref/selector or focused=true, not both",
+    };
+  }
+  if (params.focused !== true && !hasExplicitTarget) {
+    return {
+      code: "invalid_params",
+      message: "type requires a ref/selector or focused=true",
+    };
+  }
+  const ctxOrErr = lookupSession(manager, params, "type");
+  if (isRpcError(ctxOrErr)) return ctxOrErr;
+  const ctx = ctxOrErr;
+  if (throwIfAborted(deps.signal)) return { code: "cancelled", message: "type aborted" };
+  const target = await resolveTargetTab(manager, ctx, params.tab_id, deps.tabsApi);
+  if (isRpcError(target)) return target;
+  const denied = enforceAgentWindow(ctx, target, "type");
+  if (denied) return denied;
+  const dialogCursor = markDialogCursor(deps.cdp, target.tabId);
+
+  let input: InputDispatchContext | RpcError;
+  if (hasExplicitTarget) {
+    const node = await resolveBackendNode(deps.cdp, ctx, target, params, "type");
+    const explicitInput = inputContextForNode(deps.cdp, node);
+    if (isRpcError(explicitInput)) return explicitInput;
+    input = explicitInput;
+    try {
+      const described = await explicitInput.cdp.send<{ node?: DescribedNode }>(
+        target.tabId,
+        "DOM.describeNode",
+        { backendNodeId: explicitInput.backendNodeId },
+      );
+      if (!described.node || !isTypeTarget(described.node)) {
+        return rpcError(
+          "invalid_params",
+          "target_not_fillable",
+          `element ${described.node?.nodeName ?? "?"} not text-editable`,
+        );
+      }
+      deps.cdp.trackSessionTab?.(ctx.sessionId, target.tabId);
+      const scrollErr = await scrollElementAndFramesIntoView(
+        deps.cdp,
+        target.tabId,
+        explicitInput.target,
+        explicitInput.backendNodeId,
+        explicitInput.frameId,
+      );
+      if (scrollErr) return scrollErr;
+      if (throwIfAborted(deps.signal)) return { code: "cancelled", message: "type aborted" };
+      await explicitInput.cdp.send(target.tabId, "DOM.focus", {
+        backendNodeId: explicitInput.backendNodeId,
+      });
+    } catch (err) {
+      return { code: "cdp_failed", message: err instanceof Error ? err.message : String(err) };
+    }
+  } else {
+    input = await resolveFocusedInputContext(deps.cdp, target.tabId, "type");
+    if (isRpcError(input)) return input;
+  }
+
+  deps.cdp.trackSessionTab?.(ctx.sessionId, target.tabId);
+  const dispatchError = await dispatchTextInput(input.cdp, target.tabId, params.text, deps.signal);
+  if (dispatchError) return dispatchError;
+
+  return attachDialogs(deps.cdp, target.tabId, dialogCursor, {
+    tab_id: target.tabId,
+    used_ref: input.usedRef,
+    used_selector: input.usedSelector,
+    text_length: params.text.length,
+    verification: "dispatched",
+  });
 }
 
 // ---------------------------------------------------------------------------
@@ -1140,31 +1109,38 @@ export async function handlePress(
     };
   }
 
+  let input: InputDispatchContext | RpcError;
   // Optional focus before key dispatch.
   if (params.ref || params.selector) {
     const node = await resolveBackendNode(deps.cdp, ctx, target, params, "press");
-    if (isRpcError(node)) return node;
-    const nodeCdp = cdpRunnerForTarget(deps.cdp, node.cdpTarget);
+    const explicitInput = inputContextForNode(deps.cdp, node);
+    if (isRpcError(explicitInput)) return explicitInput;
+    input = explicitInput;
     try {
       deps.cdp.trackSessionTab?.(ctx.sessionId, target.tabId);
       const scrollErr = await scrollElementAndFramesIntoView(
         deps.cdp,
         target.tabId,
-        node.cdpTarget,
-        node.backendNodeId,
-        node.frameId,
+        explicitInput.target,
+        explicitInput.backendNodeId,
+        explicitInput.frameId,
       );
       if (scrollErr) return scrollErr;
       if (throwIfAborted(deps.signal)) {
         return { code: "cancelled", message: "press aborted" };
       }
-      await nodeCdp.send(target.tabId, "DOM.focus", { backendNodeId: node.backendNodeId });
+      await explicitInput.cdp.send(target.tabId, "DOM.focus", {
+        backendNodeId: explicitInput.backendNodeId,
+      });
     } catch (err) {
       return {
         code: "cdp_failed",
         message: err instanceof Error ? err.message : String(err),
       };
     }
+  } else {
+    input = await resolveFocusedInputContext(deps.cdp, target.tabId, "press");
+    if (isRpcError(input)) return input;
   }
 
   if (throwIfAborted(deps.signal)) {
@@ -1178,7 +1154,7 @@ export async function handlePress(
   try {
     let cancelled = false;
     deps.cdp.trackSessionTab?.(ctx.sessionId, target.tabId);
-    await deps.cdp.send(target.tabId, "Input.dispatchKeyEvent", {
+    await input.cdp.send(target.tabId, "Input.dispatchKeyEvent", {
       type: "rawKeyDown",
       key: descriptor.key,
       code: descriptor.code,
@@ -1192,7 +1168,7 @@ export async function handlePress(
       typeof descriptor.text === "string" &&
       descriptor.text.length > 0
     ) {
-      await deps.cdp.send(target.tabId, "Input.dispatchKeyEvent", {
+      await input.cdp.send(target.tabId, "Input.dispatchKeyEvent", {
         type: "char",
         key: descriptor.key,
         code: descriptor.code,
@@ -1209,7 +1185,7 @@ export async function handlePress(
         cancelled = true;
       }
     }
-    await deps.cdp.send(target.tabId, "Input.dispatchKeyEvent", {
+    await input.cdp.send(target.tabId, "Input.dispatchKeyEvent", {
       type: "keyUp",
       key: descriptor.key,
       code: descriptor.code,

apps/extension/src/tools/mouse-input.ts

@@ -0,0 +1,96 @@
+import type { MouseButton } from "@/transport/types";
+import type { Point } from "./geometry";
+import type { CdpRunner } from "./shared";
+
+interface MouseClickDispatchOptions {
+  button: MouseButton;
+  clickCount: number;
+  modifiers: number;
+  signal?: AbortSignal;
+  moveSettleMs?: number;
+}
+
+type MouseClickDispatchResult = "completed" | "cancelled";
+
+function waitForSettle(ms: number, signal: AbortSignal | undefined): Promise<boolean> {
+  if (ms <= 0) return Promise.resolve(signal?.aborted !== true);
+  if (signal?.aborted) return Promise.resolve(false);
+  return new Promise((resolve) => {
+    let settled = false;
+    const finish = (completed: boolean) => {
+      if (settled) return;
+      settled = true;
+      clearTimeout(timer);
+      signal?.removeEventListener("abort", onAbort);
+      resolve(completed);
+    };
+    const timer = setTimeout(() => finish(true), ms);
+    const onAbort = () => finish(false);
+    signal?.addEventListener("abort", onAbort, { once: true });
+  });
+}
+
+/**
+ * Dispatch one coherent CDP mouse click in top-viewport CSS coordinates.
+ *
+ * Keeping the sequence here prevents DOM and Surface interaction paths from
+ * duplicating cancellation and best-effort release behavior. Callers that
+ * target asynchronously hit-tested content may request a settle interval.
+ */
+export async function dispatchMouseClick(
+  cdp: CdpRunner,
+  tabId: number,
+  point: Point,
+  options: MouseClickDispatchOptions,
+): Promise<MouseClickDispatchResult> {
+  const { button, clickCount, modifiers, signal, moveSettleMs = 0 } = options;
+  let pressed = false;
+  const release = () =>
+    cdp.send(tabId, "Input.dispatchMouseEvent", {
+      type: "mouseReleased",
+      x: point.x,
+      y: point.y,
+      button,
+      clickCount,
+      modifiers,
+    });
+
+  try {
+    await cdp.send(tabId, "Input.dispatchMouseEvent", {
+      type: "mouseMoved",
+      x: point.x,
+      y: point.y,
+      modifiers,
+    });
+    if (!(await waitForSettle(moveSettleMs, signal))) return "cancelled";
+
+    await cdp.send(tabId, "Input.dispatchMouseEvent", {
+      type: "mousePressed",
+      x: point.x,
+      y: point.y,
+      button,
+      clickCount,
+      modifiers,
+    });
+    pressed = true;
+
+    if (signal?.aborted) {
+      await release();
+      pressed = false;
+      return "cancelled";
+    }
+
+    await release();
+    pressed = false;
+    return "completed";
+  } catch (error) {
+    if (pressed) {
+      try {
+        await release();
+      } catch (releaseError) {
+        console.debug("[bsk mouse-input] best-effort mouse release failed", releaseError);
+      }
+    }
+    throw error;
+  }
+}

apps/extension/src/tools/navigation.ts

@@ -28,13 +28,7 @@ import type {
   RpcError,
   WaitUntil,
 } from "@/transport/types";
-import {
-  type BrowserNavigationApi,
-  chromeBrowserNavigationApi,
-  navigateWithBrowserApi,
-} from "./browser-navigation";
 import { attachDialogs, markDialogCursor } from "./dialogs";
-import { cdpError, isCdpExtensionAccessDenied } from "./errors";
 import {
   type CdpRunner,
   type ChromeTabsApi,
@@ -48,7 +42,6 @@ import {
 export interface NavigationDeps {
   cdp: CdpRunner;
   tabsApi: ChromeTabsApi;
-  browserNavigation?: BrowserNavigationApi;
   /** Optional AbortSignal — M7 abort hook (M10.2 will wire the full chain). */
   signal?: AbortSignal;
   /** Override default timeout when the caller omits `timeout_ms`. */
@@ -486,81 +479,6 @@ async function readTabUrl(api: ChromeTabsApi, tabId: number): Promise<string | u
   }
 }
 
-/** Recover only a failed preflight; never replay an already-dispatched navigation. */
-async function recoverBrowserNavigation(
-  deps: NavigationDeps,
-  tabId: number,
-  action: (api: BrowserNavigationApi) => Promise<unknown>,
-  waitUntil: WaitUntil,
-  timeoutMs: number,
-): Promise<ReloadResult | RpcError> {
-  const abort = linkedAbortSignal(deps.signal);
-  const deadline = Date.now() + timeoutMs;
-  const api = deps.browserNavigation ?? chromeBrowserNavigationApi;
-  try {
-    let outcome = await navigateWithBrowserApi(
-      api,
-      tabId,
-      () => action(api),
-      waitUntil === "networkidle" ? "commit" : waitUntil,
-      timeoutMs,
-      abort.signal,
-    );
-    if (outcome.reached === "failed") return outcome.error;
-    if (outcome.reached === "cancelled" || abort.signal.aborted) {
-      return { code: "cancelled", message: "navigation recovery aborted" };
-    }
-    if (outcome.reached === "match") {
-      // A completed reload is not a recovery if the restricted frame remains.
-      await deps.cdp.send(tabId, "Page.enable", {});
-      if (abort.signal.aborted)
-        return { code: "cancelled", message: "navigation recovery aborted" };
-      if (waitUntil === "networkidle") {
-        // webNavigation cannot prove network idle. Subscribe on the new
-        // document before enabling CDP lifecycle events after its commit.
-        const frame = await readMainFrameInfo(deps.cdp, tabId);
-        if (!frame.frameId) return cdpError("no main frame after navigation");
-        const remaining = deadline - Date.now();
-        if (remaining <= 0) {
-          outcome = { reached: "timeout", lastLifecycle: "commit" };
-        } else {
-          const wait = startLifecycleWait(
-            deps.cdp,
-            tabId,
-            frame.frameId,
-            "networkIdle",
-            remaining,
-            abort.signal,
-            { loaderId: frame.loaderId ?? undefined },
-          );
-          await deps.cdp.send(tabId, "Page.setLifecycleEventsEnabled", { enabled: true });
-          outcome = await wait.promise;
-        }
-      }
-    }
-    if (outcome.reached === "cancelled" || abort.signal.aborted) {
-      return { code: "cancelled", message: "navigation recovery aborted" };
-    }
-    return {
-      tab_id: tabId,
-      final_url: await readTabUrl(deps.tabsApi, tabId),
-      reached: outcome.reached === "match" ? waitUntil : "timeout",
-      ...(outcome.reached === "timeout"
-        ? {
-            error_text: `timed out waiting for ${waitUntil} during navigation recovery after ${timeoutMs}ms`,
-          }
-        : {}),
-    };
-  } catch (error) {
-    return abort.signal.aborted
-      ? { code: "cancelled", message: "navigation recovery aborted" }
-      : cdpError(error);
-  } finally {
-    abort.abort();
-    abort.cleanup();
-  }
-}
-
 // ---------------------------------------------------------------------------
 // tool.navigate
 // ---------------------------------------------------------------------------
@@ -587,20 +505,7 @@ export async function handleNavigate(
 
   try {
     deps.cdp.trackSessionTab?.(ctx.sessionId, target.tabId);
-    try {
-      await ensureCdpReady(deps.cdp, target.tabId);
-    } catch (error) {
-      if (!isCdpExtensionAccessDenied(error)) throw error;
-      const recovered = await recoverBrowserNavigation(
-        deps,
-        target.tabId,
-        (api) => api.update(target.tabId, { url: params.url }),
-        waitUntil,
-        timeoutMs,
-      );
-      if (isRpcError(recovered)) return recovered;
-      return attachDialogs(deps.cdp, target.tabId, dialogCursor, { ...recovered, url: params.url });
-    }
+    await ensureCdpReady(deps.cdp, target.tabId);
     const expected = cdpLifecycleName(waitUntil);
     const beforeReadyState = await probeMainFrameReadyState(deps.cdp, target.tabId);
     const beforeFrame = await readMainFrameInfo(deps.cdp, target.tabId);
@@ -669,7 +574,10 @@ export async function handleNavigate(
       }`,
     });
   } catch (err) {
-    return cdpError(err);
+    return {
+      code: "cdp_failed",
+      message: err instanceof Error ? err.message : String(err),
+    };
   }
 }
 
@@ -819,24 +727,7 @@ export async function handleReload(
 
   try {
     deps.cdp.trackSessionTab?.(ctx.sessionId, target.tabId);
-    try {
-      await ensureCdpReady(deps.cdp, target.tabId);
-    } catch (error) {
-      if (!isCdpExtensionAccessDenied(error)) throw error;
-      const previousUrl = await readTabUrl(deps.tabsApi, target.tabId);
-      const recovered = await recoverBrowserNavigation(
-        deps,
-        target.tabId,
-        (api) => api.reload(target.tabId, { bypassCache: ignoreCache }),
-        waitUntil,
-        timeoutMs,
-      );
-      if (isRpcError(recovered)) return recovered;
-      return attachDialogs(deps.cdp, target.tabId, dialogCursor, {
-        ...recovered,
-        previous_url: previousUrl,
-      });
-    }
+    await ensureCdpReady(deps.cdp, target.tabId);
     const previousUrl = await readTabUrl(deps.tabsApi, target.tabId);
     const beforeFrame = await readMainFrameInfo(deps.cdp, target.tabId);
     const expected = cdpLifecycleName(waitUntil);
@@ -885,7 +776,10 @@ export async function handleReload(
       }`,
     });
   } catch (err) {
-    return cdpError(err);
+    return {
+      code: "cdp_failed",
+      message: err instanceof Error ? err.message : String(err),
+    };
   }
 }
 

apps/extension/src/tools/observation.ts

@@ -8,6 +8,7 @@ import {
   applyVomInteractionRecovery,
   type CondSurface,
   isVomReferenceNode,
+  type Rect,
   renderVom,
   type VomNode,
   type VomOptions,
@@ -17,7 +18,7 @@ import { ChromiumCdp } from "@/browser-driver/chromium-cdp";
 import type { CdpTarget } from "@/browser-driver/frame-graph";
 import {
   type CaptureSuppressSendToTab,
-  withExtensionOverlayHidden,
+  withOverlaysHiddenForCapture,
 } from "@/lib/capture-suppress-bridge";
 import type { SessionContext, SessionManager } from "@/session-manager/manager";
 import type {
@@ -47,21 +48,25 @@ import {
   type ToolEffect,
 } from "./shared";
 import { resolveSnapshotRef } from "./snapshot-ref";
+import { surfaceVisibleRect } from "./surface-coordinate";
+import { captureSurfaceEnvironment } from "./surface-target";
 import {
   type CapturedNode,
-  type CapturedSurfaceProbe,
   type CapturedViewModel,
   captureViewModel,
   collectOverlayExcludedBackendIds,
-  probeHoverSurfaces,
 } from "./vom/capture";
 import { type CapturedFrameDocument, captureFrameData } from "./vom/frame-capture";
-import { withOverlayBypass } from "./vom/hover-perception";
 import { probeTooltipNames } from "./vom/name-enrichment";
 import {
   type CaptureVomObservationResult,
   projectRecordSafeObservation,
 } from "./vom/record-safe-observation";
+import {
+  clusterRenderedSurfaces,
+  discoverRenderedSurfaces,
+  projectRenderedSurfaces,
+} from "./vom/rendered-surfaces";
 import {
   buildSemanticGraph,
   buildSemanticVomScene,
@@ -93,6 +98,9 @@ export const normaliseRef = sharedNormaliseRef;
 // screenshot — `tool.screenshot`
 // ---------------------------------------------------------------------------
 
+const VISUAL_SURFACE_SCREENSHOT_MAX_EDGE = 2_048;
+const VISUAL_SURFACE_SCREENSHOT_MAX_PIXELS = 4_000_000;
+
 /**
  * Strip the `data:image/...;base64,` prefix from a Chrome
  * `captureVisibleTab` dataURL and return the raw base64 payload.
@@ -185,18 +193,42 @@ async function captureElementScreenshot(
   target: CdpTarget,
   backendNodeId: number,
   frameId?: string,
+  visualSurface = false,
+  observedVisibleRect?: Rect,
   signal?: AbortSignal,
-): Promise<{ image_base64: string; width: number; height: number } | RpcError> {
+): Promise<
+  { image_base64: string; width: number; height: number; topViewportRect: Rect } | RpcError
+> {
   if (signal?.aborted) return cancelled("screenshot");
   const geometry = await resolveNodeGeometry(
     cdp,
     tabId,
     { target, backendNodeId, ...(frameId ? { frameId } : {}) },
-    { scrollIntoView: true },
+    { scrollIntoView: !visualSurface },
   );
   if (isRpcError(geometry)) return geometry;
   if (signal?.aborted) return cancelled("screenshot");
-  const rect = geometry.topBounds;
+  const topViewportRect = surfaceVisibleRect(
+    geometry.topBounds,
+    visualSurface ? observedVisibleRect : undefined,
+  );
+  if (!topViewportRect) {
+    return { code: "permission_denied", message: "surface is no longer visible" };
+  }
+  const rect = {
+    x: topViewportRect.x,
+    y: topViewportRect.y,
+    width: topViewportRect.w,
+    height: topViewportRect.h,
+  };
+  const scale = visualSurface
+    ? Math.min(
+        1,
+        VISUAL_SURFACE_SCREENSHOT_MAX_EDGE / rect.width,
+        VISUAL_SURFACE_SCREENSHOT_MAX_EDGE / rect.height,
+        Math.sqrt(VISUAL_SURFACE_SCREENSHOT_MAX_PIXELS / (rect.width * rect.height)),
+      )
+    : 1;
 
   try {
     const shot = await cdp.send<{ data?: string }>(tabId, "Page.captureScreenshot", {
@@ -206,7 +238,7 @@ async function captureElementScreenshot(
         y: rect.y,
         width: rect.width,
         height: rect.height,
-        scale: 1,
+        scale,
       },
     });
     if (signal?.aborted) return cancelled("screenshot");
@@ -218,7 +250,7 @@ async function captureElementScreenshot(
       width: Math.round(rect.width),
       height: Math.round(rect.height),
     };
-    return { image_base64, width: dims.width, height: dims.height };
+    return { image_base64, width: dims.width, height: dims.height, topViewportRect };
   } catch (err) {
     return {
       code: "cdp_failed",
@@ -313,7 +345,7 @@ export async function handleScreenshot(
     if (!deps.cdp) {
       return { code: "cdp_failed", message: "screenshot ref capture requires CDP" };
     }
-    const node = resolveSnapshotRef(ctx, ref, target.tabId);
+    const node = resolveSnapshotRef(ctx, ref, target.tabId, "screenshot");
     if (isRpcError(node)) return node;
     if (signal?.aborted) return cancelled("screenshot");
     deps.cdp.trackSessionTab?.(ctx.sessionId, target.tabId);
@@ -324,7 +356,7 @@ export async function handleScreenshot(
       tabId: target.tabId,
       ...(node.cdpSessionId ? { sessionId: node.cdpSessionId } : {}),
     };
-    const captured = await withExtensionOverlayHidden(
+    const captured = await withOverlaysHiddenForCapture(
       target.tabId,
       () =>
         captureElementScreenshot(
@@ -333,18 +365,54 @@ export async function handleScreenshot(
           nodeTarget,
           node.backendNodeId,
           node.frameId,
+          node.kind === "surface",
+          node.visibleRect,
           signal,
         ),
       deps.sendToTab,
     );
     if (isRpcError(captured)) return captured;
     if (signal?.aborted) return cancelled("screenshot");
+    const capture =
+      node.kind === "surface"
+        ? await (async () => {
+            const environment = await captureSurfaceEnvironment(cdp, target.tabId, node.frameId);
+            if (isRpcError(environment)) return environment;
+            return ctx.surfaceCaptures.create({
+              sessionId: ctx.sessionId,
+              tabId: target.tabId,
+              navigationIdentity: environment.navigationIdentity,
+              surface: {
+                ref: node.refKey,
+                ...(node.frameId ? { frameId: node.frameId } : {}),
+                backendNodeId: node.backendNodeId,
+                observationGeneration: node.generation,
+              },
+              topViewportRect: captured.topViewportRect,
+              imageWidth: captured.width,
+              imageHeight: captured.height,
+              viewportSignature: environment.viewportSignature,
+              frameProjectionSignature: environment.frameProjectionSignature,
+            });
+          })()
+        : null;
+    if (capture && isRpcError(capture)) return capture;
     return withShotDialogs({
       image_base64: captured.image_base64,
       width: captured.width,
       height: captured.height,
       format: "png",
       tab_id: target.tabId,
+      ...(capture
+        ? {
+            capture: {
+              id: capture.id,
+              surface_ref: `@${capture.surface.ref}`,
+              coordinate_space: "capture-image-pixel" as const,
+              expires_at: capture.expiresAt,
+            },
+          }
+        : {}),
     });
   }
 
@@ -356,7 +424,7 @@ export async function handleScreenshot(
     );
   }
 
-  const captured = await withExtensionOverlayHidden(
+  const captured = await withOverlaysHiddenForCapture(
     target.tabId,
     () => captureFullTabPng(deps, ctx, target, signal),
     deps.sendToTab,
@@ -528,11 +596,8 @@ function buildActiveScopeBlocks(nodes: VomNode[], signals: VomNodeDomSignals): A
   return blocks;
 }
 
-function buildConditionalSurfaces(
-  nodes: VomNode[],
-  captured: CapturedViewModel,
-  probes: CapturedSurfaceProbe[],
-): CondSurface[] {
+function buildConditionalSurfaces(nodes: VomNode[], captured: CapturedViewModel): CondSurface[] {
+  const probes = captured.surfaceProbes ?? [];
   if (probes.length === 0) return [];
 
   const surfaces: CondSurface[] = [];
@@ -655,7 +720,6 @@ function findSurfaceNodeByPoint(
 export interface BuildVomSceneOptions {
   pageUrl?: string;
   supplementalNames?: ReadonlyMap<string, string>;
-  surfaceProbes?: CapturedSurfaceProbe[];
 }
 
 export type VomFrameDocument = CapturedFrameDocument<CdpAxNode>;
@@ -824,19 +888,18 @@ export function buildFrameVomScene(
     excludedBackendNodeIds: captured.excludedBackendNodeIds,
     supplementalNames: options.supplementalNames,
   });
-  return attachCapturedSceneAnnotations(scene, documents, captured, options.surfaceProbes ?? []);
+  return attachCapturedSceneAnnotations(scene, documents, captured);
 }
 
 function attachCapturedSceneAnnotations(
   scene: VomScene,
   documents: VomFrameDocument[],
   captured: CapturedViewModel,
-  surfaceProbes: CapturedSurfaceProbe[],
 ): VomScene {
   const rootDocument = documents.find((document) => document.frameId === scene.rootFrameId);
   const signals = capturedOnlySignals(rootDocument?.domNodes ?? captured.nodes);
   const activeScopeBlocks = buildActiveScopeBlocks(scene.nodes, signals);
-  const surfaces = buildConditionalSurfaces(scene.nodes, captured, surfaceProbes);
+  const surfaces = buildConditionalSurfaces(scene.nodes, captured);
   return {
     ...scene,
     ...(surfaces.length > 0 ? { surfaces } : {}),
@@ -1018,71 +1081,20 @@ async function captureForVom(
   options: CaptureVomObservationOptions,
 ): Promise<CapturedViewModel> {
   try {
-    return await captureViewModel(cdp, tabId, { signal: options.signal });
+    return await captureViewModel(cdp, tabId, {
+      conditionalSurfaceProbe: options.conditionalSurfaceProbe ?? false,
+      hoverProbeBypassOverlay: options.hoverProbeBypassOverlay,
+      signal: options.signal,
+    });
   } catch (error) {
     if (isAbortError(error)) throw error;
     return fallbackCapturedViewModel(cdp, tabId, options.signal);
   }
 }
 
-export interface HoverProbeOutcome {
-  /** Whether any active hover was dispatched during this observation. */
-  performed: boolean;
-  /**
-   * Whether hovering surfaced content absent from the static snapshot. Cached
-   * tooltip names count, so this may over-report on repeat observations — it
-   * errs towards telling the caller the snapshot is less fresh than it looks.
-   */
-  revealedContent: boolean;
-  surfaceProbes: CapturedSurfaceProbe[];
-  tooltipNames: Map<string, string>;
-}
-
-const NO_HOVER_PROBES: HoverProbeOutcome = {
-  performed: false,
-  revealedContent: false,
-  surfaceProbes: [],
-  tooltipNames: new Map(),
-};
-
-/**
- * Runs both active-hover chains back to back.
- *
- * Ordering matters: this happens after DOM *and* accessibility capture so a
- * hover that opens a menu cannot leave one half of the observation describing
- * the page before the change and the other half after it. Sharing one overlay
- * bypass span also means the agent overlay toggles once per observation
- * instead of once per chain.
- */
-async function runHoverProbes(
-  cdp: CdpRunner,
-  tabId: number,
-  captured: CapturedViewModel,
-  documents: VomFrameDocument[],
-  staticSemantics: ReturnType<typeof resolveSemanticGraph>,
-  options: CaptureVomObservationOptions,
-): Promise<HoverProbeOutcome> {
-  if (!options.conditionalSurfaceProbe) return NO_HOVER_PROBES;
-
-  return withOverlayBypass(options.hoverProbeBypassOverlay, tabId, async () => {
-    const surfaceProbes = await probeHoverSurfaces(cdp, tabId, captured.nodes, {
-      signal: options.signal,
-    });
-    throwIfAborted(options.signal, "observation");
-    const tooltipNames = await probeTooltipNames(cdp, tabId, documents, staticSemantics, {
-      signal: options.signal,
-    });
-    return {
-      performed: true,
-      revealedContent: surfaceProbes.length > 0 || tooltipNames.size > 0,
-      surfaceProbes,
-      tooltipNames,
-    };
-  });
-}
-
 export interface CaptureVomObservationOptions extends VomOptions {
   conditionalSurfaceProbe?: boolean;
+  renderedSurfaceDiscovery?: boolean;
   hoverProbeBypassOverlay?: (tabId: number, enabled: boolean) => Promise<void>;
   signal?: AbortSignal;
 }
@@ -1111,27 +1123,33 @@ export async function captureVomObservation(
     excludedBackendNodeIds: captured.excludedBackendNodeIds,
   });
   const staticSemantics = resolveSemanticGraph(semanticGraph, { identifierFallback: false });
-  const hoverProbes = await runHoverProbes(
-    cdp,
-    tabId,
-    captured,
-    normalizedDocuments,
-    staticSemantics,
-    options,
-  );
+  const tooltipNames = options.conditionalSurfaceProbe
+    ? await probeTooltipNames(cdp, tabId, normalizedDocuments, staticSemantics, {
+        signal: options.signal,
+        hoverProbeBypassOverlay: options.hoverProbeBypassOverlay,
+      })
+    : new Map<string, string>();
   throwIfAborted(options.signal, "observation");
   const scene = projectSemanticGraph(
     normalizeSemanticStructure(
-      resolveSemanticGraph(semanticGraph, { supplementalNames: hoverProbes.tooltipNames }),
+      resolveSemanticGraph(semanticGraph, { supplementalNames: tooltipNames }),
     ),
   );
-  const decoratedScene = attachCapturedSceneAnnotations(
-    scene,
-    normalizedDocuments,
-    captured,
-    hoverProbes.surfaceProbes,
-  );
-  const rendered = renderVom(decoratedScene, {
+  const decoratedScene = attachCapturedSceneAnnotations(scene, normalizedDocuments, captured);
+  const observedScene = options.renderedSurfaceDiscovery
+    ? projectRenderedSurfaces(
+        decoratedScene,
+        normalizedDocuments,
+        clusterRenderedSurfaces(
+          discoverRenderedSurfaces(
+            normalizedDocuments,
+            captured.viewport,
+            captured.excludedBackendNodeIds,
+          ),
+        ),
+      )
+    : decoratedScene;
+  const rendered = renderVom(observedScene, {
     maxDepth: options.maxDepth,
     maxTokens: options.maxTokens,
     redactValues: options.redactValues,
@@ -1142,11 +1160,7 @@ export async function captureVomObservation(
     rootFrameId: captured.rootFrameId ?? normalizedDocuments[0]?.frameId ?? "root",
     frameDocuments: normalizedDocuments,
     rendered,
-    surfaceProbes: hoverProbes.surfaceProbes,
-    hoverProbe: {
-      performed: hoverProbes.performed,
-      revealedContent: hoverProbes.revealedContent,
-    },
+    surfaceProbes: captured.surfaceProbes,
   });
 }
 
@@ -1186,6 +1200,7 @@ async function handleVomObservation(
       maxTokens: params.max_tokens,
       activeRegionPolicy: true,
       conditionalSurfaceProbe: effectiveConditionalSurfaceProbe,
+      renderedSurfaceDiscovery: toolName === "observe",
       hoverProbeBypassOverlay: deps.hoverProbeBypassOverlay,
       signal,
     });
@@ -1203,6 +1218,9 @@ async function handleVomObservation(
             tabId: target.tabId,
             ...(ref.frameId ? { frameId: ref.frameId } : {}),
             ...(refTarget?.sessionId ? { cdpSessionId: refTarget.sessionId } : {}),
+            ...(ref.visibleRect ? { visibleRect: ref.visibleRect } : {}),
+            ...(ref.kind ? { kind: ref.kind } : {}),
+            ...(ref.capabilities ? { capabilities: ref.capabilities } : {}),
           },
         ] as const;
       }),
@@ -1212,14 +1230,6 @@ async function handleVomObservation(
       ref_count: observation.refs.length,
       tab_id: target.tabId,
       truncated: observation.truncated,
-      ...(toolName === "observe" && observation.hoverProbe?.performed
-        ? {
-            hover_probe: {
-              performed: true,
-              revealed_content: observation.hoverProbe.revealedContent,
-            },
-          }
-        : {}),
       ...(toolName === "observe" && (params as ObserveParams).debug_surfaces
         ? {
             debug: {
@@ -1258,13 +1268,5 @@ export async function handleObserve(
   deps: SnapshotDeps = getDefaultDeps(),
   signal?: AbortSignal,
 ): Promise<ObserveResult | RpcError> {
-  return handleVomObservation(
-    manager,
-    params,
-    "observe",
-    "transient_input",
-    params.probe_hover === true,
-    deps,
-    signal,
-  );
+  return handleVomObservation(manager, params, "observe", "transient_input", true, deps, signal);
 }

apps/extension/src/tools/session.ts

@@ -2,7 +2,6 @@ import { type SessionManager, SessionStartCleanupError } from "@/session-manager
 import type { RpcError } from "@/transport/types";
 import { rpcError } from "./errors";
 import { clearRecordingForSession } from "./record";
-import type { CdpRunner, ChromeTabsApi } from "./shared";
 import { isRpcError } from "./shared";
 import { returnBorrowedTab, type TabManagementDeps } from "./tabs";
 
@@ -74,17 +73,11 @@ export interface SessionStopResult {
   /** Tab ids whose return path failed; those entries remain borrowed so
    *  shutdown can be retried without closing the Agent Window. */
   return_failures?: Array<{ tab_id: number; code: string; message: string }>;
-  /**
-   * True when the Agent Window was *released* to the user (instead of being
-   * closed) because it still contained user-created tabs after the agent's own
-   * tabs were closed. Single-session scope.
-   */
-  window_released?: boolean;
 }
 
 export interface SessionStopDeps {
   signal?: AbortSignal;
-  cdp?: Pick<CdpRunner, "releaseSessionTab"> & {
+  cdp?: {
     detachSession(sessionId: string): Promise<void>;
   };
   /**
@@ -94,13 +87,6 @@ export interface SessionStopDeps {
    * without a real browser.
    */
   tabManagement?: TabManagementDeps;
-  /**
-   * Read-only Chrome tabs API used to *query* remaining tabs before deciding
-   * whether to release the window. Kept separate from `tabManagement.tabs`
-   * (a `TabMutationApi` that has no `query`) so the mutation interface stays
-   * pure. Tests can inject a fake `ChromeTabsApi` too.
-   */
-  tabsQuery?: ChromeTabsApi;
 }
 
 /**
@@ -166,10 +152,8 @@ export async function handleSessionStart(
  *      cleanly (review parity with M6).
  *   3. Detach CDP sessions the extension still holds for this
  *      session (no-op if M6/M7 didn't attach to any tab).
- *   4. Close the agent-created tabs and the home tab (design §3.2).
- *   5. If any tab remains — those are user-created tabs — release the
- *      window to the user (dropOnly) instead of closing it, so user
- *      tabs survive. Otherwise close the window as before.
+ *   4. Close the Agent Window. SessionManager.stop() removes the
+ *      Chrome window and forgets the context.
  *
  * Failures in step 1 keep the Agent Window open: a failed borrowed tab
  * may still be there, so closing the window would risk losing user
@@ -204,19 +188,15 @@ export async function handleSessionStop(
   const returnFailures: SessionStopResult["return_failures"] = [];
   const borrowedIds = Array.from(ctx.borrowedTabs.keys());
   for (const tabId of borrowedIds) {
-    // A close event may have removed this borrow while another tab returned.
-    if (!ctx.borrowedTabs.has(tabId)) continue;
     try {
       const tabManagement = {
         ...(deps.tabManagement ?? {}),
-        cdp: deps.cdp ?? deps.tabManagement?.cdp,
         signal: deps.signal,
         isAgentWindowId:
           deps.tabManagement?.isAgentWindowId ??
           ((windowId: number) => manager.findByWindowId(windowId) !== null),
       };
       const outcome = await returnBorrowedTab(ctx, tabId, tabManagement);
-      if (!ctx.borrowedTabs.has(tabId)) continue;
       if (typeof outcome === "object" && "code" in outcome) {
         console.warn(`[bsk session_stop] auto-return failed for tab ${tabId}`, outcome);
         returnFailures?.push({
@@ -229,7 +209,6 @@ export async function handleSessionStop(
         ctx.borrowedTabs.delete(tabId);
       }
     } catch (err) {
-      if (!ctx.borrowedTabs.has(tabId)) continue;
       const message = err instanceof Error ? err.message : String(err);
       console.warn(`[bsk session_stop] auto-return threw for tab ${tabId}: ${message}`);
       returnFailures?.push({
@@ -240,15 +219,8 @@ export async function handleSessionStop(
     }
   }
 
-  // A previously failed tab may have closed during a later return. Only
-  // confirmed close events remove borrows; ordinary move failures stay retryable.
-  const pendingReturnFailures = returnFailures.filter((failure) =>
-    ctx.borrowedTabs.has(failure.tab_id),
-  );
   if (deps.signal?.aborted) {
-    const nonCancelFailures = pendingReturnFailures.filter(
-      (failure) => failure.code !== "cancelled",
-    );
+    const nonCancelFailures = returnFailures?.filter((failure) => failure.code !== "cancelled");
     if (nonCancelFailures && nonCancelFailures.length > 0) {
       return {
         ...(returnedTabIds.length > 0 ? { returned_tab_ids: returnedTabIds } : {}),
@@ -260,8 +232,8 @@ export async function handleSessionStop(
 
   const result: SessionStopResult = {};
   if (returnedTabIds.length > 0) result.returned_tab_ids = returnedTabIds;
-  if (pendingReturnFailures.length > 0) {
-    result.return_failures = pendingReturnFailures;
+  if (returnFailures && returnFailures.length > 0) {
+    result.return_failures = returnFailures;
     // A failed return means at least one borrowed user tab may still be
     // inside the Agent Window. Keep the session/window alive so the user
     // can retry `bsk session stop` or explicitly `bsk tab return` after the
@@ -271,6 +243,7 @@ export async function handleSessionStop(
 
   // Step 2: clear the per-session RefStore (review M6/M7 parity).
   ctx.refStore.clear();
+  ctx.surfaceCaptures.clear();
 
   clearRecordingForSession(params.session_id);
 
@@ -281,66 +254,8 @@ export async function handleSessionStop(
     return { code: "cancelled", message: "session_stop aborted before window close" };
   }
 
-  // Step 4: close every tab explicitly created by the agent, including the
-  // home tab. `tabsApi` is a
-  // TabMutationApi (remove only); `queryApi` is a separate read-only
-  // ChromeTabsApi. If neither is injected, we conservatively fall back to
-  // closing the window (see Step 5).
-  const tabsApi = deps.tabManagement?.tabs;
-  const queryApi = deps.tabsQuery;
-
-  if (tabsApi) {
-    // Close each agent-created tab that still exists.
-    const agentCreatedTabIds = Array.from(ctx.agentCreatedTabs);
-    for (const tabId of agentCreatedTabIds) {
-      try {
-        await tabsApi.remove(tabId);
-        ctx.agentCreatedTabs.delete(tabId);
-      } catch (err) {
-        // Tab may already be gone (closed by the user). Non-fatal.
-        console.warn(`[bsk session_stop] failed to close agent tab ${tabId}`, err);
-      }
-    }
-  }
-
-  // Step 5: decide whether to release (keep) the window or close it.
-  let shouldRelease = false;
-  if (queryApi) {
-    try {
-      const liveWindowTabs = await queryApi.query({ windowId: ctx.agentWindowId });
-      // Only genuine *user* tabs count toward keeping the window open.
-      // An agent tab that failed to close in Step 4 may still be present
-      // here; if we counted it as a reason to release (dropOnly), the
-      // window would be kept and that agent tab would leak (issue #57
-      // regression). Exclude any id still tracked in agentCreatedTabs.
-      const userTabs = liveWindowTabs.filter((t) => {
-        if (t.id === undefined) return false;
-        return !ctx.agentCreatedTabs.has(t.id);
-      });
-      const leakedAgentTabs = liveWindowTabs.filter(
-        (t) => t.id !== undefined && ctx.agentCreatedTabs.has(t.id),
-      );
-      if (leakedAgentTabs.length > 0) {
-        console.warn(
-          `[bsk session_stop] ${leakedAgentTabs.length} agent tab(s) failed to close; forcing window close instead of release`,
-          leakedAgentTabs.map((t) => t.id),
-        );
-      }
-      shouldRelease = userTabs.length > 0;
-    } catch {
-      // Query failed (e.g. window already gone) — conservatively close it.
-      shouldRelease = false;
-    }
-  }
-
-  if (shouldRelease) {
-    // Keep the window + its user tabs; only drop the session binding.
-    await manager.stop(params.session_id, { dropOnly: true });
-    result.window_released = true;
-  } else {
-    // Window is empty (or we couldn't verify state) — close it.
-    await manager.stop(params.session_id);
-  }
+  // Step 4: close the Agent Window and drop the context.
+  await manager.stop(params.session_id);
 
   return result;
 }

apps/extension/src/tools/shared.ts

@@ -53,11 +53,9 @@ export type { DialogCursor };
 export interface CdpRunner {
   send<T = unknown>(tabId: number, method: string, params?: object): Promise<T>;
   sendToTarget?<T = unknown>(target: CdpTarget, method: string, params?: object): Promise<T>;
-  detach?(tabId: number): Promise<void>;
   getFrameGraph?(tabId: number): Promise<CdpFrameGraph>;
   ensureAttachedToUrl?(tabId: number, expectedUrl: string | undefined): Promise<void>;
   trackSessionTab?(sessionId: string, tabId: number): void;
-  releaseSessionTab?(sessionId: string, tabId: number): Promise<void>;
   onEvent?(handler: (source: CdpDebuggee, method: string, params: unknown) => void): {
     dispose(): void;
   };

apps/extension/src/tools/snapshot-ref.ts

@@ -2,8 +2,9 @@
 // compound CDP node identity via the session RefStore, and build stable
 // `ref_not_found` errors for hard-failure tool paths.
 
+import type { Rect } from "@browser-skill/vom";
 import type { SessionContext } from "@/session-manager/manager";
-import { normaliseRef } from "@/session-manager/ref-store";
+import { normaliseRef, type RefCapability, type RefTargetKind } from "@/session-manager/ref-store";
 import type { RpcError } from "@/transport/types";
 import { rpcError } from "./errors";
 
@@ -12,6 +13,10 @@ export interface SnapshotRefLookup {
   refKey: string;
   frameId?: string;
   cdpSessionId?: string;
+  visibleRect?: Rect;
+  kind: RefTargetKind;
+  capabilities: RefCapability[];
+  generation: number;
 }
 
 function refEntryForTab(ctx: SessionContext, refKey: string, tabId: number) {
@@ -37,6 +42,10 @@ export function lookupSnapshotRef(
     refKey,
     ...(entry.frameId ? { frameId: entry.frameId } : {}),
     ...(entry.cdpSessionId ? { cdpSessionId: entry.cdpSessionId } : {}),
+    ...(entry.visibleRect ? { visibleRect: entry.visibleRect } : {}),
+    kind: entry.kind,
+    capabilities: entry.capabilities,
+    generation: entry.generation,
   };
 }
 
@@ -49,6 +58,7 @@ export function resolveSnapshotRef(
   ctx: SessionContext,
   ref: string,
   tabId: number,
+  requiredCapability?: RefCapability,
 ): SnapshotRefLookup | RpcError {
   const looked = lookupSnapshotRef(ctx, ref, tabId);
   if (looked === null) {
@@ -58,5 +68,13 @@ export function resolveSnapshotRef(
       `ref ${ref} unknown for tab ${tabId} in session ${ctx.sessionId}`,
     );
   }
+  if (requiredCapability && !looked.capabilities.includes(requiredCapability)) {
+    return rpcError(
+      "permission_denied",
+      "ref_capability_denied",
+      `ref ${ref} does not support ${requiredCapability}`,
+      { ref: looked.refKey, required_capability: requiredCapability, kind: looked.kind },
+    );
+  }
   return looked;
 }

apps/extension/src/tools/surface-coordinate.ts

@@ -0,0 +1,88 @@
+import type { Rect } from "@browser-skill/vom";
+import type { Point, Region, ViewportRect } from "./geometry";
+
+const RECT_EPSILON = 0.01;
+
+export function surfaceVisibleRect(
+  liveRect: ViewportRect,
+  observedRect: Rect | undefined,
+): Rect | null {
+  const observed = observedRect ?? {
+    x: liveRect.x,
+    y: liveRect.y,
+    w: liveRect.width,
+    h: liveRect.height,
+  };
+  const x = Math.max(liveRect.x, observed.x);
+  const y = Math.max(liveRect.y, observed.y);
+  const right = Math.min(liveRect.x + liveRect.width, observed.x + observed.w);
+  const bottom = Math.min(liveRect.y + liveRect.height, observed.y + observed.h);
+  return right > x && bottom > y ? { x, y, w: right - x, h: bottom - y } : null;
+}
+
+export function sameRect(a: Rect, b: Rect): boolean {
+  return (
+    Math.abs(a.x - b.x) <= RECT_EPSILON &&
+    Math.abs(a.y - b.y) <= RECT_EPSILON &&
+    Math.abs(a.w - b.w) <= RECT_EPSILON &&
+    Math.abs(a.h - b.h) <= RECT_EPSILON
+  );
+}
+
+export function mapImagePointToViewport(
+  rect: Rect,
+  imageWidth: number,
+  imageHeight: number,
+  imageX: number,
+  imageY: number,
+): Point | null {
+  if (
+    !Number.isFinite(imageX) ||
+    !Number.isFinite(imageY) ||
+    !Number.isSafeInteger(imageWidth) ||
+    !Number.isSafeInteger(imageHeight) ||
+    imageWidth <= 0 ||
+    imageHeight <= 0 ||
+    imageX < 0 ||
+    imageY < 0 ||
+    imageX >= imageWidth ||
+    imageY >= imageHeight ||
+    !Number.isFinite(rect.x) ||
+    !Number.isFinite(rect.y) ||
+    !Number.isFinite(rect.w) ||
+    !Number.isFinite(rect.h) ||
+    rect.w <= 0 ||
+    rect.h <= 0
+  ) {
+    return null;
+  }
+  return {
+    x: rect.x + (imageX / imageWidth) * rect.w,
+    y: rect.y + (imageY / imageHeight) * rect.h,
+  };
+}
+
+function pointInPolygon(point: Point, polygon: Point[]): boolean {
+  for (let index = 0; index < polygon.length; index += 1) {
+    const a = polygon[index];
+    const b = polygon[(index + 1) % polygon.length];
+    const cross = (point.x - a.x) * (b.y - a.y) - (point.y - a.y) * (b.x - a.x);
+    const withinX = point.x >= Math.min(a.x, b.x) - 1e-9 && point.x <= Math.max(a.x, b.x) + 1e-9;
+    const withinY = point.y >= Math.min(a.y, b.y) - 1e-9 && point.y <= Math.max(a.y, b.y) + 1e-9;
+    if (Math.abs(cross) <= 1e-9 && withinX && withinY) return true;
+  }
+  let inside = false;
+  for (let index = 0, previous = polygon.length - 1; index < polygon.length; previous = index++) {
+    const a = polygon[index];
+    const b = polygon[previous];
+    const intersects =
+      a.y > point.y !== b.y > point.y &&
+      point.x < ((b.x - a.x) * (point.y - a.y)) / (b.y - a.y) + a.x;
+    if (intersects) inside = !inside;
+  }
+  return inside;
+}
+
+export function pointInRegion(point: Point, region: Region): boolean {
+  return region.some((polygon) => polygon.length >= 3 && pointInPolygon(point, polygon));
+}

apps/extension/src/tools/surface-target.ts

@@ -0,0 +1,260 @@
+import type { CdpFrameGraph } from "@/browser-driver/frame-graph";
+import type { SessionContext } from "@/session-manager/manager";
+import type { RpcError } from "@/transport/types";
+import { rpcError } from "./errors";
+import { resolveNodeGeometry } from "./frame-geometry";
+import { type CdpRunner, isRpcError, normaliseRef } from "./shared";
+import { resolveSnapshotRef } from "./snapshot-ref";
+import {
+  mapImagePointToViewport,
+  pointInRegion,
+  sameRect,
+  surfaceVisibleRect,
+} from "./surface-coordinate";
+
+export interface SurfaceCaptureEnvironment {
+  navigationIdentity: string;
+  viewportSignature: string;
+  frameProjectionSignature: string;
+}
+
+export interface SurfaceTargetResolverDeps {
+  cdp: CdpRunner;
+  signal?: AbortSignal;
+}
+
+export interface SurfacePointerTargetInput {
+  ref?: string;
+  selector?: string;
+  captureId?: string;
+  imageX?: number;
+  imageY?: number;
+}
+
+export interface ResolvedSurfacePointerTarget {
+  point: { x: number; y: number };
+  usedRef: string;
+  moveSettleMs: number;
+  capture: {
+    id: string;
+    imageX: number;
+    imageY: number;
+  };
+}
+
+// Canvas-style controls commonly update their pointer hit-test on the next
+// animation frame. Leave a scheduling boundary after moving and before pressing.
+export const SURFACE_POINTER_MOVE_SETTLE_MS = 32;
+
+function stale(message: string): RpcError {
+  return rpcError("permission_denied", "surface_capture_stale", message);
+}
+
+function finiteObject(value: unknown): unknown {
+  if (Array.isArray(value)) return value.map(finiteObject);
+  if (typeof value !== "object" || value === null) return value;
+  return Object.fromEntries(
+    Object.entries(value)
+      .filter(
+        ([, entry]) => entry !== undefined && (typeof entry !== "number" || Number.isFinite(entry)),
+      )
+      .sort(([a], [b]) => a.localeCompare(b))
+      .map(([key, entry]) => [key, finiteObject(entry)]),
+  );
+}
+
+function framePathSignature(graph: CdpFrameGraph, frameId: string | undefined): string | null {
+  if (!frameId) return "top";
+  const frames = new Map(graph.frames.map((frame) => [frame.frameId, frame]));
+  const path: Array<{
+    frameId: string;
+    parentFrameId?: string;
+    ownerBackendNodeId?: number;
+    targetSessionId?: string;
+  }> = [];
+  const seen = new Set<string>();
+  let current = frames.get(frameId);
+  while (current) {
+    if (seen.has(current.frameId)) return null;
+    seen.add(current.frameId);
+    path.push({
+      frameId: current.frameId,
+      ...(current.parentFrameId ? { parentFrameId: current.parentFrameId } : {}),
+      ...(current.ownerBackendNodeId !== undefined
+        ? { ownerBackendNodeId: current.ownerBackendNodeId }
+        : {}),
+      ...(current.target.sessionId ? { targetSessionId: current.target.sessionId } : {}),
+    });
+    if (!current.parentFrameId) return JSON.stringify(path);
+    const parent = frames.get(current.parentFrameId);
+    if (!parent) return null;
+    current = parent;
+  }
+  return null;
+}
+
+export async function captureSurfaceEnvironment(
+  cdp: CdpRunner,
+  tabId: number,
+  frameId: string | undefined,
+): Promise<SurfaceCaptureEnvironment | RpcError> {
+  try {
+    const [tree, metrics, graph] = await Promise.all([
+      cdp.send<{
+        frameTree?: { frame?: { id?: string; loaderId?: string; url?: string } };
+      }>(tabId, "Page.getFrameTree", {}),
+      cdp.send<Record<string, unknown>>(tabId, "Page.getLayoutMetrics", {}),
+      frameId && cdp.getFrameGraph ? cdp.getFrameGraph(tabId) : Promise.resolve(null),
+    ]);
+    const root = tree.frameTree?.frame;
+    if (!root?.id || !root.loaderId) {
+      return { code: "cdp_failed", message: "could not identify the current navigation" };
+    }
+    const frameProjectionSignature = frameId
+      ? graph
+        ? framePathSignature(graph, frameId)
+        : null
+      : "top";
+    if (!frameProjectionSignature) {
+      return { code: "cdp_failed", message: `could not identify frame projection for ${frameId}` };
+    }
+    return {
+      navigationIdentity: JSON.stringify({ id: root.id, loaderId: root.loaderId, url: root.url }),
+      viewportSignature: JSON.stringify(
+        finiteObject({
+          cssLayoutViewport: metrics.cssLayoutViewport,
+          cssVisualViewport: metrics.cssVisualViewport,
+        }),
+      ),
+      frameProjectionSignature,
+    };
+  } catch (error) {
+    return { code: "cdp_failed", message: error instanceof Error ? error.message : String(error) };
+  }
+}
+
+function validatePointParams(input: SurfacePointerTargetInput): RpcError | null {
+  if (!input.captureId || typeof input.captureId !== "string") {
+    return { code: "invalid_params", message: "surface point click requires capture_id" };
+  }
+  if (typeof input.imageX !== "number" || !Number.isFinite(input.imageX)) {
+    return { code: "invalid_params", message: "image_x must be a finite number" };
+  }
+  if (typeof input.imageY !== "number" || !Number.isFinite(input.imageY)) {
+    return { code: "invalid_params", message: "image_y must be a finite number" };
+  }
+  if (!input.ref || input.selector) {
+    return {
+      code: "invalid_params",
+      message: "surface point click requires one Surface ref and does not accept a selector",
+    };
+  }
+  return null;
+}
+
+function captureConsumeError(reason: "not_found" | "expired" | "consumed"): RpcError {
+  const code = reason === "not_found" ? "not_found" : "permission_denied";
+  return rpcError(code, `surface_capture_${reason}`, `surface capture is ${reason}`);
+}
+
+function aborted(signal: AbortSignal | undefined): RpcError | null {
+  return signal?.aborted
+    ? { code: "cancelled", message: "surface target resolution aborted" }
+    : null;
+}
+
+/** Resolve one fresh Surface capture coordinate into a validated top-viewport pointer target. */
+export async function resolveSurfacePointerTarget(
+  ctx: SessionContext,
+  target: { tabId: number; url?: string },
+  input: SurfacePointerTargetInput,
+  deps: SurfaceTargetResolverDeps,
+): Promise<ResolvedSurfacePointerTarget | RpcError> {
+  const invalid = validatePointParams(input);
+  if (invalid) return invalid;
+  const earlyAbort = aborted(deps.signal);
+  if (earlyAbort) return earlyAbort;
+
+  const consumed = ctx.surfaceCaptures.consume(input.captureId as string);
+  if (!consumed.ok) return captureConsumeError(consumed.reason);
+  const capture = consumed.capture;
+  const postConsumeAbort = aborted(deps.signal);
+  if (postConsumeAbort) return postConsumeAbort;
+
+  if (capture.sessionId !== ctx.sessionId || capture.tabId !== target.tabId) {
+    return stale("surface capture belongs to a different session or tab");
+  }
+  if (normaliseRef(input.ref as string) !== capture.surface.ref) {
+    return stale("surface ref does not match the capture");
+  }
+  const node = resolveSnapshotRef(ctx, input.ref as string, target.tabId, "screenshot");
+  if (isRpcError(node)) return stale(node.message);
+  if (
+    node.kind !== "surface" ||
+    node.backendNodeId !== capture.surface.backendNodeId ||
+    node.frameId !== capture.surface.frameId ||
+    node.generation !== capture.surface.observationGeneration
+  ) {
+    return stale("surface observation generation or node identity changed");
+  }
+
+  deps.cdp.trackSessionTab?.(ctx.sessionId, target.tabId);
+  try {
+    await deps.cdp.ensureAttachedToUrl?.(target.tabId, target.url);
+  } catch (error) {
+    return { code: "cdp_failed", message: error instanceof Error ? error.message : String(error) };
+  }
+  const environment = await captureSurfaceEnvironment(deps.cdp, target.tabId, node.frameId);
+  if (isRpcError(environment)) return environment;
+  if (
+    environment.navigationIdentity !== capture.navigationIdentity ||
+    environment.viewportSignature !== capture.viewportSignature ||
+    environment.frameProjectionSignature !== capture.frameProjectionSignature
+  ) {
+    return stale("navigation, viewport, zoom, scroll, or frame projection changed");
+  }
+
+  const geometry = await resolveNodeGeometry(
+    deps.cdp,
+    target.tabId,
+    {
+      target: {
+        tabId: target.tabId,
+        ...(node.cdpSessionId ? { sessionId: node.cdpSessionId } : {}),
+      },
+      backendNodeId: node.backendNodeId,
+      ...(node.frameId ? { frameId: node.frameId } : {}),
+    },
+    { scrollIntoView: false },
+  );
+  if (isRpcError(geometry)) return stale(geometry.message);
+  const currentRect = surfaceVisibleRect(geometry.topBounds, node.visibleRect);
+  if (!currentRect || !sameRect(currentRect, capture.topViewportRect)) {
+    return stale("surface visible region changed");
+  }
+  const point = mapImagePointToViewport(
+    capture.topViewportRect,
+    capture.imageWidth,
+    capture.imageHeight,
+    input.imageX as number,
+    input.imageY as number,
+  );
+  if (!point || !pointInRegion(point, geometry.topVisibleRegions)) {
+    return rpcError(
+      "invalid_params",
+      "surface_coordinate_invalid",
+      "image coordinate is outside the captured Surface",
+    );
+  }
+
+  return {
+    point,
+    usedRef: capture.surface.ref,
+    moveSettleMs: SURFACE_POINTER_MOVE_SETTLE_MS,
+    capture: {
+      id: capture.id,
+      imageX: input.imageX as number,
+      imageY: input.imageY as number,
+    },
+  };
+}

apps/extension/src/tools/tabs.ts

@@ -10,7 +10,7 @@ import {
 import type { SessionContext, SessionManager } from "@/session-manager/manager";
 import type { RpcError } from "@/transport/types";
 import { rpcError } from "./errors";
-import { type CdpRunner, isRpcError, lookupSession } from "./shared";
+import { isRpcError, lookupSession } from "./shared";
 
 export type TabScope = "user" | "agent" | "all";
 
@@ -287,10 +287,6 @@ export interface TabManagementDeps {
   approveBorrow?: BorrowConfirmationApprover;
   /** Clears Agent-scoped overlays after a borrowed tab is returned. */
   agentOverlayReset?: AgentOverlayResetApi;
-  /** Releases this session's CDP claim after a borrowed tab is returned. */
-  cdp?: Pick<CdpRunner, "releaseSessionTab">;
-  /** Runs after tab_return validation, before moving the borrowed tab. */
-  beforeReturn?: (sessionId: string, tabId: number) => Promise<void>;
   /**
    * Reports whether `windowId` is any live session's Agent Window.
    * `tab_return`'s fallback window picker uses this to avoid moving a
@@ -373,7 +369,6 @@ function buildCreateProps(
  * Returns the created tab on success, or an `RpcError` on failure.
  */
 async function createTabAndCleanup(
-  ctx: SessionContext,
   deps: TabManagementDeps,
   createProps: chrome.tabs.CreateProperties,
 ): Promise<CreatedChromeTab | RpcError> {
@@ -386,28 +381,25 @@ async function createTabAndCleanup(
       message: err instanceof Error ? err.message : String(err),
     };
   }
-  if (typeof tab.id !== "number") {
-    return { code: "protocol_error", message: "chrome.tabs.create returned no tab id" };
-  }
-  // Claim the concrete id returned by Chrome. Ownership never depends on
-  // matching this request to an asynchronous onCreated event.
-  ctx.agentCreatedTabs.add(tab.id);
   if (aborted(deps.signal, "tab_create")) {
-    try {
-      await getTabsApi(deps).remove(tab.id);
-      ctx.agentCreatedTabs.delete(tab.id);
-    } catch (cleanupErr) {
-      // Keep the claim when cleanup fails so session_stop can retry instead
-      // of releasing an agent-owned tab to the user.
-      return rpcError(
-        "protocol_error",
-        "cleanup_failed",
-        `tab_create aborted but cleanup of tab ${tab.id} failed: ${describeError(cleanupErr)}`,
-        { resource_type: "tab", resource_id: tab.id },
-      );
+    // We already opened the tab; close it on abort so we don't leak.
+    if (typeof tab.id === "number") {
+      try {
+        await getTabsApi(deps).remove(tab.id);
+      } catch (cleanupErr) {
+        return rpcError(
+          "protocol_error",
+          "cleanup_failed",
+          `tab_create aborted but cleanup of tab ${tab.id} failed: ${describeError(cleanupErr)}`,
+          { resource_type: "tab", resource_id: tab.id },
+        );
+      }
     }
     return { code: "cancelled", message: "tab_create aborted" };
   }
+  if (typeof tab.id !== "number") {
+    return { code: "protocol_error", message: "chrome.tabs.create returned no tab id" };
+  }
   return { ...tab, id: tab.id };
 }
 
@@ -434,7 +426,7 @@ export async function handleTabCreate(
   const paramErr = validateTabCreateParams(params);
   if (paramErr) return paramErr;
 
-  const tab = await createTabAndCleanup(ctx, deps, buildCreateProps(ctx, params));
+  const tab = await createTabAndCleanup(deps, buildCreateProps(ctx, params));
   if (isRpcError(tab)) return tab;
 
   return {
@@ -534,9 +526,6 @@ export async function handleTabClose(
   }
   try {
     await getTabsApi(deps).remove(params.tab_id);
-    // Keep the tracking set accurate so session_stop won't try to close a
-    // tab that's already gone (design §3.1).
-    ctx.agentCreatedTabs.delete(params.tab_id);
   } catch (err) {
     return {
       code: "protocol_error",
@@ -764,6 +753,13 @@ async function executeBorrowCore(
   );
   if (moveErr) return moveErr;
 
+  // Best-effort activation — failure is non-fatal.
+  try {
+    await p.tabsApi.update(p.tabId, { active: true });
+  } catch (err) {
+    console.debug("[bsk tab_borrow] activate after move failed", err);
+  }
+
   return { originalWindowId, originalIndex };
 }
 
@@ -827,12 +823,6 @@ export async function handleTabBorrow(
         message: err instanceof Error ? err.message : String(err),
       };
     }
-    // Activate after commit so overlay/event observers see the tab as claimed.
-    try {
-      await tabsApi.update(params.tab_id, { active: true });
-    } catch (err) {
-      console.debug("[bsk tab_borrow] activate after move failed", err);
-    }
     return {
       tab_id: params.tab_id,
       original_window_id: coreResult.originalWindowId,
@@ -949,22 +939,16 @@ async function cleanupUnusedFallbackWindow(
   }
 }
 
-async function releaseReturnedTabState(
+function resetAgentOverlaysInReturnedTab(
   ctx: SessionContext,
   tabId: number,
   deps: TabManagementDeps,
-): Promise<void> {
+): void {
   void getAgentOverlayResetApi(deps)
     .resetAgentOverlays(tabId, ctx.sessionId)
     .catch((err) => {
       console.debug("[bsk tab_return] agent overlay reset failed", err);
     });
-  try {
-    await deps.cdp?.releaseSessionTab?.(ctx.sessionId, tabId);
-  } catch (err) {
-    // A successful move must not be retried because debugger cleanup failed.
-    console.debug("[bsk tab_return] CDP release failed", err);
-  }
 }
 
 /**
@@ -1038,7 +1022,7 @@ export async function returnBorrowedTab(
     });
     const movedTab = Array.isArray(moved) ? moved[0] : moved;
     const finalIndex = typeof movedTab?.index === "number" ? movedTab.index : targetIndex;
-    await releaseReturnedTabState(ctx, tabId, deps);
+    resetAgentOverlaysInReturnedTab(ctx, tabId, deps);
     return {
       tabId,
       toWindowId: targetWindowId,
@@ -1080,7 +1064,7 @@ export async function returnBorrowedTab(
         });
         const movedTab = Array.isArray(moved) ? moved[0] : moved;
         const finalIndex = typeof movedTab?.index === "number" ? movedTab.index : target.index;
-        await releaseReturnedTabState(ctx, tabId, deps);
+        resetAgentOverlaysInReturnedTab(ctx, tabId, deps);
         return {
           tabId,
           toWindowId: target.windowId,
@@ -1126,7 +1110,6 @@ export async function handleTabReturn(
       message: `tab_return: tab ${params.tab_id} is not borrowed by this session`,
     };
   }
-  await deps.beforeReturn?.(ctx.sessionId, params.tab_id);
   const outcome = await returnBorrowedTab(ctx, params.tab_id, {
     ...deps,
     isAgentWindowId:

apps/extension/src/tools/text-input.ts

@@ -0,0 +1,194 @@
+// Browser-native text entry for live editing sessions. Unlike fill, this
+// module does not assume that the focused editable retains a DOM value: a
+// Canvas editor may use an input/textarea only as a transient keyboard/IME
+// sink and clear it after consuming each event.
+
+import type { RpcError } from "@/transport/types";
+import type { CdpRunner } from "./shared";
+
+function cancelled(signal: AbortSignal | undefined): RpcError | null {
+  return signal?.aborted ? { code: "cancelled", message: "type aborted" } : null;
+}
+
+interface KeyDescriptor {
+  key: string;
+  code: string;
+  windowsVirtualKeyCode: number;
+  nativeVirtualKeyCode?: number;
+  modifiers: number;
+}
+
+function asciiKeyDescriptor(character: string): KeyDescriptor {
+  if (/^[a-z]$/.test(character)) {
+    return {
+      key: character,
+      code: `Key${character.toUpperCase()}`,
+      windowsVirtualKeyCode: character.toUpperCase().charCodeAt(0),
+      modifiers: 0,
+    };
+  }
+  if (/^[A-Z]$/.test(character)) {
+    return {
+      key: character,
+      code: `Key${character}`,
+      windowsVirtualKeyCode: character.charCodeAt(0),
+      modifiers: 8,
+    };
+  }
+  if (/^[0-9]$/.test(character)) {
+    return {
+      key: character,
+      code: `Digit${character}`,
+      windowsVirtualKeyCode: character.charCodeAt(0),
+      modifiers: 0,
+    };
+  }
+  if (character === " ") {
+    return { key: " ", code: "Space", windowsVirtualKeyCode: 32, modifiers: 0 };
+  }
+  return {
+    key: character,
+    code: "",
+    windowsVirtualKeyCode: character.charCodeAt(0),
+    modifiers: 0,
+  };
+}
+
+/**
+ * Run one complete browser key transaction and always release the key. The
+ * commit callback owns the text-producing phase: a `char` event for a
+ * keyboard character, or `Input.insertText` for committed IME text.
+ */
+async function dispatchKeyTransaction(
+  cdp: CdpRunner,
+  tabId: number,
+  descriptor: KeyDescriptor,
+  commit: () => Promise<void>,
+  signal?: AbortSignal,
+): Promise<RpcError | null> {
+  let keyDown = false;
+  try {
+    await cdp.send(tabId, "Input.dispatchKeyEvent", {
+      type: "rawKeyDown",
+      ...descriptor,
+    });
+    keyDown = true;
+    const abortedAfterDown = cancelled(signal);
+    if (!abortedAfterDown) await commit();
+    await cdp.send(tabId, "Input.dispatchKeyEvent", {
+      type: "keyUp",
+      ...descriptor,
+    });
+    keyDown = false;
+    return abortedAfterDown ?? cancelled(signal);
+  } catch (err) {
+    if (keyDown) {
+      try {
+        await cdp.send(tabId, "Input.dispatchKeyEvent", {
+          type: "keyUp",
+          ...descriptor,
+        });
+      } catch {
+        // Best-effort cleanup keeps the page from observing a stuck key.
+      }
+    }
+    throw err;
+  }
+}
+
+async function typeAsciiCharacter(
+  cdp: CdpRunner,
+  tabId: number,
+  character: string,
+  signal?: AbortSignal,
+): Promise<RpcError | null> {
+  const descriptor = asciiKeyDescriptor(character);
+  return dispatchKeyTransaction(
+    cdp,
+    tabId,
+    descriptor,
+    async () => {
+      await cdp.send(tabId, "Input.dispatchKeyEvent", {
+        type: "char",
+        ...descriptor,
+        text: character,
+        unmodifiedText: character,
+      });
+    },
+    signal,
+  );
+}
+
+async function typeImeSegment(
+  cdp: CdpRunner,
+  tabId: number,
+  text: string,
+  signal?: AbortSignal,
+): Promise<RpcError | null> {
+  const abortedBefore = cancelled(signal);
+  if (abortedBefore) return abortedBefore;
+
+  // A committed IME value is still part of a keyboard transaction. The
+  // Process key (VK_PROCESSKEY / 229) is the platform-neutral signal that an
+  // IME owns the keystroke. It lets selection-based editors enter text mode
+  // before insertText commits the Unicode value, without inventing a visible
+  // ASCII prefix or leaving a composition transaction open.
+  const processKey: KeyDescriptor = {
+    key: "Process",
+    code: "",
+    windowsVirtualKeyCode: 229,
+    nativeVirtualKeyCode: 229,
+    modifiers: 0,
+  };
+  return dispatchKeyTransaction(
+    cdp,
+    tabId,
+    processKey,
+    () => cdp.send(tabId, "Input.insertText", { text }).then(() => undefined),
+    signal,
+  );
+}
+
+/**
+ * Dispatch text through the browser's keyboard/IME pipelines. ASCII runs are
+ * delivered as physical-key-shaped events; non-ASCII runs use CDP's committed
+ * IME/software-keyboard insertion path. Successful return means the browser
+ * accepted the input commands, not that a DOM value retained the text.
+ */
+export async function dispatchTextInput(
+  cdp: CdpRunner,
+  tabId: number,
+  text: string,
+  signal?: AbortSignal,
+): Promise<RpcError | null> {
+  const abortedBefore = cancelled(signal);
+  if (abortedBefore) return abortedBefore;
+
+  const segments: Array<{ ime: boolean; text: string }> = [];
+  for (const character of text) {
+    const ime = character.codePointAt(0)! > 0x7f;
+    const previous = segments.at(-1);
+    if (previous?.ime === ime) previous.text += character;
+    else segments.push({ ime, text: character });
+  }
+
+  try {
+    for (const segment of segments) {
+      if (segment.ime) {
+        const error = await typeImeSegment(cdp, tabId, segment.text, signal);
+        if (error) return error;
+        continue;
+      }
+      for (const character of segment.text) {
+        const error = await typeAsciiCharacter(cdp, tabId, character, signal);
+        if (error) return error;
+      }
+    }
+    return null;
+  } catch (err) {
+    return {
+      code: "cdp_failed",
+      message: err instanceof Error ? err.message : String(err),
+    };
+  }
+}

apps/extension/src/tools/transfer-transaction.ts

@@ -1,42 +0,0 @@
-// Shared time boundary for browser-side file-transfer transactions. Concrete
-// upload/download mechanisms own their state machines and effect semantics.
-
-export class BoundedWaitError extends Error {
-  constructor(
-    readonly kind: "timeout" | "aborted",
-    message: string,
-  ) {
-    super(message);
-  }
-}
-
-export function remainingMs(deadline: number): number {
-  return Math.max(0, deadline - Date.now());
-}
-
-export async function waitBounded<T>(
-  promise: Promise<T>,
-  deadline: number,
-  signal: AbortSignal | undefined,
-  timeoutMessage: string,
-): Promise<T> {
-  const remaining = remainingMs(deadline);
-  if (signal?.aborted) throw new BoundedWaitError("aborted", "file transfer aborted");
-  if (remaining === 0) throw new BoundedWaitError("timeout", timeoutMessage);
-
-  let timer: ReturnType<typeof setTimeout> | undefined;
-  let onAbort: (() => void) | undefined;
-  const boundary = new Promise<never>((_resolve, reject) => {
-    timer = setTimeout(() => reject(new BoundedWaitError("timeout", timeoutMessage)), remaining);
-    if (signal) {
-      onAbort = () => reject(new BoundedWaitError("aborted", "file transfer aborted"));
-      signal.addEventListener("abort", onAbort, { once: true });
-    }
-  });
-  try {
-    return await Promise.race([promise, boundary]);
-  } finally {
-    if (timer) clearTimeout(timer);
-    if (signal && onAbort) signal.removeEventListener("abort", onAbort);
-  }
-}

apps/extension/src/tools/upload.ts

@@ -1,85 +0,0 @@
-// Upload orchestration: validate the session-scoped request, resolve its click
-// target, then delegate the browser protocol transaction to the file-input
-// transaction module.
-
-import {
-  type CaptureSuppressSendToTab,
-  withExtensionOverlayHidden,
-} from "@/lib/capture-suppress-bridge";
-import type { SessionManager } from "@/session-manager/manager";
-import type { RpcError, UploadParams, UploadResult } from "@/transport/types";
-import { uploadThroughFileDrop } from "./file-drop-transaction";
-import { uploadThroughActivatedFileInput } from "./file-input-transaction";
-import { clickResolvedTarget, type InteractionDeps, resolveActionTarget } from "./interaction";
-import {
-  type CdpRunner,
-  enforceAgentWindow,
-  isRpcError,
-  lookupSession,
-  resolveTargetTab,
-} from "./shared";
-
-const DEFAULT_TIMEOUT_MS = 120_000;
-
-export interface UploadDeps extends InteractionDeps {
-  cdp: CdpRunner;
-  sendToTab?: CaptureSuppressSendToTab;
-}
-
-export async function handleUpload(
-  manager: SessionManager,
-  params: UploadParams,
-  deps: UploadDeps,
-): Promise<UploadResult | RpcError> {
-  const ctx = lookupSession(manager, params, "upload");
-  if (isRpcError(ctx)) return ctx;
-  const target = await resolveTargetTab(manager, ctx, params.tab_id, deps.tabsApi);
-  if (isRpcError(target)) return target;
-  const denied = enforceAgentWindow(ctx, target, "upload");
-  if (denied) return denied;
-  if (
-    params.files.length === 0 ||
-    params.files.length > 20 ||
-    params.files.some((file) => !file.staged_path)
-  ) {
-    return { code: "invalid_params", message: "upload requires daemon-staged files" };
-  }
-  const address = await resolveActionTarget(deps.cdp, ctx, target, params, "upload");
-  if (isRpcError(address)) return address;
-  const timeoutMs = params.timeout_ms ?? DEFAULT_TIMEOUT_MS;
-  const files = params.files.map((file) => file.staged_path as string);
-  const mode = params.mode ?? "input";
-  let usedRef = address.usedRef;
-  let usedSelector = address.usedSelector;
-  if (mode === "drop") {
-    deps.cdp.trackSessionTab?.(ctx.sessionId, target.tabId);
-    const transaction = await uploadThroughFileDrop({
-      cdp: deps.cdp,
-      actionTarget: address,
-      files,
-      timeoutMs,
-      signal: deps.signal,
-      withOverlayHidden: (operation) =>
-        withExtensionOverlayHidden(target.tabId, operation, deps.sendToTab),
-    });
-    if (isRpcError(transaction)) return transaction;
-  } else {
-    const transaction = await uploadThroughActivatedFileInput({
-      cdp: deps.cdp,
-      actionTarget: address,
-      files,
-      timeoutMs,
-      signal: deps.signal,
-      trigger: () => clickResolvedTarget(ctx, address, {}, deps),
-    });
-    if (isRpcError(transaction)) return transaction;
-    usedRef = transaction.click.used_ref;
-    usedSelector = transaction.click.used_selector;
-  }
-  return {
-    tab_id: target.tabId,
-    used_ref: usedRef,
-    used_selector: usedSelector,
-    file_names: params.files.map((file) => file.name),
-  };
-}

apps/extension/src/tools/vom/__tests__/capture.test.ts

@@ -1,6 +1,6 @@
 import { describe, expect, it, vi } from "vitest";
 import { OVERLAY_HOST_MARKER_ATTR, OVERLAY_HOST_NAME } from "../../../lib/overlay-bridge";
-import { captureViewModel, collectOverlayExcludedBackendIds, probeHoverSurfaces } from "../capture";
+import { captureViewModel, collectOverlayExcludedBackendIds } from "../capture";
 
 // Minimal but format-accurate captureSnapshot reply: a body with one
 // fixed full-screen overlay div carrying a password input.
@@ -215,29 +215,21 @@ function makeCdp(snapshot: unknown) {
       if (method === "Runtime.evaluate") {
         return {
           result: {
-            deepSerializedValue: {
-              type: "array",
-              value: [
+            value: {
+              controls: [
                 {
-                  type: "array",
-                  value: [
-                    { type: "node", value: { backendNodeId: 13 } },
-                    {
-                      type: "string",
-                      value: JSON.stringify({
-                        state: "filled",
-                        sensitive: true,
-                        placeholder: "secret",
-                      }),
-                    },
-                  ],
+                  state: "filled",
+                  sensitive: true,
+                  defaultValue: "hunter2",
+                  value: "hunter2",
+                  placeholder: "secret",
                 },
               ],
+              childFrames: [],
             },
           },
         };
       }
-      if (method === "Runtime.releaseObjectGroup") return {};
       throw new Error(`unexpected ${method}`);
     }) as unknown as <T>(tabId: number, method: string, params?: object) => Promise<T>,
   };
@@ -259,10 +251,11 @@ describe("captureViewModel", () => {
     expect(input?.attrs).toEqual({ type: "password" });
   });
 
-  it("never hovers the page while capturing", async () => {
+  it("does not run conditional surface probes by default", async () => {
     const cdp = makeCdp(fakeSnapshotReply());
-    await captureViewModel(cdp, 4);
+    const result = await captureViewModel(cdp, 4);
 
+    expect(result.surfaceProbes).toEqual([]);
     expect(cdp.send).not.toHaveBeenCalledWith(4, "Input.dispatchMouseEvent", expect.anything());
   });
 
@@ -281,11 +274,8 @@ describe("captureViewModel", () => {
     expect(cdp.send).toHaveBeenCalledWith(
       4,
       "Runtime.evaluate",
-      expect.objectContaining({
-        serializationOptions: expect.objectContaining({ serialization: "deep" }),
-      }),
+      expect.objectContaining({ returnByValue: true }),
     );
-    expect(cdp.send).toHaveBeenCalledWith(4, "Runtime.releaseObjectGroup", expect.anything());
     expect(cdp.send).not.toHaveBeenCalledWith(4, "DOM.resolveNode", expect.anything());
     expect(cdp.send).not.toHaveBeenCalledWith(4, "Runtime.callFunctionOn", expect.anything());
   });
@@ -329,10 +319,9 @@ describe("captureViewModel", () => {
       }) as unknown as <T>(tabId: number, method: string, params?: object) => Promise<T>,
     };
 
-    const captured = await captureViewModel(cdp, 4);
-    const surfaceProbes = await probeHoverSurfaces(cdp, 4, captured.nodes);
+    const result = await captureViewModel(cdp, 4, { conditionalSurfaceProbe: true });
 
-    expect(surfaceProbes).toEqual([
+    expect(result.surfaceProbes).toEqual([
       {
         triggerBackendNodeId: 12,
         triggerPoint: { x: 956, y: 32 },
@@ -389,10 +378,9 @@ describe("captureViewModel", () => {
       }) as unknown as <T>(tabId: number, method: string, params?: object) => Promise<T>,
     };
 
-    const captured = await captureViewModel(cdp, 4);
-    const surfaceProbes = await probeHoverSurfaces(cdp, 4, captured.nodes);
+    const result = await captureViewModel(cdp, 4, { conditionalSurfaceProbe: true });
 
-    expect(surfaceProbes).toEqual([
+    expect(result.surfaceProbes).toEqual([
       expect.objectContaining({
         triggerBackendNodeId: 12,
         subItems: ["My profile", "Sign out"],
@@ -434,8 +422,7 @@ describe("captureViewModel", () => {
       }) as unknown as <T>(tabId: number, method: string, params?: object) => Promise<T>,
     };
 
-    const captured = await captureViewModel(cdp, 4);
-    await probeHoverSurfaces(cdp, 4, captured.nodes);
+    await captureViewModel(cdp, 4, { conditionalSurfaceProbe: true });
 
     expect(hoverMoves).toBe(1);
   });
@@ -735,6 +722,44 @@ describe("captureViewModel", () => {
     expect(nodes.find((node) => node.backendNodeId === 14)?.rendered).toBe(false);
   });
 
+  it("propagates fully transparent ancestor suppression to painted descendants", async () => {
+    const S = ["html", "body", "div", "canvas", "static", "auto", "visible", "1", "0"];
+    const i = (value: string) => S.indexOf(value);
+    const snapshot = {
+      strings: S,
+      documents: [
+        {
+          nodes: {
+            parentIndex: [-1, 0, 1, 2],
+            nodeName: [i("html"), i("body"), i("div"), i("canvas")],
+            backendNodeId: [10, 11, 12, 13],
+            attributes: [[], [], [], []],
+          },
+          layout: {
+            nodeIndex: [1, 2, 3],
+            styles: [
+              [i("static"), i("auto"), i("auto"), i("visible"), i("1")],
+              [i("static"), i("auto"), i("auto"), i("visible"), i("0")],
+              [i("static"), i("auto"), i("auto"), i("visible"), i("1")],
+            ],
+            bounds: [
+              [0, 0, 1000, 800],
+              [10, 10, 200, 100],
+              [10, 10, 200, 100],
+            ],
+            paintOrders: [0, 1, 2],
+          },
+        },
+      ],
+    };
+
+    const { nodes } = await captureViewModel(makeCdp(snapshot), 4);
+    const canvas = nodes.find((node) => node.backendNodeId === 13);
+
+    expect(canvas?.rendered).toBe(true);
+    expect(canvas?.visuallySuppressed).toBe(true);
+  });
+
   it("subtracts scroll offset to make rects viewport-relative", async () => {
     const cdp = {
       send: vi.fn(async (_t: number, method: string) => {

apps/extension/src/tools/vom/__tests__/form-state.test.ts

@@ -1,270 +0,0 @@
-import { describe, expect, it, vi } from "vitest";
-import { type CdpRunner, cdpRunnerForTarget } from "../../shared";
-import { captureViewModel } from "../capture";
-
-interface Control {
-  id: number;
-  value: string;
-  type?: string;
-}
-
-function snapshot(groups: Control[][]) {
-  const strings: string[] = [];
-  const str = (value: string) => {
-    let index = strings.indexOf(value);
-    if (index < 0) index = strings.push(value) - 1;
-    return index;
-  };
-  const documents = groups.map((controls, groupIndex) => ({
-    frameId: `frame-${groupIndex}`,
-    nodes: {
-      parentIndex: [-1, 0, ...controls.map(() => 1)],
-      nodeName: [str("html"), str("body"), ...controls.map(() => str("input"))],
-      backendNodeId: [8000 + groupIndex * 2, 8001 + groupIndex * 2, ...controls.map((c) => c.id)],
-      attributes: [
-        [],
-        [],
-        ...controls.map((c) => [
-          str("type"),
-          str(c.type ?? "text"),
-          str("placeholder"),
-          str("snapshot placeholder"),
-        ]),
-      ],
-      inputValue: {
-        index: controls.map((_, i) => i + 2),
-        value: controls.map((c) => str(c.value)),
-      },
-      contentDocumentIndex: { index: [] as number[], value: [] as number[] },
-    },
-  }));
-  const root = documents[0].nodes;
-  for (let i = 1; i < documents.length; i++) {
-    root.contentDocumentIndex.index.push(root.backendNodeId.length);
-    root.contentDocumentIndex.value.push(i);
-    root.parentIndex.push(1);
-    root.nodeName.push(str("iframe"));
-    root.backendNodeId.push(9000 + i);
-    root.attributes.push([]);
-  }
-  return { strings, documents };
-}
-
-function state(value: string) {
-  return {
-    value,
-    defaultValue: "",
-    placeholder: "live placeholder",
-    state: value ? "filled" : "empty",
-    sensitive: false,
-  };
-}
-
-function reply(entries: Array<[number, unknown]>) {
-  return {
-    result: {
-      deepSerializedValue: {
-        type: "array",
-        value: entries.map(([backendNodeId, fields]) => ({
-          type: "array",
-          value: [
-            { type: "node", value: { backendNodeId } },
-            { type: "string", value: JSON.stringify(fields) },
-          ],
-        })),
-      },
-    },
-  };
-}
-
-function fakeCdp(groups: Control[][], evaluate: (params: Record<string, unknown>) => unknown) {
-  const snap = snapshot(groups);
-  const send = vi.fn(async (_tabId: number, method: string, params?: object) => {
-    if (method === "DOMSnapshot.captureSnapshot") return snap;
-    if (method === "Runtime.evaluate") return evaluate(params as Record<string, unknown>);
-    return {};
-  });
-  return { send, cdp: { send: send as CdpRunner["send"] } };
-}
-
-describe("form state identity", () => {
-  it("matches reordered controls by identity and preserves controls missing from the batch", async () => {
-    // The snapshot includes a shadow input. The runtime query skips it,
-    // encounters a new input, and sees the remaining inputs in a new order.
-    const { cdp } = fakeCdp(
-      [
-        [
-          { id: 10, value: "" },
-          { id: 11, value: "old email" },
-          { id: 12, value: "old city" },
-        ],
-      ],
-      () =>
-        reply([
-          [99, state("new control")],
-          [12, state("上海")],
-          [11, state("alice@example.com")],
-        ]),
-    );
-    const captured = await captureViewModel(cdp, 7);
-    expect(captured.nodes.find((n) => n.backendNodeId === 10)).toMatchObject({
-      formValue: "",
-      formState: "empty",
-      formPlaceholder: "snapshot placeholder",
-    });
-    expect(captured.nodes.find((n) => n.backendNodeId === 11)).toMatchObject({
-      formValue: "alice@example.com",
-    });
-    expect(captured.nodes.find((n) => n.backendNodeId === 12)).toMatchObject({ formValue: "上海" });
-    expect(captured.nodes.some((n) => n.backendNodeId === 99)).toBe(false);
-  });
-
-  it("does not shift state between frames when one frame is absent from the runtime batch", async () => {
-    const { cdp } = fakeCdp(
-      [
-        [],
-        [{ id: 11, value: "inaccessible frame" }],
-        [{ id: 21, value: "old second" }],
-        [{ id: 31, value: "old third" }],
-      ],
-      () =>
-        reply([
-          [31, state("third frame")],
-          [21, state("second frame")],
-        ]),
-    );
-    const captured = await captureViewModel(cdp, 7);
-    const controls = [...captured.iframeNodes.values()].flat();
-    expect(controls.find((n) => n.backendNodeId === 11)?.formValue).toBe("inaccessible frame");
-    expect(controls.find((n) => n.backendNodeId === 21)?.formValue).toBe("second frame");
-    expect(controls.find((n) => n.backendNodeId === 31)?.formValue).toBe("third frame");
-  });
-
-  it("keeps identical backend ids in different OOPIF targets separate", async () => {
-    const sendToTarget = vi.fn(async (target, method) => {
-      if (method === "DOMSnapshot.captureSnapshot") return snapshot([[{ id: 11, value: "old" }]]);
-      if (method === "Runtime.evaluate") return reply([[11, state(target.sessionId)]]);
-      return {};
-    });
-    const send = vi.fn(async () => {
-      throw new Error("unexpected top-level target");
-    });
-    const cdp = { send, sendToTarget } as unknown as CdpRunner;
-    for (const sessionId of ["left-frame", "right-frame"]) {
-      const captured = await captureViewModel(cdpRunnerForTarget(cdp, { tabId: 7, sessionId }), 7);
-      expect(captured.nodes.find((n) => n.backendNodeId === 11)?.formValue).toBe(sessionId);
-      expect(sendToTarget).toHaveBeenCalledWith(
-        { tabId: 7, sessionId },
-        "Runtime.releaseObjectGroup",
-        expect.anything(),
-      );
-    }
-    expect(send).not.toHaveBeenCalled();
-  });
-
-  it.each([
-    "unsupported",
-    "legacy",
-    "malformed",
-    "invalid-state",
-  ])("preserves snapshot data when enrichment is %s", async (mode) => {
-    const { cdp, send } = fakeCdp([[{ id: 11, value: "snapshot value" }]], () => {
-      if (mode === "unsupported") throw new Error("serialization unsupported");
-      if (mode === "legacy") return { result: { value: { controls: [state("wrong value")] } } };
-      const result = reply([
-        [11, mode === "invalid-state" ? { value: "wrong value" } : state("wrong value")],
-      ]);
-      if (mode === "malformed")
-        result.result.deepSerializedValue.value[0].value[1].value = "not json";
-      return result;
-    });
-    const captured = await captureViewModel(cdp, 7);
-    expect(captured.nodes.find((n) => n.backendNodeId === 11)).toMatchObject({
-      formValue: "snapshot value",
-      formPlaceholder: "snapshot placeholder",
-    });
-    expect(send).toHaveBeenCalledWith(7, "Runtime.releaseObjectGroup", expect.anything());
-  });
-
-  it("releases the batch's remote objects when capture is cancelled", async () => {
-    const controller = new AbortController();
-    const { cdp, send } = fakeCdp([[{ id: 11, value: "old" }]], () => {
-      controller.abort();
-      return reply([[11, state("new")]]);
-    });
-    await expect(captureViewModel(cdp, 7, { signal: controller.signal })).rejects.toMatchObject({
-      name: "AbortError",
-    });
-    expect(send).toHaveBeenCalledWith(7, "Runtime.releaseObjectGroup", expect.anything());
-  });
-
-  it("skips entries without node identity while still applying valid entries", async () => {
-    const { cdp } = fakeCdp(
-      [
-        [
-          { id: 11, value: "snapshot value" },
-          { id: 12, value: "old value" },
-        ],
-      ],
-      () => {
-        const result = reply([
-          [11, state("wrong value")],
-          [12, state("live value")],
-        ]);
-        const element = result.result.deepSerializedValue.value[0].value[0];
-        delete (element.value as { backendNodeId?: number }).backendNodeId;
-        return result;
-      },
-    );
-    const captured = await captureViewModel(cdp, 7);
-    expect(captured.nodes.find((n) => n.backendNodeId === 11)?.formValue).toBe("snapshot value");
-    expect(captured.nodes.find((n) => n.backendNodeId === 12)?.formValue).toBe("live value");
-  });
-
-  it.each([
-    "password",
-    "text",
-  ])("redacts a sensitive %s control after identity matching", async (type) => {
-    const { cdp } = fakeCdp([[{ id: 11, value: "secret", type }]], () =>
-      reply([[11, { ...state("secret"), sensitive: true }]]),
-    );
-    const captured = await captureViewModel(cdp, 7);
-    const control = captured.nodes.find((n) => n.backendNodeId === 11);
-    expect(control?.formState).toBe("filled");
-    expect(control?.formValue).toBeUndefined();
-    expect(control?.formDefaultValue).toBeUndefined();
-    expect(control?.attrs.value).toBeUndefined();
-  });
-
-  it("keeps one bounded batch without per-control CDP queries", async () => {
-    const controls = Array.from({ length: 251 }, (_, i) => ({ id: i + 1, value: "old" }));
-    const elements = controls.map((c) => ({
-      backendNodeId: c.id,
-      tagName: "INPUT",
-      type: "text",
-      value: `live ${c.id}`,
-      defaultValue: "",
-      placeholder: "",
-    }));
-    const { cdp, send } = fakeCdp([controls], (params) => {
-      const doc = {
-        querySelectorAll: (selector: string) =>
-          selector === "input,textarea,select" ? elements : [],
-      };
-      const entries = new Function("document", `return ${params.expression}`)(doc) as Array<
-        [{ backendNodeId: number }, string]
-      >;
-      expect(entries).toHaveLength(250);
-      return reply(entries.map(([element, json]) => [element.backendNodeId, JSON.parse(json)]));
-    });
-    const captured = await captureViewModel(cdp, 7);
-    expect(captured.nodes.find((n) => n.backendNodeId === 250)?.formValue).toBe("live 250");
-    expect(captured.nodes.find((n) => n.backendNodeId === 251)?.formValue).toBe("old");
-    expect(send.mock.calls.filter(([, method]) => method === "Runtime.evaluate")).toHaveLength(1);
-    expect(
-      send.mock.calls.filter(([, method]) => method === "Runtime.releaseObjectGroup"),
-    ).toHaveLength(1);
-    expect(send).not.toHaveBeenCalledWith(7, "DOM.resolveNode", expect.anything());
-    expect(send).not.toHaveBeenCalledWith(7, "Runtime.callFunctionOn", expect.anything());
-  });
-});

apps/extension/src/tools/vom/__tests__/hover-perception.test.ts

@@ -1,84 +0,0 @@
-import { afterEach, describe, expect, it, vi } from "vitest";
-import { ProbeBudget, resetOverlayBypassDepth, withOverlayBypass } from "../hover-perception";
-
-afterEach(() => {
-  resetOverlayBypassDepth();
-});
-
-describe("ProbeBudget", () => {
-  it("refuses a candidate that cannot finish inside the budget", () => {
-    const budget = new ProbeBudget(500);
-
-    expect(budget.canAfford(400)).toBe(true);
-    expect(budget.canAfford(600)).toBe(false);
-  });
-
-  it("stops admitting candidates once the remaining budget is smaller than one", async () => {
-    const budget = new ProbeBudget(60);
-    await new Promise((resolve) => setTimeout(resolve, 40));
-
-    // A top-of-loop `elapsed > total` check would still admit this candidate
-    // and then overshoot by its full cost.
-    expect(budget.canAfford(50)).toBe(false);
-  });
-});
-
-describe("withOverlayBypass", () => {
-  it("toggles the overlay once around nested probe phases", async () => {
-    const bypass = vi.fn(async () => {});
-
-    await withOverlayBypass(bypass, 4, async () => {
-      await withOverlayBypass(bypass, 4, async () => {
-        expect(bypass).toHaveBeenCalledTimes(1);
-      });
-      expect(bypass).toHaveBeenCalledTimes(1);
-    });
-
-    expect(bypass.mock.calls).toEqual([
-      [4, true],
-      [4, false],
-    ]);
-  });
-
-  it("restores the overlay when the probe throws", async () => {
-    const bypass = vi.fn(async () => {});
-
-    await expect(
-      withOverlayBypass(bypass, 4, async () => {
-        throw new Error("probe exploded");
-      }),
-    ).rejects.toThrow("probe exploded");
-
-    expect(bypass).toHaveBeenLastCalledWith(4, false);
-  });
-
-  it("reports a failed restore instead of swallowing it", async () => {
-    const onRestoreFailure = vi.fn();
-    const bypass = vi.fn(async (_tabId: number, enabled: boolean) => {
-      if (!enabled) throw new Error("tab is gone");
-    });
-
-    await withOverlayBypass(bypass, 4, async () => undefined, { onRestoreFailure });
-
-    expect(onRestoreFailure).toHaveBeenCalledTimes(1);
-  });
-
-  it("does not pin the overlay off for the tab when a restore fails", async () => {
-    const failing = vi.fn(async (_tabId: number, enabled: boolean) => {
-      if (!enabled) throw new Error("tab is gone");
-    });
-    await withOverlayBypass(failing, 4, async () => undefined, { onRestoreFailure: () => {} });
-
-    const bypass = vi.fn(async () => {});
-    await withOverlayBypass(bypass, 4, async () => undefined);
-
-    expect(bypass.mock.calls).toEqual([
-      [4, true],
-      [4, false],
-    ]);
-  });
-
-  it("runs the probe untouched when no bypass is wired up", async () => {
-    await expect(withOverlayBypass(undefined, 4, async () => "done")).resolves.toBe("done");
-  });
-});

apps/extension/src/tools/vom/__tests__/name-enrichment.test.ts

@@ -132,37 +132,6 @@ describe("tooltip name enrichment", () => {
     expect(cdp.sendToTarget).not.toHaveBeenCalled();
   });
 
-  it("remembers controls that revealed no tooltip so the miss is paid once", async () => {
-    const cdp: CdpRunner = {
-      send: vi.fn(async () => ({})) as CdpRunner["send"],
-      sendToTarget: vi.fn(async (_target, method) => {
-        if (method === "Page.createIsolatedWorld") return { executionContextId: 31 };
-        if (method === "DOM.resolveNode") return { object: { objectId: "button-801" } };
-        if (method === "Runtime.callFunctionOn") return { result: { value: [] } };
-        return {};
-      }) as CdpRunner["sendToTarget"],
-    };
-
-    const documents = [
-      document({ tabId: 11, sessionId: "miss-session" }, 801, undefined, "miss-frame"),
-    ];
-    const graph = semantics(documents);
-
-    const first = await probeTooltipNames(cdp, 11, documents, graph);
-    expect(first.size).toBe(0);
-    const hoversAfterFirst = vi
-      .mocked(cdp.send)
-      .mock.calls.filter(([, method]) => method === "Input.dispatchMouseEvent").length;
-    expect(hoversAfterFirst).toBeGreaterThan(0);
-
-    const second = await probeTooltipNames(cdp, 11, documents, graph);
-    expect(second.size).toBe(0);
-    expect(
-      vi.mocked(cdp.send).mock.calls.filter(([, method]) => method === "Input.dispatchMouseEvent")
-        .length,
-    ).toBe(hoversAfterFirst);
-  });
-
   it("keeps tooltip evidence isolated across sibling frame sessions", async () => {
     let hoveredX = -10;
     const cdp: CdpRunner = {

apps/extension/src/tools/vom/__tests__/rendered-surfaces.test.ts

@@ -0,0 +1,256 @@
+import type { VomScene } from "@browser-skill/vom";
+import { describe, expect, it } from "vitest";
+import type { CapturedNode } from "../capture";
+import type { CapturedFrameDocument } from "../frame-capture";
+import {
+  clusterRenderedSurfaces,
+  discoverRenderedSurfaces,
+  projectRenderedSurfaces,
+} from "../rendered-surfaces";
+import type { SemanticAxNode } from "../semantic-graph";
+
+function domNode(backendNodeId: number, overrides: Partial<CapturedNode> = {}): CapturedNode {
+  return {
+    backendNodeId,
+    parentBackendNodeId: null,
+    frameId: "main",
+    ownerFrameBackendNodeId: null,
+    tag: "canvas",
+    attrs: {},
+    rect: { x: 10, y: 20, w: 800, h: 500 },
+    localRect: { x: 10, y: 20, w: 800, h: 500 },
+    paintOrder: 1,
+    position: "static",
+    pointerEvents: "auto",
+    rendered: true,
+    ...overrides,
+  };
+}
+
+function document(
+  domNodes: CapturedNode[],
+  axNodes: SemanticAxNode[] = [],
+): CapturedFrameDocument<SemanticAxNode> {
+  return {
+    frameId: domNodes[0]?.frameId ?? "main",
+    contextScopeId: domNodes[0]?.frameId ?? "main",
+    target: { tabId: 4 },
+    domNodes,
+    axNodes,
+  };
+}
+
+describe("rendered surface discovery", () => {
+  it("finds prominent rendered canvases and excludes hidden descendants", () => {
+    const visible = domNode(2, { parentBackendNodeId: 1 });
+    const hidden = domNode(3, { parentBackendNodeId: 4 });
+    const surfaces = discoverRenderedSurfaces(
+      [
+        document([
+          domNode(1, { tag: "main", rect: null, localRect: null }),
+          visible,
+          domNode(4, { tag: "section", attrs: { "aria-hidden": "true" }, rect: null }),
+          hidden,
+        ]),
+      ],
+      { width: 1280, height: 720 },
+    );
+
+    expect(surfaces).toHaveLength(1);
+    expect(surfaces[0]).toMatchObject({ backendNodeId: 2 });
+  });
+
+  it("keeps tiny canvases as individually addressable exact discoveries", () => {
+    const surfaces = discoverRenderedSurfaces(
+      [document([domNode(2, { rect: { x: 1, y: 1, w: 8, h: 8 } })])],
+      { width: 1280, height: 720 },
+    );
+    expect(surfaces).toEqual([expect.objectContaining({ backendNodeId: 2 })]);
+  });
+
+  it("keeps a surface when the canvas only has partial native accessibility content", () => {
+    const axNodes = [
+      {
+        nodeId: "canvas",
+        backendDOMNodeId: 2,
+        role: { type: "role", value: "presentation" },
+        childIds: ["button"],
+        frameId: "main",
+      },
+      {
+        nodeId: "button",
+        backendDOMNodeId: 3,
+        role: { type: "role", value: "button" },
+        frameId: "main",
+      },
+    ] as SemanticAxNode[];
+    expect(
+      discoverRenderedSurfaces([document([domNode(2)], axNodes)], { width: 1280, height: 720 }),
+    ).toEqual([expect.objectContaining({ backendNodeId: 2 })]);
+  });
+
+  it("discovers unnamed canvases without guessing labels from engineering attributes", () => {
+    const canvas = domNode(2, {
+      attrs: {
+        id: "must-not-be-a-label",
+        class: "must-not-be-a-label",
+        "data-testid": "must-not-be-a-label",
+      },
+    });
+
+    expect(discoverRenderedSurfaces([document([canvas])], { width: 1280, height: 720 })).toEqual([
+      expect.not.objectContaining({
+        label: expect.anything(),
+      }),
+    ]);
+  });
+
+  it("treats blank explicit Canvas names as unnamed", () => {
+    expect(
+      discoverRenderedSurfaces(
+        [document([domNode(2, { attrs: { "aria-label": "   ", title: "\n\t" } })])],
+        { width: 1280, height: 720 },
+      ),
+    ).toEqual([expect.not.objectContaining({ label: expect.anything() })]);
+  });
+
+  it("excludes canvases inside BrowserSkill-owned DOM subtrees", () => {
+    const owner = domNode(1, { tag: "div", rect: null, localRect: null });
+    const canvas = domNode(2, { parentBackendNodeId: 1 });
+
+    expect(
+      discoverRenderedSurfaces(
+        [document([owner, canvas])],
+        { width: 1280, height: 720 },
+        new Set([1]),
+      ),
+    ).toEqual([]);
+  });
+
+  it("excludes canvases suppressed by a fully transparent ancestor", () => {
+    const owner = domNode(1, {
+      tag: "div",
+      rect: { x: 0, y: 0, w: 900, h: 600 },
+      visuallySuppressed: true,
+    });
+    const canvas = domNode(2, { parentBackendNodeId: 1 });
+
+    expect(
+      discoverRenderedSurfaces([document([owner, canvas])], { width: 1280, height: 720 }),
+    ).toEqual([]);
+  });
+
+  it("clips canvases to overflow ancestors and drops fully clipped canvases", () => {
+    const owner = domNode(1, {
+      tag: "div",
+      rect: { x: 10, y: 20, w: 100, h: 70 },
+      overflowX: "hidden",
+      overflowY: "hidden",
+    });
+    const partial = domNode(2, {
+      parentBackendNodeId: 1,
+      rect: { x: 55, y: 30, w: 140, h: 60 },
+    });
+    const fullyClipped = domNode(3, {
+      parentBackendNodeId: 1,
+      rect: { x: 140, y: 30, w: 80, h: 50 },
+    });
+
+    expect(
+      discoverRenderedSurfaces([document([owner, partial, fullyClipped])], {
+        width: 1280,
+        height: 720,
+      }),
+    ).toEqual([
+      expect.objectContaining({
+        backendNodeId: 2,
+        visibleRect: { x: 55, y: 30, w: 55, h: 60 },
+      }),
+    ]);
+  });
+});
+
+describe("rendered surface clustering and projection", () => {
+  it("collapses overlapping canvas layers and attaches them to the nearest retained parent", () => {
+    const body = domNode(1, { tag: "body", rect: null, localRect: null });
+    const surfaces = discoverRenderedSurfaces(
+      [
+        document([
+          body,
+          domNode(2, { parentBackendNodeId: 1, paintOrder: 1 }),
+          domNode(3, { parentBackendNodeId: 1, paintOrder: 2 }),
+        ]),
+      ],
+      { width: 1280, height: 720 },
+    );
+    const groups = clusterRenderedSurfaces(surfaces);
+    const scene: VomScene = {
+      viewport: { width: 1280, height: 720 },
+      nodes: [
+        {
+          id: 10,
+          parentId: null,
+          backendNodeId: 1,
+          frameId: "main",
+          tag: "body",
+          role: "RootWebArea",
+          rect: null,
+          paintOrder: 0,
+          position: "static",
+          pointerEvents: "auto",
+        },
+      ],
+    };
+
+    expect(groups).toHaveLength(1);
+    expect(groups[0]?.members).toHaveLength(2);
+    expect(groups[0]?.label).toBeUndefined();
+    expect(
+      projectRenderedSurfaces(
+        scene,
+        [document([body, ...surfaces.map((s) => domNode(s.backendNodeId))])],
+        groups,
+      ),
+    ).toMatchObject({
+      visualSurfaces: [expect.objectContaining({ parentId: 10, backendNodeId: 3, memberCount: 2 })],
+    });
+  });
+
+  it("does not merge a contained canvas with a larger independent canvas", () => {
+    const groups = clusterRenderedSurfaces([
+      ...discoverRenderedSurfaces(
+        [
+          document([
+            domNode(1, { tag: "main", rect: null, localRect: null }),
+            domNode(2, { parentBackendNodeId: 1 }),
+            domNode(3, {
+              parentBackendNodeId: 1,
+              rect: { x: 30, y: 40, w: 200, h: 100 },
+            }),
+          ]),
+        ],
+        { width: 1280, height: 720 },
+      ),
+    ]);
+
+    expect(groups).toHaveLength(2);
+  });
+
+  it("does not merge coincident canvases owned by different containers", () => {
+    const groups = clusterRenderedSurfaces(
+      discoverRenderedSurfaces(
+        [
+          document([
+            domNode(1, { tag: "section", rect: null, localRect: null }),
+            domNode(2, { tag: "section", rect: null, localRect: null }),
+            domNode(3, { parentBackendNodeId: 1 }),
+            domNode(4, { parentBackendNodeId: 2 }),
+          ]),
+        ],
+        { width: 1280, height: 720 },
+      ),
+    );
+
+    expect(groups).toHaveLength(2);
+  });
+});

apps/extension/src/tools/vom/capture.ts

@@ -9,9 +9,17 @@ import { evaluateHoverTrigger } from "@/lib/hover-trigger-policy";
 import { isOverlayHostNode, OVERLAY_HOST_SELECTOR } from "../../lib/overlay-bridge";
 import { childFrameProjection, type GeometryProjection, projectRectToViewport } from "../geometry";
 import type { CdpRunner } from "../shared";
-import { clearHover, ProbeBudget, waitForHover } from "./hover-perception";
+import { clearHover, waitForHover } from "./hover-perception";
 
-const REQUESTED_STYLES = ["position", "pointer-events", "cursor", "visibility", "opacity"] as const;
+const REQUESTED_STYLES = [
+  "position",
+  "pointer-events",
+  "cursor",
+  "visibility",
+  "opacity",
+  "overflow-x",
+  "overflow-y",
+] as const;
 const STYLE_COL = Object.fromEntries(
   REQUESTED_STYLES.map((name, index) => [name, index]),
 ) as Record<(typeof REQUESTED_STYLES)[number], number>;
@@ -39,6 +47,10 @@ export interface CapturedNode {
    * `textContent`: the live parser always sets it, hand-built fixtures may not.
    */
   cursor?: string;
+  overflowX?: string;
+  overflowY?: string;
+  /** True when this node or one of its DOM ancestors makes the subtree fully transparent/hidden. */
+  visuallySuppressed?: boolean;
   /**
    * Whether the live DOM snapshot provides a painted, non-hidden box for this
    * node. Semantic resolution uses this only for DOM fallback nodes; AX-backed
@@ -71,11 +83,14 @@ export interface CapturedViewModel {
   /** Explicit DOMSnapshot frame ancestry; never inferred from backend node ids. */
   frameParentIds?: Map<string, string>;
   rootFrameId?: string;
+  surfaceProbes?: CapturedSurfaceProbe[];
   /** Backend node ids belonging to the agent overlay host + its shadow subtree. */
   excludedBackendNodeIds: Set<number>;
 }
 
 export interface CaptureViewModelOptions {
+  conditionalSurfaceProbe?: boolean;
+  hoverProbeBypassOverlay?: (tabId: number, enabled: boolean) => Promise<void>;
   signal?: AbortSignal;
 }
 
@@ -190,9 +205,9 @@ interface CapturedFormState {
   sensitive?: boolean;
 }
 
-interface DeepSerializedValue {
-  type: string;
-  value?: unknown;
+interface CapturedFrameFormState {
+  controls: CapturedFormState[];
+  childFrames: CapturedFrameFormState[];
 }
 
 function formStateBatchExpression(maxControls: number): string {
@@ -215,63 +230,43 @@ function formStateBatchExpression(maxControls: number): string {
         ...(sensitive ? {} : { value: rawValue, defaultValue }),
       };
     };
-    const controls = [];
     const collect = (doc) => {
+      const controls = [];
       for (const el of Array.from(doc.querySelectorAll(controlSelector))) {
         if (remaining <= 0) break;
-        // Deep serialization supplies the node's backend id. Keep the
-        // state as JSON so decoding needs no general-purpose V8 deserializer.
-        controls.push([el, JSON.stringify(controlState(el))]);
+        controls.push(controlState(el));
         remaining -= 1;
       }
+      const childFrames = [];
       for (const frame of Array.from(doc.querySelectorAll("iframe"))) {
-        if (remaining <= 0) break;
         let childDoc = null;
         try { childDoc = frame.contentDocument; } catch { childDoc = null; }
-        if (childDoc) collect(childDoc);
+        childFrames.push(childDoc && remaining > 0 ? collect(childDoc) : { controls: [], childFrames: [] });
       }
+      return { controls, childFrames };
     };
-    collect(document);
-    return controls;
+    return collect(document);
   })()`;
 }
 
-function formStatesByBackendId(result: RuntimeEvaluateReply): Map<number, CapturedFormState> {
-  const states = new Map<number, CapturedFormState>();
-  const serialized = result.result?.deepSerializedValue;
-  if (serialized?.type !== "array" || !Array.isArray(serialized.value)) return states;
-  for (const entry of serialized.value as DeepSerializedValue[]) {
-    if (entry?.type !== "array" || !Array.isArray(entry.value)) continue;
-    const [element, json] = entry.value as DeepSerializedValue[];
-    if (element?.type !== "node" || json?.type !== "string" || typeof json.value !== "string") {
-      continue;
-    }
-    const backendNodeId = (element.value as { backendNodeId?: number } | undefined)?.backendNodeId;
-    if (typeof backendNodeId !== "number" || !Number.isSafeInteger(backendNodeId)) continue;
-    try {
-      const state = JSON.parse(json.value) as CapturedFormState | null;
-      if (
-        !state ||
-        !["empty", "filled", "default"].includes(state.state ?? "") ||
-        typeof state.sensitive !== "boolean" ||
-        typeof state.placeholder !== "string" ||
-        (!state.sensitive &&
-          (typeof state.value !== "string" || typeof state.defaultValue !== "string"))
-      ) {
-        continue;
-      }
-      states.set(backendNodeId, state);
-    } catch {
-      // A malformed entry must not overwrite the snapshot's own state.
-    }
+function flattenFrameFormStates(
+  frame: CapturedFrameFormState | undefined,
+  out: CapturedFormState[][] = [],
+): CapturedFormState[][] {
+  if (!frame) return out;
+  out.push(Array.isArray(frame.controls) ? frame.controls : []);
+  for (const child of Array.isArray(frame.childFrames) ? frame.childFrames : []) {
+    flattenFrameFormStates(child, out);
   }
-  return states;
+  return out;
 }
 
-function applyFormStates(nodes: CapturedNode[], states: Map<number, CapturedFormState>): void {
+function applyFormStates(nodes: CapturedNode[], states: CapturedFormState[]): void {
+  let index = 0;
   for (const node of nodes) {
     if (!isFormControlTag(node.tag)) continue;
-    const state = states.get(node.backendNodeId);
+    const state = states[index];
+    index += 1;
     if (!state) continue;
     node.formState = state.state;
     node.formPlaceholder = state.placeholder ?? "";
@@ -296,37 +291,26 @@ async function enrichFormControlStates(
     nodes.some((node) => isFormControlTag(node.tag)),
   );
   if (!hasControls) return;
-  const objectGroup = `bsk-vom-forms-${crypto.randomUUID()}`;
   try {
     throwIfAborted(signal);
     const result = await cdp.send<RuntimeEvaluateReply>(tabId, "Runtime.evaluate", {
       expression: formStateBatchExpression(MAX_FORM_ENRICH_CONTROLS),
-      objectGroup,
-      serializationOptions: {
-        serialization: "deep",
-        maxDepth: 3,
-        additionalParameters: { maxNodeDepth: 0, includeShadowTree: "none" },
-      },
+      returnByValue: true,
     });
     throwIfAborted(signal);
-    // Backend ids are scoped to this capture's CDP target. OOPIF targets
-    // use their own captureViewModel call and never share this lookup.
-    const states = formStatesByBackendId(result);
-    for (const nodes of frameNodeGroups) {
-      applyFormStates(nodes, states);
+    const frameStates = flattenFrameFormStates(runtimeValue<CapturedFrameFormState>(result));
+    for (let i = 0; i < frameNodeGroups.length; i += 1) {
+      applyFormStates(frameNodeGroups[i], frameStates[i] ?? []);
     }
   } catch (error) {
     if (isAbortError(error)) throw error;
     // Best-effort enrichment. DOMSnapshot/AX data still carries the nodes.
-  } finally {
-    await cdp.send(tabId, "Runtime.releaseObjectGroup", { objectGroup }).catch(() => undefined);
   }
 }
 
 interface RuntimeEvaluateReply {
   result?: {
     value?: unknown;
-    deepSerializedValue?: DeepSerializedValue;
   };
 }
 
@@ -396,16 +380,7 @@ function collectBackendIdsFromDomNode(node: CdpDomNode | undefined, out: Set<num
   }
 }
 
-/**
- * Ceiling for the whole hover-surface phase.
- *
- * The previous 2000 was only a floor: the loop checked elapsed time at the top,
- * so a candidate could start with 1ms left and still run its two settle
- * windows, pushing real cost to ~2.6s. This is that true ceiling, now actually
- * enforced by an up-front affordability check, so the same number of candidates
- * get probed under a bound that no longer lies.
- */
-const MAX_HOVER_PROBE_MS = 2_600;
+const MAX_HOVER_PROBE_MS = 2_000;
 const MAX_HOVER_TRIGGERS = 6;
 const MAX_HOVER_SURFACES = 3;
 const HOVER_SETTLE_MS = 300;
@@ -651,35 +626,13 @@ function confidenceForHover(
   return "low";
 }
 
-/**
- * One candidate costs two settle windows (baseline + post-hover) plus a few
- * CDP round trips. Used to decide whether the next candidate still fits the
- * budget before paying for it.
- */
-const HOVER_CANDIDATE_COST_MS = HOVER_SETTLE_MS * 2;
-
-export interface HoverSurfaceProbeOptions {
-  signal?: AbortSignal;
-}
-
-/**
- * Hovers a bounded set of likely menu triggers and reports the sub-items each
- * one reveals.
- *
- * Runs after DOM *and* accessibility capture. Hovering can open menus and
- * change layout, so probing between the two captures would leave the DOM half
- * of an observation describing the page before the change and the AX half
- * describing it after.
- *
- * The caller owns the overlay bypass span (see `withOverlayBypass`).
- */
-export async function probeHoverSurfaces(
+async function probeHoverSurfaces(
   cdp: CdpRunner,
   tabId: number,
   nodes: CapturedNode[],
-  options: HoverSurfaceProbeOptions = {},
+  options: CaptureViewModelOptions,
 ): Promise<CapturedSurfaceProbe[]> {
-  const budget = new ProbeBudget(MAX_HOVER_PROBE_MS);
+  const started = Date.now();
   try {
     throwIfAborted(options.signal);
     const cssScan = await cdp.send<RuntimeEvaluateReply>(tabId, "Runtime.evaluate", {
@@ -694,54 +647,60 @@ export async function probeHoverSurfaces(
     const results: CapturedSurfaceProbe[] = [];
     const seen = new Set<number>();
     throwIfAborted(options.signal);
-    for (const candidate of candidates.slice(0, MAX_HOVER_TRIGGERS)) {
+    await options.hoverProbeBypassOverlay?.(tabId, true).catch(() => undefined);
+    try {
       throwIfAborted(options.signal);
-      if (!budget.canAfford(HOVER_CANDIDATE_COST_MS)) break;
-      if (results.length >= MAX_HOVER_SURFACES) break;
-      if (seen.has(candidate.backendNodeId)) continue;
-      try {
-        await clearHover(cdp, tabId);
+      for (const candidate of candidates.slice(0, MAX_HOVER_TRIGGERS)) {
         throwIfAborted(options.signal);
-        await waitForHover(HOVER_SETTLE_MS, options.signal);
-        const baselineReply = await cdp.send<RuntimeEvaluateReply>(tabId, "Runtime.evaluate", {
-          expression: hoverStateExpression(),
-          returnByValue: true,
-        });
-        throwIfAborted(options.signal);
-        const baselineItems = runtimeValue<HoverRuntimeItem[]>(baselineReply) ?? [];
+        if (Date.now() - started > MAX_HOVER_PROBE_MS) break;
+        if (results.length >= MAX_HOVER_SURFACES) break;
+        if (seen.has(candidate.backendNodeId)) continue;
+        try {
+          await clearHover(cdp, tabId);
+          throwIfAborted(options.signal);
+          await waitForHover(HOVER_SETTLE_MS, options.signal);
+          const baselineReply = await cdp.send<RuntimeEvaluateReply>(tabId, "Runtime.evaluate", {
+            expression: hoverStateExpression(),
+            returnByValue: true,
+          });
+          throwIfAborted(options.signal);
+          const baselineItems = runtimeValue<HoverRuntimeItem[]>(baselineReply) ?? [];
 
-        throwIfAborted(options.signal);
-        await cdp.send(tabId, "Input.dispatchMouseEvent", {
-          type: "mouseMoved",
-          x: candidate.x,
-          y: candidate.y,
-        });
-        throwIfAborted(options.signal);
-        await waitForHover(HOVER_SETTLE_MS, options.signal);
-        const collected = await cdp.send<RuntimeEvaluateReply>(tabId, "Runtime.evaluate", {
-          expression: hoverStateExpression(),
-          returnByValue: true,
-        });
-        throwIfAborted(options.signal);
-        const subItems = diffHoverItems(
-          baselineItems,
-          runtimeValue<HoverRuntimeItem[]>(collected) ?? [],
-        );
-        if (subItems.length === 0) continue;
-        seen.add(candidate.backendNodeId);
-        results.push({
-          triggerBackendNodeId: candidate.backendNodeId,
-          triggerPoint: { x: candidate.x, y: candidate.y },
-          triggerAction: "hover",
-          subItems,
-          confidence: confidenceForHover(candidate, subItems),
-        });
-      } catch (error) {
-        if (isAbortError(error)) throw error;
-        continue;
-      } finally {
-        await clearHover(cdp, tabId);
+          throwIfAborted(options.signal);
+          await cdp.send(tabId, "Input.dispatchMouseEvent", {
+            type: "mouseMoved",
+            x: candidate.x,
+            y: candidate.y,
+          });
+          throwIfAborted(options.signal);
+          await waitForHover(HOVER_SETTLE_MS, options.signal);
+          const collected = await cdp.send<RuntimeEvaluateReply>(tabId, "Runtime.evaluate", {
+            expression: hoverStateExpression(),
+            returnByValue: true,
+          });
+          throwIfAborted(options.signal);
+          const subItems = diffHoverItems(
+            baselineItems,
+            runtimeValue<HoverRuntimeItem[]>(collected) ?? [],
+          );
+          if (subItems.length === 0) continue;
+          seen.add(candidate.backendNodeId);
+          results.push({
+            triggerBackendNodeId: candidate.backendNodeId,
+            triggerPoint: { x: candidate.x, y: candidate.y },
+            triggerAction: "hover",
+            subItems,
+            confidence: confidenceForHover(candidate, subItems),
+          });
+        } catch (error) {
+          if (isAbortError(error)) throw error;
+          continue;
+        } finally {
+          await clearHover(cdp, tabId);
+        }
       }
+    } finally {
+      await options.hoverProbeBypassOverlay?.(tabId, false).catch(() => undefined);
     }
     return results;
   } catch (err) {
@@ -849,6 +808,8 @@ function parseDocumentNodes(
   const cursorCol = STYLE_COL.cursor;
   const visibilityCol = STYLE_COL.visibility;
   const opacityCol = STYLE_COL.opacity;
+  const overflowXCol = STYLE_COL["overflow-x"];
+  const overflowYCol = STYLE_COL["overflow-y"];
   const inputValues = sparseIndexMap(dn.inputValue);
   const textValues = sparseIndexMap(dn.textValue);
   const checkedInputs = new Set(dn.inputChecked?.index ?? []);
@@ -874,6 +835,7 @@ function parseDocumentNodes(
 
   // Build CapturedNodes.
   const nodes: CapturedNode[] = [];
+  const visuallySuppressedByNodeIndex = new Array<boolean>(count).fill(false);
   for (let n = 0; n < count; n++) {
     if (excluded[n]) continue;
     const backendNodeId = dn.backendNodeId[n];
@@ -893,7 +855,11 @@ function parseDocumentNodes(
     let position = "static";
     let pointerEvents = "auto";
     let cursor = "auto";
+    let overflowX = "visible";
+    let overflowY = "visible";
     let rendered = false;
+    let visuallySuppressed =
+      parentIdx >= 0 ? visuallySuppressedByNodeIndex[parentIdx] === true : false;
     const li = layoutByNode.get(n);
     if (li !== undefined) {
       const b = dl?.bounds?.[li];
@@ -916,12 +882,19 @@ function parseDocumentNodes(
       cursor = str(strings, styleRow[cursorCol]) || "auto";
       const visibility = str(strings, styleRow[visibilityCol]) || "visible";
       const opacity = str(strings, styleRow[opacityCol]) || "1";
+      overflowX = str(strings, styleRow[overflowXCol]) || "visible";
+      overflowY = str(strings, styleRow[overflowYCol]) || "visible";
+      visuallySuppressed ||=
+        visibility === "hidden" ||
+        visibility === "collapse" ||
+        (Number.parseFloat(opacity) || 0) <= 0;
       rendered =
         localRect !== null &&
         visibility !== "hidden" &&
         visibility !== "collapse" &&
         (Number.parseFloat(opacity) || 0) > 0;
     }
+    visuallySuppressedByNodeIndex[n] = visuallySuppressed;
 
     // Skip non-element nodes (#text, #cdata-section, etc.) — they carry no
     // geometry and are never queried by callers.
@@ -956,6 +929,9 @@ function parseDocumentNodes(
       position,
       pointerEvents,
       cursor,
+      overflowX,
+      overflowY,
+      visuallySuppressed,
       rendered,
       textContent,
       ...(formValue !== undefined ? { formValue } : {}),
@@ -1095,6 +1071,7 @@ export async function captureViewModel(
       nodes: [],
       viewport,
       iframeNodes: new Map(),
+      surfaceProbes: [],
       excludedBackendNodeIds: new Set(),
     };
   }
@@ -1123,6 +1100,11 @@ export async function captureViewModel(
   await enrichFormControlStates(cdp, tabId, [nodes, ...iframeNodes.values()], options.signal);
   throwIfAborted(options.signal);
 
+  const surfaceProbes = options.conditionalSurfaceProbe
+    ? await probeHoverSurfaces(cdp, tabId, nodes, options)
+    : [];
+  throwIfAborted(options.signal);
+
   const frameNodes = new Map<string, CapturedNode[]>();
   if (topContext.frameId) frameNodes.set(topContext.frameId, nodes);
   for (const iframeFrameNodes of iframeNodes.values()) {
@@ -1143,6 +1125,7 @@ export async function captureViewModel(
     frameOwnerBackendNodeIds: frameParsed.frameOwnerBackendNodeIds,
     frameParentIds: frameParsed.frameParentIds,
     ...(topContext.frameId ? { rootFrameId: topContext.frameId } : {}),
+    surfaceProbes,
     excludedBackendNodeIds,
   };
 }

apps/extension/src/tools/vom/hover-perception.ts

@@ -35,91 +35,3 @@ export async function clearHover(cdp: CdpRunner, tabId: number): Promise<void> {
       }),
     );
 }
-
-/**
- * Wall-clock budget for a probe phase.
- *
- * Callers ask whether the *next* candidate still fits before starting it,
- * rather than checking elapsed time at the top of the loop. A top-of-loop
- * check lets a candidate start with 1ms of budget left and then run to
- * completion, overshooting by a full settle window.
- */
-export class ProbeBudget {
-  private readonly startedAt = Date.now();
-
-  constructor(private readonly totalMs: number) {}
-
-  get elapsedMs(): number {
-    return Date.now() - this.startedAt;
-  }
-
-  /** True when a candidate costing `estimatedMs` still fits in the budget. */
-  canAfford(estimatedMs: number): boolean {
-    return this.elapsedMs + estimatedMs <= this.totalMs;
-  }
-}
-
-export type OverlayBypass = (tabId: number, enabled: boolean) => Promise<void>;
-
-export interface OverlayBypassScopeOptions {
-  /**
-   * Called when the overlay could not be restored. Defaults to `console.error`
-   * so a stuck-bypassed overlay is never silent; tests inject a spy.
-   */
-  onRestoreFailure?: (error: unknown) => void;
-}
-
-const overlayBypassDepth = new Map<number, number>();
-
-/**
- * Runs `body` with the agent overlay bypassed, restoring it afterwards.
- *
- * Reference counted per tab so nested probe phases share a single bypass span
- * instead of toggling the overlay once per phase. Restore failures are
- * reported rather than swallowed, and the depth is released even when restore
- * throws so a failure cannot pin the overlay off for the tab's lifetime.
- */
-export async function withOverlayBypass<T>(
-  bypass: OverlayBypass | undefined,
-  tabId: number,
-  body: () => Promise<T>,
-  options: OverlayBypassScopeOptions = {},
-): Promise<T> {
-  if (!bypass) return body();
-
-  const depth = overlayBypassDepth.get(tabId) ?? 0;
-  overlayBypassDepth.set(tabId, depth + 1);
-  if (depth === 0) {
-    try {
-      await bypass(tabId, true);
-    } catch (error) {
-      // Failing to hide the overlay only risks probing our own UI, so continue.
-      console.debug("[bsk hover] overlay bypass could not be enabled", error);
-    }
-  }
-
-  try {
-    return await body();
-  } finally {
-    const current = overlayBypassDepth.get(tabId) ?? 1;
-    if (current <= 1) {
-      overlayBypassDepth.delete(tabId);
-      try {
-        await bypass(tabId, false);
-      } catch (error) {
-        (options.onRestoreFailure ?? defaultRestoreFailure)(error);
-      }
-    } else {
-      overlayBypassDepth.set(tabId, current - 1);
-    }
-  }
-}
-
-function defaultRestoreFailure(error: unknown): void {
-  console.error("[bsk hover] agent overlay could not be restored after probing", error);
-}
-
-/** Test seam: drops any bypass depth left behind by a crashed probe. */
-export function resetOverlayBypassDepth(): void {
-  overlayBypassDepth.clear();
-}

apps/extension/src/tools/vom/name-enrichment.ts

@@ -1,7 +1,7 @@
 import { isVomReferenceNode } from "@browser-skill/vom";
 import { type CdpRunner, sendToCdpTarget } from "../shared";
 import type { FrameDocument, FrameOwnedAxNode } from "./frame-document";
-import { clearHover, ProbeBudget, waitForHover } from "./hover-perception";
+import { clearHover, waitForHover } from "./hover-perception";
 import { stableIdentifierName } from "./semantic-graph/name-evidence";
 import { frameBackendKey, type ResolvedSemanticGraph } from "./semantic-graph/types";
 
@@ -9,18 +9,11 @@ const MAX_TOOLTIP_CANDIDATES = 8;
 const MAX_TOOLTIP_PROBE_MS = 2_400;
 const TOOLTIP_SETTLE_MS = 250;
 const CACHE_TTL_MS = 5 * 60_000;
-/**
- * Misses expire sooner than hits: a control that gains a tooltip once the app
- * finishes booting should still be discoverable, but not at the cost of
- * re-probing every unlabelled control on every observation.
- */
-const MISS_CACHE_TTL_MS = 60_000;
 const MAX_CACHE_ENTRIES = 500;
 const FORM_CONTROL_TAGS = new Set(["input", "select", "textarea"]);
 
 interface TooltipCacheEntry {
-  /** `null` records "probed, revealed no tooltip" so the miss is not re-paid. */
-  name: string | null;
+  name: string;
   expiresAt: number;
 }
 
@@ -34,6 +27,7 @@ interface TooltipCandidate<T extends FrameOwnedAxNode> {
 
 export interface TooltipNameProbeOptions {
   signal?: AbortSignal;
+  hoverProbeBypassOverlay?: (tabId: number, enabled: boolean) => Promise<void>;
 }
 
 const tooltipCache = new Map<string, TooltipCacheEntry>();
@@ -45,11 +39,10 @@ function cacheKey<T extends FrameOwnedAxNode>(
   return `${document.target.tabId}\u0000${document.frameId}\u0000${document.url ?? ""}\u0000${backendNodeId}`;
 }
 
-/** Returns the cache entry, or `undefined` when this control was never probed. */
-function cachedEntry<T extends FrameOwnedAxNode>(
+function cachedName<T extends FrameOwnedAxNode>(
   document: FrameDocument<T>,
   backendNodeId: number,
-): TooltipCacheEntry | undefined {
+): string | undefined {
   const key = cacheKey(document, backendNodeId);
   const entry = tooltipCache.get(key);
   if (!entry) return undefined;
@@ -57,13 +50,13 @@ function cachedEntry<T extends FrameOwnedAxNode>(
     tooltipCache.delete(key);
     return undefined;
   }
-  return entry;
+  return entry.name;
 }
 
 function storeName<T extends FrameOwnedAxNode>(
   document: FrameDocument<T>,
   backendNodeId: number,
-  name: string | null,
+  name: string,
 ): void {
   if (tooltipCache.size >= MAX_CACHE_ENTRIES) {
     const oldest = tooltipCache.keys().next().value;
@@ -71,7 +64,7 @@ function storeName<T extends FrameOwnedAxNode>(
   }
   tooltipCache.set(cacheKey(document, backendNodeId), {
     name,
-    expiresAt: Date.now() + (name === null ? MISS_CACHE_TTL_MS : CACHE_TTL_MS),
+    expiresAt: Date.now() + CACHE_TTL_MS,
   });
 }
 
@@ -244,27 +237,22 @@ export async function probeTooltipNames<T extends FrameOwnedAxNode>(
   const names = new Map<string, string>();
   const pending: TooltipCandidate<T>[] = [];
   for (const candidate of candidates(documents, graph)) {
-    const cached = cachedEntry(candidate.document, candidate.backendNodeId);
+    const cached = cachedName(candidate.document, candidate.backendNodeId);
     if (cached) {
-      // A cached miss is still a decision: skip the control instead of re-probing.
-      if (cached.name !== null) {
-        names.set(
-          frameBackendKey(candidate.document.frameId, candidate.backendNodeId),
-          cached.name,
-        );
-      }
+      names.set(frameBackendKey(candidate.document.frameId, candidate.backendNodeId), cached);
     } else if (pending.length < MAX_TOOLTIP_CANDIDATES) {
       pending.push(candidate);
     }
   }
   if (pending.length === 0) return names;
 
-  const budget = new ProbeBudget(MAX_TOOLTIP_PROBE_MS);
+  const started = Date.now();
   const contexts = new Map<string, number>();
+  await options.hoverProbeBypassOverlay?.(tabId, true).catch(() => undefined);
   try {
     for (const candidate of pending) {
       if (options.signal?.aborted) throw new DOMException("observation aborted", "AbortError");
-      if (!budget.canAfford(TOOLTIP_SETTLE_MS)) break;
+      if (Date.now() - started >= MAX_TOOLTIP_PROBE_MS) break;
       let objectId: string | undefined;
       try {
         let contextId = contexts.get(candidate.document.frameId);
@@ -290,9 +278,9 @@ export async function probeTooltipNames<T extends FrameOwnedAxNode>(
         await waitForHover(TOOLTIP_SETTLE_MS, options.signal);
         const after = await tooltipTexts(cdp, candidate.document, objectId);
         const name = cleanTooltipDiff(before, after);
-        storeName(candidate.document, candidate.backendNodeId, name ?? null);
         if (!name) continue;
         names.set(frameBackendKey(candidate.document.frameId, candidate.backendNodeId), name);
+        storeName(candidate.document, candidate.backendNodeId, name);
       } catch (error) {
         if (error instanceof Error && error.name === "AbortError") throw error;
       } finally {
@@ -302,6 +290,7 @@ export async function probeTooltipNames<T extends FrameOwnedAxNode>(
     }
   } finally {
     await clearHover(cdp, tabId);
+    await options.hoverProbeBypassOverlay?.(tabId, false).catch(() => undefined);
   }
   return names;
 }

apps/extension/src/tools/vom/record-safe-observation.ts

@@ -37,17 +37,6 @@ export interface CaptureVomObservationResult {
   frames: CaptureVomFrame[];
   matchNodes: CaptureVomMatchNode[];
   surfaceProbes?: CapturedSurfaceProbe[];
-  /**
-   * Whether this observation actively hovered the page, and whether that
-   * revealed content the static tree does not contain. Lets callers judge how
-   * far the page may have drifted from the returned snapshot.
-   */
-  hoverProbe?: HoverProbeReport;
-}
-
-export interface HoverProbeReport {
-  performed: boolean;
-  revealedContent: boolean;
 }
 
 function projectFrames(documents: CapturedFrameDocument<FrameAxNode>[]): CaptureVomFrame[] {
@@ -86,7 +75,6 @@ export function projectRecordSafeObservation(input: {
   frameDocuments: CapturedFrameDocument<FrameAxNode>[];
   rendered: VomResult;
   surfaceProbes?: CapturedSurfaceProbe[];
-  hoverProbe?: HoverProbeReport;
 }): CaptureVomObservationResult {
   return {
     text: input.rendered.text,
@@ -96,6 +84,5 @@ export function projectRecordSafeObservation(input: {
     frames: projectFrames(input.frameDocuments),
     matchNodes: projectMatchNodes(input.frameDocuments),
     surfaceProbes: input.surfaceProbes,
-    ...(input.hoverProbe ? { hoverProbe: input.hoverProbe } : {}),
   };
 }

apps/extension/src/tools/vom/rendered-surfaces/cluster.ts

@@ -0,0 +1,65 @@
+import type { Rect } from "@browser-skill/vom";
+import type { RenderedSurface, RenderedSurfaceGroup } from "./types";
+
+const STACK_IOU_RATIO = 0.9;
+const STACK_SIZE_RATIO = 0.9;
+
+function area(rect: Rect): number {
+  return Math.max(0, rect.w) * Math.max(0, rect.h);
+}
+
+function intersectionArea(a: Rect, b: Rect): number {
+  const width = Math.max(0, Math.min(a.x + a.w, b.x + b.w) - Math.max(a.x, b.x));
+  const height = Math.max(0, Math.min(a.y + a.h, b.y + b.h) - Math.max(a.y, b.y));
+  return width * height;
+}
+
+function sameVisualStack(a: RenderedSurface, b: RenderedSurface): boolean {
+  if (a.frameId !== b.frameId) return false;
+  if (a.parentBackendNodeId !== b.parentBackendNodeId) return false;
+  const aArea = area(a.visibleRect);
+  const bArea = area(b.visibleRect);
+  const intersection = intersectionArea(a.visibleRect, b.visibleRect);
+  const union = aArea + bArea - intersection;
+  const widthRatio =
+    Math.min(a.visibleRect.w, b.visibleRect.w) / Math.max(a.visibleRect.w, b.visibleRect.w);
+  const heightRatio =
+    Math.min(a.visibleRect.h, b.visibleRect.h) / Math.max(a.visibleRect.h, b.visibleRect.h);
+  return (
+    union > 0 &&
+    intersection / union >= STACK_IOU_RATIO &&
+    widthRatio >= STACK_SIZE_RATIO &&
+    heightRatio >= STACK_SIZE_RATIO
+  );
+}
+
+function toGroup(members: RenderedSurface[]): RenderedSurfaceGroup {
+  const representative = [...members].sort(
+    (a, b) => area(b.visibleRect) - area(a.visibleRect) || b.paintOrder - a.paintOrder,
+  )[0];
+  return {
+    frameId: representative.frameId,
+    members,
+    representative,
+    label: members.map((surface) => surface.label).find((label) => label !== undefined),
+  };
+}
+
+export function clusterRenderedSurfaces(surfaces: RenderedSurface[]): RenderedSurfaceGroup[] {
+  const groups: RenderedSurface[][] = [];
+  for (const surface of surfaces) {
+    const matching = groups.find((group) =>
+      group.some((member) => sameVisualStack(member, surface)),
+    );
+    if (matching) matching.push(surface);
+    else groups.push([surface]);
+  }
+  return groups
+    .map(toGroup)
+    .sort(
+      (a, b) =>
+        a.frameId.localeCompare(b.frameId) ||
+        a.representative.visibleRect.y - b.representative.visibleRect.y ||
+        a.representative.visibleRect.x - b.representative.visibleRect.x,
+    );
+}

apps/extension/src/tools/vom/rendered-surfaces/discovery.ts

@@ -0,0 +1,113 @@
+import type { Rect, Viewport } from "@browser-skill/vom";
+import type { CapturedNode } from "../capture";
+import type { CapturedFrameDocument } from "../frame-capture";
+import type { SemanticAxNode } from "../semantic-graph";
+import type { RenderedSurface } from "./types";
+
+function clean(value: string | undefined): string | undefined {
+  const normalized = value?.replace(/\s+/g, " ").trim();
+  return normalized || undefined;
+}
+
+function intersectViewport(rect: Rect, viewport: Viewport): Rect | null {
+  return intersectRect(rect, { x: 0, y: 0, w: viewport.width, h: viewport.height });
+}
+
+function intersectRect(a: Rect, b: Rect): Rect | null {
+  const x = Math.max(a.x, b.x);
+  const y = Math.max(a.y, b.y);
+  const right = Math.min(a.x + a.w, b.x + b.w);
+  const bottom = Math.min(a.y + a.h, b.y + b.h);
+  if (right <= x || bottom <= y) return null;
+  return { x, y, w: right - x, h: bottom - y };
+}
+
+const CLIPPING_OVERFLOW = new Set(["auto", "clip", "hidden", "scroll"]);
+
+function clipToOverflowAncestors(
+  node: CapturedNode,
+  rect: Rect,
+  nodesByBackend: ReadonlyMap<number, CapturedNode>,
+): Rect | null {
+  let clipped: Rect | null = rect;
+  let parentBackendNodeId = node.parentBackendNodeId;
+  let guard = 0;
+  while (clipped && parentBackendNodeId !== null && guard <= nodesByBackend.size) {
+    const parent = nodesByBackend.get(parentBackendNodeId);
+    if (!parent) break;
+    if (parent.rect) {
+      const clipX = CLIPPING_OVERFLOW.has(parent.overflowX ?? "visible");
+      const clipY = CLIPPING_OVERFLOW.has(parent.overflowY ?? "visible");
+      if (clipX || clipY) {
+        const clipRect = {
+          x: clipX ? parent.rect.x : clipped.x,
+          y: clipY ? parent.rect.y : clipped.y,
+          w: clipX ? parent.rect.w : clipped.w,
+          h: clipY ? parent.rect.h : clipped.h,
+        };
+        clipped = intersectRect(clipped, clipRect);
+      }
+    }
+    parentBackendNodeId = parent.parentBackendNodeId;
+    guard += 1;
+  }
+  return clipped;
+}
+
+function excludedByDomPolicy(
+  node: CapturedNode,
+  nodesByBackend: Map<number, CapturedNode>,
+  excludedBackendNodeIds: ReadonlySet<number>,
+): boolean {
+  let current: CapturedNode | undefined = node;
+  let guard = 0;
+  while (current && guard <= nodesByBackend.size) {
+    if (excludedBackendNodeIds.has(current.backendNodeId)) return true;
+    if (current.visuallySuppressed === true) return true;
+    const attrs = current.attrs;
+    if (
+      Object.prototype.hasOwnProperty.call(attrs, "hidden") ||
+      Object.prototype.hasOwnProperty.call(attrs, "inert") ||
+      (attrs["aria-hidden"] ?? "").toLowerCase() === "true"
+    ) {
+      return true;
+    }
+    current =
+      current.parentBackendNodeId === null
+        ? undefined
+        : nodesByBackend.get(current.parentBackendNodeId);
+    guard += 1;
+  }
+  return false;
+}
+
+export function discoverRenderedSurfaces(
+  documents: CapturedFrameDocument<SemanticAxNode>[],
+  viewport: Viewport,
+  excludedBackendNodeIds: ReadonlySet<number> = new Set(),
+): RenderedSurface[] {
+  const surfaces: RenderedSurface[] = [];
+  for (const document of documents) {
+    const nodesByBackend = new Map(document.domNodes.map((node) => [node.backendNodeId, node]));
+    for (const node of document.domNodes) {
+      if (node.tag.toLowerCase() !== "canvas" || node.rendered !== true || !node.rect) continue;
+      if (excludedByDomPolicy(node, nodesByBackend, excludedBackendNodeIds)) continue;
+      const viewportRect = intersectViewport(node.rect, viewport);
+      const visibleRect = viewportRect
+        ? clipToOverflowAncestors(node, viewportRect, nodesByBackend)
+        : null;
+      if (!visibleRect) continue;
+      const label = clean(node.attrs["aria-label"] ?? node.attrs.title);
+      surfaces.push({
+        renderingKind: "canvas",
+        frameId: document.frameId,
+        backendNodeId: node.backendNodeId,
+        parentBackendNodeId: node.parentBackendNodeId,
+        visibleRect,
+        paintOrder: node.paintOrder,
+        ...(label ? { label } : {}),
+      });
+    }
+  }
+  return surfaces;
+}

apps/extension/src/tools/vom/rendered-surfaces/index.ts

@@ -0,0 +1,4 @@
+export { clusterRenderedSurfaces } from "./cluster";
+export { discoverRenderedSurfaces } from "./discovery";
+export { projectRenderedSurfaces } from "./project";
+export type { RenderedSurface, RenderedSurfaceGroup } from "./types";

apps/extension/src/tools/vom/rendered-surfaces/project.ts

@@ -0,0 +1,100 @@
+import type { VomScene, VomVisualSurface } from "@browser-skill/vom";
+import type { CapturedFrameDocument } from "../frame-capture";
+import type { SemanticAxNode } from "../semantic-graph";
+import type { RenderedSurfaceGroup } from "./types";
+
+function frameBackendKey(frameId: string, backendNodeId: number): string {
+  return `${frameId}\u0000${backendNodeId}`;
+}
+
+interface ProjectionIndex {
+  sceneByBackend: ReadonlyMap<string, number>;
+  documentByFrame: ReadonlyMap<string, CapturedFrameDocument<SemanticAxNode>>;
+  domByFrame: ReadonlyMap<
+    string,
+    ReadonlyMap<number, CapturedFrameDocument<SemanticAxNode>["domNodes"][number]>
+  >;
+  rootSceneIdByFrame: ReadonlyMap<string, number>;
+}
+
+function nearestSceneParent(index: ProjectionIndex, group: RenderedSurfaceGroup): number | null {
+  const document = index.documentByFrame.get(group.frameId);
+  const domByBackend = index.domByFrame.get(group.frameId) ?? new Map();
+  let parentBackendNodeId = group.representative.parentBackendNodeId;
+  let guard = 0;
+  while (parentBackendNodeId !== null && guard <= domByBackend.size) {
+    const sceneId = index.sceneByBackend.get(frameBackendKey(group.frameId, parentBackendNodeId));
+    if (sceneId !== undefined) return sceneId;
+    parentBackendNodeId = domByBackend.get(parentBackendNodeId)?.parentBackendNodeId ?? null;
+    guard += 1;
+  }
+
+  if (document?.parentFrameId && document.ownerBackendNodeId !== undefined) {
+    const owner = index.sceneByBackend.get(
+      frameBackendKey(document.parentFrameId, document.ownerBackendNodeId),
+    );
+    if (owner !== undefined) return owner;
+  }
+
+  return index.rootSceneIdByFrame.get(group.frameId) ?? null;
+}
+
+function buildProjectionIndex(
+  scene: VomScene,
+  documents: CapturedFrameDocument<SemanticAxNode>[],
+): ProjectionIndex {
+  const rootSceneIdByFrame = new Map<string, number>();
+  for (const node of scene.nodes) {
+    if (
+      node.frameId &&
+      !rootSceneIdByFrame.has(node.frameId) &&
+      ["rootwebarea", "webarea"].includes(node.role?.toLowerCase() ?? "")
+    ) {
+      rootSceneIdByFrame.set(node.frameId, node.id);
+    }
+  }
+  return {
+    sceneByBackend: new Map(
+      scene.nodes.flatMap((node) =>
+        node.backendNodeId === undefined
+          ? []
+          : [[frameBackendKey(node.frameId ?? "", node.backendNodeId), node.id] as const],
+      ),
+    ),
+    documentByFrame: new Map(documents.map((document) => [document.frameId, document])),
+    domByFrame: new Map(
+      documents.map((document) => [
+        document.frameId,
+        new Map(document.domNodes.map((node) => [node.backendNodeId, node])),
+      ]),
+    ),
+    rootSceneIdByFrame,
+  };
+}
+
+export function projectRenderedSurfaces(
+  scene: VomScene,
+  documents: CapturedFrameDocument<SemanticAxNode>[],
+  groups: RenderedSurfaceGroup[],
+): VomScene {
+  const index = buildProjectionIndex(scene, documents);
+  const visualSurfaces: VomVisualSurface[] = [];
+
+  for (const group of groups) {
+    const parentId = nearestSceneParent(index, group);
+    visualSurfaces.push({
+      parentId,
+      backendNodeId: group.representative.backendNodeId,
+      frameId: group.frameId,
+      renderingKind: "canvas",
+      visibleRect: group.representative.visibleRect,
+      ...(group.label ? { label: group.label } : {}),
+      memberCount: group.members.length,
+    });
+  }
+
+  return {
+    ...scene,
+    ...(visualSurfaces.length > 0 ? { visualSurfaces } : {}),
+  };
+}

apps/extension/src/tools/vom/rendered-surfaces/types.ts

@@ -0,0 +1,18 @@
+import type { Rect } from "@browser-skill/vom";
+
+export interface RenderedSurface {
+  renderingKind: "canvas";
+  frameId: string;
+  backendNodeId: number;
+  parentBackendNodeId: number | null;
+  visibleRect: Rect;
+  paintOrder: number;
+  label?: string;
+}
+
+export interface RenderedSurfaceGroup {
+  frameId: string;
+  members: RenderedSurface[];
+  representative: RenderedSurface;
+  label?: string;
+}

apps/extension/src/transport/types.ts

@@ -23,41 +23,28 @@ export type ErrorCode =
 export type RpcErrorReason =
   | "agent_window_scope"
   | "element_not_visible"
+  | "ref_capability_denied"
   | "ref_not_found"
   | "selector_not_found"
   | "target_not_fillable"
-  | "fill_value_invalid"
-  | "fill_target_changed"
-  | "fill_focus_lost"
-  | "fill_value_mismatch"
-  | "fill_failed"
+  | "focus_target_unresolved"
+  | "input_not_applied"
   | "target_not_select"
   | "option_not_found"
   | "single_select_value_count"
   | "tab_not_active"
   | "restricted_tab_url"
-  | "cdp_extension_access_denied"
   | "borrow_conflict"
   | "screenshot_capture_failed"
-  | "file_input_probe_failed"
-  | "file_input_not_activated"
-  | "set_file_input_failed"
-  | "upload_mechanism_unsupported"
-  | "file_drop_target_unavailable"
-  | "file_drop_failed"
-  | "download_capture_failed"
-  | "transfer_outcome_unknown"
-  | "transfer_timeout"
+  | "surface_capture_not_found"
+  | "surface_capture_expired"
+  | "surface_capture_consumed"
+  | "surface_capture_stale"
+  | "surface_coordinate_invalid"
   | "cleanup_failed";
 
-export type TransferEffectState = "none" | "committed" | "unknown";
-export type TransferCleanupState = "complete" | "failed";
-
 export interface RpcErrorData {
   reason?: RpcErrorReason;
-  effect_state?: TransferEffectState;
-  phase?: string;
-  cleanup_state?: TransferCleanupState;
   [key: string]: unknown;
 }
 
@@ -319,9 +306,17 @@ export interface ScreenshotResult {
   height: number;
   format: string;
   tab_id: number;
+  capture?: SurfaceCaptureInfo;
   dialogs?: JavaScriptDialogInfo[];
 }
 
+export interface SurfaceCaptureInfo {
+  id: string;
+  surface_ref: string;
+  coordinate_space: "capture-image-pixel";
+  expires_at: number;
+}
+
 export interface SnapshotParams {
   session_id: string;
   tab_id?: number;
@@ -339,14 +334,9 @@ export interface SnapshotResult {
 
 export interface ObserveParams extends SnapshotParams {
   debug_surfaces?: boolean;
-  probe_hover?: boolean;
 }
 
 export interface ObserveResult extends SnapshotResult {
-  hover_probe?: {
-    performed: boolean;
-    revealed_content: boolean;
-  };
   debug?: {
     surface_probes?: Array<{
       trigger_backend_node_id: number;
@@ -440,6 +430,9 @@ export interface ClickParams {
   click_count?: number;
   modifiers?: KeyModifier[];
   timeout_ms?: number;
+  capture_id?: string;
+  image_x?: number;
+  image_y?: number;
 }
 
 export interface ClickResult {
@@ -448,6 +441,9 @@ export interface ClickResult {
   used_selector?: string;
   x: number;
   y: number;
+  capture_id?: string;
+  image_x?: number;
+  image_y?: number;
   dialogs?: JavaScriptDialogInfo[];
 }
 
@@ -488,6 +484,26 @@ export interface FillResult {
   dialogs?: JavaScriptDialogInfo[];
 }
 
+export interface TypeParams {
+  session_id: string;
+  text: string;
+  ref?: string;
+  selector?: string;
+  /** Explicitly dispatch to the unique focused text-entry target. */
+  focused?: boolean;
+  tab_id?: number;
+  timeout_ms?: number;
+}
+
+export interface TypeResult {
+  tab_id: number;
+  used_ref?: string;
+  used_selector?: string;
+  text_length: number;
+  verification: "dispatched";
+  dialogs?: JavaScriptDialogInfo[];
+}
+
 export interface PressParams {
   session_id: string;
   key: string;
@@ -526,53 +542,6 @@ export interface SelectResult {
   dialogs?: JavaScriptDialogInfo[];
 }
 
-export interface UploadFile {
-  transfer_id: string;
-  name: string;
-  staged_path?: string;
-}
-
-export type UploadMode = "input" | "drop";
-
-export interface UploadParams {
-  session_id: string;
-  ref?: string;
-  selector?: string;
-  tab_id?: number;
-  files: UploadFile[];
-  mode?: UploadMode;
-  timeout_ms?: number;
-}
-
-export interface UploadResult {
-  tab_id: number;
-  used_ref?: string;
-  used_selector?: string;
-  file_names: string[];
-}
-
-export interface DownloadParams {
-  session_id: string;
-  ref?: string;
-  selector?: string;
-  tab_id?: number;
-  timeout_ms?: number;
-  browser_relative_dir?: string;
-  max_byte_size?: number;
-}
-
-export interface DownloadResult {
-  tab_id: number;
-  used_ref?: string;
-  used_selector?: string;
-  suggested_filename: string;
-  byte_size: number;
-  mime?: string;
-  danger?: string;
-  browser_path?: string;
-  transfer_id?: string;
-}
-
 // --------------------------------------------------------------------------
 // M9 tool payloads — evaluate / wait_for_navigation / wait_ms
 // --------------------------------------------------------------------------

apps/extension/test-fixtures/rendered-surfaces/README.md

@@ -0,0 +1,147 @@
+# Rendered Surface browser fixtures
+
+These pages exercise Canvas discovery through a real browser DOMSnapshot and AX tree. They are deliberately framework- and site-independent.
+
+## Serve
+
+From the repository root:
+
+```bash
+python3 -m http.server 4173 --bind 127.0.0.1 \
+  --directory apps/extension/test-fixtures/rendered-surfaces
+```
+
+Use an Agent Window large enough to keep the primary matrix visible:
+
+```bash
+bsk session start --width 1440 --height 1100
+bsk navigate http://127.0.0.1:4173/index.html --session <id>
+bsk observe --session <id>
+```
+
+Always stop the session after the run.
+
+## Pages
+
+### `index.html`
+
+Expected individual Surface labels:
+
+- `fixture-basic-large`
+- `fixture-tiny-8px`
+- `fixture-pointer-events-none`
+- `fixture-partial-ax`
+- `fixture-contained-outer`
+- `fixture-contained-inner`
+- `fixture-different-parent-a`
+- `fixture-different-parent-b`
+
+Expected one grouped Surface:
+
+- `fixture-layered-group`, with `layers=2`
+
+Expected absent:
+
+- `fixture-display-none`
+- `fixture-visibility-hidden`
+- `fixture-opacity-zero`
+- `fixture-parent-opacity-zero`
+- `fixture-hidden-attribute`
+- `fixture-inert-parent`
+- `fixture-aria-hidden-parent`
+- `fixture-zero-size`
+- `fixture-offscreen-right`
+- `fixture-fully-overflow-clipped`
+
+`fixture-partially-overflow-clipped` must remain present because part of it is visible. Its screenshot must contain only the browser's final composited view.
+
+### `unlabeled.html`
+
+This page contains eight Canvas elements but must render seven generic Surface markers because
+the two coincident layers form one group.
+
+- Every marker uses the stable fallback label `canvas visual surface`.
+- Exactly one marker has `layers=2`.
+- The two contained canvases remain separate markers.
+- `id`, `class`, `data-testid`, nearby prose, and fallback AX content never become a Surface label.
+- Blank `aria-label` and `title` values behave as no label.
+- Fallback AX content does not suppress the corresponding Surface.
+
+### `frames.html`
+
+Expected Surface labels:
+
+- `fixture-same-origin-frame`
+- `fixture-cross-origin-frame`
+- `fixture-nested-frame`
+
+The cross-origin frame uses `http://localhost:4173` while the parent uses `http://127.0.0.1:4173`. With Chromium site isolation this exercises the OOPIF path without an external service.
+
+Each frame Canvas also exposes a `pointer status` live region. A successful
+screenshot-bound point click must report
+`down button=0 buttons=1 hoverReady=true; click received` and turn the Canvas
+yellow. The hover flag is set on the animation frame after `pointermove`, so the
+fixture verifies both child-document delivery and the scheduling boundary
+required before `pointerdown`, rather than merely validating projected
+coordinates or a successful CDP response.
+
+### `viewport.html`
+
+Run with a `900x700` Agent Window.
+
+- `fixture-partial-viewport` is partly visible and must be represented.
+- `fixture-below-fold` is absent initially.
+- After scrolling it into view and observing again, `fixture-below-fold` must be represented with a fresh ref.
+
+### `input-frames.html`
+
+This page exercises generic frame-aware input routing. It contains a top-level
+input plus same-origin, cross-origin (OOPIF), and nested frame inputs. Each
+frame includes an input, textarea, contenteditable editor, an input that
+replaces its DOM node after accepted input, an input that rejects synthetic
+DOM events, a transient keyboard/IME input proxy that consumes and clears its
+DOM value, and a controlled input that resets its value on the next animation
+frame.
+
+- `fill` must route focus, text insertion, DOM events, and readback through the
+  ref's owning CDP target.
+- `press --ref` must focus and dispatch every key event through that same target.
+- Untargeted `press` must follow the actual focused target or refuse when focus
+  ownership cannot be determined uniquely.
+- The remounting input must retain the value through its replacement node.
+- The trusted input must be filled without synthetic `input` / `change` events.
+- `type` must deliver ASCII key events and committed non-ASCII IME/software
+  keyboard insertion to the input proxy. Its DOM value remains empty while
+  `proxy output` retains the consumed text; the result reports
+  `verification=dispatched`, not a retained DOM value.
+- `type` must reject an omitted target unless `focused=true` is explicit, and
+  focused mode must reject focus that is not owned by a text-entry target.
+- The rejecting input must return `input_not_applied`, never a successful
+  requested-value length.
+
+### `canvas-input.html`
+
+This is the PR3 end-to-end fixture. Its cross-origin frame contains a
+visual-only three-row Canvas. A screenshot-bound click on row 3 exposes and
+focuses a standard textbox; ordinary `fill` must then write and read back its
+real value through the textbox's owning target. The fixture deliberately does
+not provide a Canvas-specific fill action.
+
+### `dynamic.html`
+
+Use the normal DOM buttons and re-observe after each action:
+
+- Initial state: `fixture-dynamic` absent.
+- After **Insert canvas**: present.
+- After **Resize to zero**: absent.
+- After **Restore size**: present with a fresh ref.
+- After **Remove canvas**: absent.
+
+## Invariants
+
+- `snapshot` must not emit any of these Surface markers.
+- `observe` must emit only currently visible Canvas surfaces.
+- Existing DOM/AX controls remain readable alongside Surface markers.
+- Missing labels never prevent discovery; naming and discovery are independent.
+- A Surface ref allows `screenshot --ref` and rejects ordinary click/fill/hover/select.
+- No fixture URL, id, label, or layout constant may appear in production discovery code.

apps/extension/test-fixtures/rendered-surfaces/canvas-input.html

@@ -0,0 +1,20 @@
+<!doctype html>
+<html lang="en">
+  <head>
+    <meta charset="utf-8" />
+    <meta name="viewport" content="width=device-width, initial-scale=1" />
+    <title>Canvas to standard input fixture</title>
+    <style>
+      body { padding: 16px; font: 14px system-ui, sans-serif; }
+      iframe { width: 300px; height: 220px; border: 2px solid #5d6d7e; }
+    </style>
+  </head>
+  <body>
+    <h1>Canvas to standard input fixture</h1>
+    <p>Select row 3 in the visual-only Canvas, then fill the exposed standard textbox.</p>
+    <iframe
+      title="cross-origin Canvas editor"
+      src="http://localhost:4173/frame.html?label=canvas-input-grid&interactive=1&editor=1"
+    ></iframe>
+  </body>
+</html>

apps/extension/test-fixtures/rendered-surfaces/dynamic.html

@@ -0,0 +1,46 @@
+<!doctype html>
+<html lang="en">
+  <head>
+    <meta charset="utf-8" />
+    <meta name="viewport" content="width=device-width, initial-scale=1" />
+    <title>Rendered Surface dynamic matrix</title>
+    <style>
+      html, body { margin: 0; font: 14px system-ui, sans-serif; }
+      body { padding: 16px; }
+      .controls { display: flex; gap: 8px; margin-bottom: 16px; }
+      button { padding: 8px 12px; }
+      canvas { display: block; width: 260px; height: 120px; background: #e8daef; border: 2px solid #7d3c98; }
+    </style>
+  </head>
+  <body>
+    <h1>Dynamic Canvas lifecycle</h1>
+    <div class="controls">
+      <button id="insert" type="button">Insert canvas</button>
+      <button id="zero" type="button">Resize to zero</button>
+      <button id="restore" type="button">Restore size</button>
+      <button id="remove" type="button">Remove canvas</button>
+    </div>
+    <div id="mount"></div>
+    <script>
+      const mount = document.querySelector("#mount");
+      const getCanvas = () => mount.querySelector("canvas");
+      document.querySelector("#insert").addEventListener("click", () => {
+        if (getCanvas()) return;
+        const canvas = document.createElement("canvas");
+        canvas.width = 260;
+        canvas.height = 120;
+        canvas.setAttribute("aria-label", "fixture-dynamic");
+        mount.append(canvas);
+      });
+      document.querySelector("#zero").addEventListener("click", () => {
+        const canvas = getCanvas();
+        if (canvas) canvas.style.cssText = "width:0;height:0;border:0";
+      });
+      document.querySelector("#restore").addEventListener("click", () => {
+        const canvas = getCanvas();
+        if (canvas) canvas.removeAttribute("style");
+      });
+      document.querySelector("#remove").addEventListener("click", () => getCanvas()?.remove());
+    </script>
+  </body>
+</html>

apps/extension/test-fixtures/rendered-surfaces/frame.html

@@ -0,0 +1,80 @@
+<!doctype html>
+<html lang="en">
+  <head>
+    <meta charset="utf-8" />
+    <meta name="viewport" content="width=device-width, initial-scale=1" />
+    <style>
+      html, body { margin: 0; font: 14px system-ui, sans-serif; }
+      body { padding: 12px; background: #fdfefe; }
+      canvas { display: block; width: 220px; height: 90px; background: #d5f5e3; border: 2px solid #1e8449; }
+      iframe { width: 280px; height: 150px; margin-top: 10px; border: 2px solid #85929e; }
+      #pointer-status { margin: 8px 0 0; min-height: 20px; }
+    </style>
+  </head>
+  <body>
+    <canvas id="frame-canvas" width="220" height="90"></canvas>
+    <script>
+      const params = new URLSearchParams(location.search);
+      const canvas = document.querySelector("#frame-canvas");
+      if (params.get("canvas") === "0") canvas.remove();
+      else {
+        canvas.setAttribute("aria-label", params.get("label") || "fixture-frame-default");
+        if (params.get("editor") === "1") {
+          const context = canvas.getContext("2d");
+          context.font = "14px system-ui, sans-serif";
+          context.textBaseline = "middle";
+          context.strokeStyle = "#1e8449";
+          context.fillStyle = "#145a32";
+          for (let row = 0; row < 3; row += 1) {
+            const top = row * 30;
+            context.strokeRect(0.5, top + 0.5, 219, 29);
+            context.fillText(`Row ${row + 1}`, 12, top + 15);
+          }
+        }
+        if (params.get("interactive") === "1") {
+          const status = document.createElement("p");
+          status.id = "pointer-status";
+          status.setAttribute("aria-live", "polite");
+          status.textContent = "pointer status: idle";
+          canvas.insertAdjacentElement("afterend", status);
+          let hoverReady = false;
+          canvas.addEventListener("pointermove", () => {
+            hoverReady = false;
+            requestAnimationFrame(() => {
+              hoverReady = true;
+            });
+          });
+          canvas.addEventListener("pointerdown", (event) => {
+            status.textContent = `pointer status: down button=${event.button} buttons=${event.buttons} hoverReady=${hoverReady}`;
+            if (event.buttons === 1 && hoverReady) canvas.style.background = "#f9e79f";
+          });
+          canvas.addEventListener("click", () => {
+            status.textContent += "; click received";
+          });
+          if (params.get("editor") === "1") {
+            canvas.addEventListener("click", (event) => {
+              const rect = canvas.getBoundingClientRect();
+              const canvasY = ((event.clientY - rect.top) / rect.height) * canvas.height;
+              const row = Math.min(2, Math.max(0, Math.floor(canvasY / 30)));
+              status.textContent += `; selected row=${row + 1}`;
+              document.querySelector("#canvas-cell-editor")?.remove();
+              if (row !== 2) return;
+              const editor = document.createElement("input");
+              editor.id = "canvas-cell-editor";
+              editor.setAttribute("aria-label", "selected canvas cell editor");
+              editor.placeholder = "Type into selected row 3";
+              canvas.insertAdjacentElement("afterend", editor);
+              editor.focus();
+            });
+          }
+        }
+      }
+      if (params.get("nested") === "1") {
+        const iframe = document.createElement("iframe");
+        iframe.title = "nested fixture frame";
+        iframe.src = "frame.html?label=fixture-nested-frame&interactive=1";
+        document.body.append(iframe);
+      }
+    </script>
+  </body>
+</html>

apps/extension/test-fixtures/rendered-surfaces/frames.html

@@ -0,0 +1,22 @@
+<!doctype html>
+<html lang="en">
+  <head>
+    <meta charset="utf-8" />
+    <meta name="viewport" content="width=device-width, initial-scale=1" />
+    <title>Rendered Surface frame matrix</title>
+    <style>
+      html, body { margin: 0; font: 14px system-ui, sans-serif; }
+      body { padding: 16px; }
+      .frames { display: grid; grid-template-columns: repeat(3, 320px); gap: 16px; }
+      iframe { width: 300px; height: 300px; border: 2px solid #5d6d7e; }
+    </style>
+  </head>
+  <body>
+    <h1>Frame matrix</h1>
+    <div class="frames">
+      <iframe title="same-origin fixture" src="frame.html?label=fixture-same-origin-frame&interactive=1"></iframe>
+      <iframe title="cross-origin fixture" src="http://localhost:4173/frame.html?label=fixture-cross-origin-frame&interactive=1"></iframe>
+      <iframe title="nested fixture" src="frame.html?canvas=0&nested=1"></iframe>
+    </div>
+  </body>
+</html>