Compare commits

..
Author SHA1 Message Date
drakezhang 5cd43553ed docs(eval): clarify agent-neutral scope 2026-08-28 15:36:38 +08:00
drakezhang 9ad33801c7 test: add extensible browser evaluation harness 2026-08-28 12:16:39 +08:00
Zhang GH 7aaa5a2d01 Merge pull request #147 from Tencent/fix/dsh-plugin-session-lifecycle
fix(dsh-plugin): stabilize skill loading and browser sessions
2026-08-28 12:11:08 +08:00
drakezhang 210b21b637 style(dsh-plugin): format observation 2026-08-28 12:06:47 +08:00
Hannah Liu 48f64ec0df Merge pull request #122 from Tencent/feat/vom-functional-refactor
feat(vom): refactor semantic structure generation for grouped observations | 重构语义结构生成,支持分组式 VOM 观察
2026-08-28 11:24:34 +08:00
drakezhang 1a70d2200e fix(dsh-plugin): stabilize skill loading and browser sessions
Register bundled browser instructions at agent scope, prioritize foreground commands over thumbnail capture, and unify cancellation and session cleanup.
2026-08-27 19:59:20 +08:00
Zhang GH 554861d6b1 Merge pull request #131 from Tencent/feat/dsh-browser-tools-parity
feat(dsh-plugin): add Browser tool parity and dedicated Harness skill
2026-08-27 16:29:30 +08:00
Ljy-0827 0d7f976335 fix(vom): fix repeat names 2026-08-26 15:28:27 +08:00
Ljy-0827 d1d3eb6a94 fix(vom): bug fix 2026-08-26 15:28:27 +08:00
Ljy-0827 38fd7d3f4c fix(vom): safety policy and rebase bug fix 2026-08-26 15:28:27 +08:00
Ljy-0827 b5062ac19b feat(vom): fix name lost 2026-08-26 15:28:27 +08:00
Ljy-0827 41cb576256 feat(vom): function refactor 2026-08-26 15:28:27 +08:00
Hannah Liu f060a7c074 Merge pull request #142 from Tencent/chore/release-ext-0.1.7
chore(release): extension 0.1.7
2026-08-26 15:19:33 +08:00
Ljy-0827 60b62d5420 chore(release): extension 0.1.7 2026-08-26 15:12:44 +08:00
Hannah Liu 70b851a3e8 Merge pull request #141 from Tencent/chore/release-bsk-0.1.11
chore(release): bsk 0.1.11
2026-08-26 15:06:30 +08:00
Ljy-0827 da3534d066 chore(release): bsk 0.1.11 2026-08-26 14:59:09 +08:00
BB-fat a004291848 feat(dsh-plugin): better-sidebar 集成、会话关停与归档清理、悬浮窗修复 (#130)
* better-sidebar 集成:observation 视图迁入 better-sidebar 载体,按会话作用域展示 bsk 会话
* 会话关停与归档清理:支持会话 dispose 与归档自动清理
* 悬浮窗修复:悬浮 overlay 拖拽/缩放/PiP 弹出等问题修复
* 修复 Document PiP 空白窗口:载体卸载(关 tab、切会话、插件 HMR)时主动 close() PiP 窗口,避免残留空白窗口
2026-08-25 10:33:46 +08:00
Hannah Liu 4b6cdde168 Merge pull request #115 from Tencent/fix/recorder-iframe
fix(recording): support user action recording in iframes and OOPIFs
2026-08-24 13:18:05 +08:00
Ljy-0827 4aade4b94f fix(recorder): bug fix, oopif load too late 2026-08-24 10:42:46 +08:00
Ljy-0827 616af1099f fix(record): saftey policy and rebase bug fix 2026-08-24 10:42:46 +08:00
Ljy-0827 3656f0dd3c fix(recorder): ci fix 2026-08-24 10:42:46 +08:00
Ljy-0827 9742d83213 fix(recorder): bug fix 2026-08-24 10:42:46 +08:00
Ljy-0827 615e866608 fix(recorder): add iframe support 2026-08-24 10:42:46 +08:00
Ljy-0827 c08fed5de3 fix(record): saftey policy and rebase bug fix 2026-08-24 10:42:45 +08:00
“haonan” 887d1ed0b9 fix(record): drop 'hover' in tracev2 for compatible 2026-08-24 10:42:45 +08:00
Ljy-0827 1a3d3b7de8 fix(record): bug fix 2026-08-24 10:42:45 +08:00
“haonan”andCursor 72834980d2 fix(record): record switch_tab and keep Trace v3 states per tab
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-24 10:42:45 +08:00
“haonan” 38b0debba9 fix(test): share one BSK_HOME lock across lib tests 2026-08-24 10:40:37 +08:00
“haonan” 4677149e5e fix(record): recover failed Trace exports and hint --url when the default start page fails
Reject legacy JSON --output before recording starts, persist the completed Trace before writing the bundle so `bsk record stop --output <dir>` can retry after an export failure, and tell users to pass --url when https://example.com/ does not load.
2026-08-23 21:26:40 +08:00
Ljy-0827 8ea5477e11 fix(record): safety problem and bug fix 2026-08-23 21:26:40 +08:00
“haonan” 17af1af65a feat(record): export Trace v3 bundles with settled page states 2026-08-23 21:26:40 +08:00
Ljy-0827 6300eda0aa fix(record): rebase and dto policy 2026-08-23 21:26:40 +08:00
Ljy-0827 87cdcfe5ca fix(recording): using unified geometry 2026-08-23 21:26:40 +08:00
“haonan”andCursor 21476c3884 feat(record): capture Trace v3 states after pages settle
Wait for DOM quiet after actions, match targets to VOM refs, and reduce
drafts into deduped state-linked Trace v3 while keeping the v2 reducer.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-23 21:26:40 +08:00
drakezhang a4b6a61d29 docs(dsh-plugin): clarify package description 2026-08-22 21:29:33 +08:00
drakezhang cad0b5cbc9 feat(dsh-plugin): add Browser tool parity 2026-08-22 21:15:53 +08:00
192 changed files with 20543 additions and 2052 deletions
+4
View File
@@ -11,6 +11,10 @@ dist/
# Logs
*.log
# Local browser evaluation artifacts
evals/browser/results/
evals/browser/agents.local.json
# macOS
.DS_Store
Generated
+2 -2
View File
@@ -134,7 +134,7 @@ dependencies = [
[[package]]
name = "bsk"
version = "0.1.10"
version = "0.1.11"
dependencies = [
"anyhow",
"bsk-protocol",
@@ -168,7 +168,7 @@ dependencies = [
[[package]]
name = "bsk-protocol"
version = "0.1.10"
version = "0.1.11"
dependencies = [
"schemars",
"semver",
+1 -1
View File
@@ -7,7 +7,7 @@ members = [
default-members = ["crates/bsk-cli"]
[workspace.package]
version = "0.1.10"
version = "0.1.11"
edition = "2024"
rust-version = "1.85"
license = "MIT"
+1
View File
@@ -208,6 +208,7 @@ The repository is a Cargo + pnpm workspace:
- `apps/extension` — browser extension
- `packages/ui` and `packages/i18n` — shared extension UI support
- `packages/dsh-plugin-browserskill` — DeepSeek Harness plugin (`@wxg-prc-cpg/browser-skill-dsh-plugin`)
- [`evals/browser`](evals/browser/README.md) — deterministic local pages and agent-neutral browser capability evaluation
## License
+1
View File
@@ -178,6 +178,7 @@ Agent 不直接与浏览器通信。它通过 `bsk` CLI 下发浏览器任务;
- `apps/extension` — 浏览器扩展
- `packages/ui` 和 `packages/i18n` — 扩展 UI 共享支持
- `packages/dsh-plugin-browserskill` — DeepSeek Harness 插件(`@wxg-prc-cpg/browser-skill-dsh-plugin`)
- [`evals/browser`](evals/browser/README.zh-CN.md) — 确定性本地页面与 Agent 无关的浏览器能力测试台
## 许可证
+1 -1
View File
@@ -1,6 +1,6 @@
{
"name": "@browser-skill/extension",
"version": "0.1.6",
"version": "0.1.7",
"private": true,
"type": "module",
"scripts": {
@@ -1,10 +1,12 @@
import { beforeEach, describe, expect, it, vi } from "vitest";
import { RECORD_STOP, type RecordStepPayload } from "@/lib/record-bridge";
import { handleRecordContentMessage, startRecordCapture } from "../record-capture";
import type { RecordStepPayload } from "@/lib/record-bridge";
import { startRecordCapture } from "../record-capture";
vi.stubGlobal("chrome", {
runtime: {
sendMessage: vi.fn(() => Promise.resolve()),
sendMessage: vi.fn((message: { sequence?: number }) =>
Promise.resolve({ ok: true, sequence: message.sequence }),
),
},
});
@@ -46,58 +48,6 @@ function click(el: Element): void {
el.dispatchEvent(new MouseEvent("click", { bubbles: true, button: 0, detail: 1 }));
}
describe("handleRecordContentMessage stop/cancel", () => {
it("ignores STOP when no recording is active", () => {
const dispose = vi.fn();
const onStop = vi.fn();
const setActiveRequestId = vi.fn();
const setCapture = vi.fn();
const sendResponse = vi.fn();
const needsAsync = handleRecordContentMessage(
{ type: RECORD_STOP, requestId: "rec-stale" },
{
activeRequestId: null,
capture: { dispose },
setActiveRequestId,
setCapture,
onStart: vi.fn(),
onStop,
},
sendResponse,
);
expect(needsAsync).toBe(false);
expect(dispose).not.toHaveBeenCalled();
expect(onStop).not.toHaveBeenCalled();
expect(setActiveRequestId).not.toHaveBeenCalled();
expect(setCapture).not.toHaveBeenCalled();
expect(sendResponse).not.toHaveBeenCalled();
});
it("ignores STOP for a mismatched requestId", () => {
const dispose = vi.fn();
const onStop = vi.fn();
const needsAsync = handleRecordContentMessage(
{ type: RECORD_STOP, requestId: "rec-other" },
{
activeRequestId: "rec-1",
capture: { dispose },
setActiveRequestId: vi.fn(),
setCapture: vi.fn(),
onStart: vi.fn(),
onStop,
},
vi.fn(),
);
expect(needsAsync).toBe(false);
expect(dispose).not.toHaveBeenCalled();
expect(onStop).not.toHaveBeenCalled();
});
});
describe("record-capture semantic", () => {
let steps: RecordStepPayload[];
@@ -174,6 +124,19 @@ describe("record-capture semantic", () => {
]);
});
it("does not emit page navigation from a child Document capture", () => {
const originalUrl = location.href;
const capture = startRecordCapture("rec-child", (step) => steps.push(step), {
captureNavigation: false,
});
history.pushState({}, "", "#inside-frame");
expect(steps).toEqual([]);
capture.dispose();
history.replaceState({}, "", originalUrl);
});
it("does not record clicks on anonymous layout divs", () => {
document.body.innerHTML = `
<div id="chrome">page chrome</div>
@@ -228,6 +191,7 @@ describe("record-capture semantic", () => {
expect(steps[0]).toMatchObject({
op: "hover",
target: { role: "button", name: "Open user navigation menu" },
geometry: { tag: "button", rect: { x: 900, y: 8, w: 32, h: 32 } },
});
expect(steps[1]).toMatchObject({
op: "click",
@@ -0,0 +1,102 @@
import { afterEach, describe, expect, it, vi } from "vitest";
import type { RecordFramePortMessage } from "@/lib/recording/frame-bridge";
import { RECORD_FRAME_START } from "@/lib/recording/frame-bridge";
import { RECORD_DOCUMENT_ATTRIBUTE } from "@/shared/recording-document-identity";
import { RecordFrameAgent } from "../recording/frame-agent";
class PortListeners<T extends (...args: never[]) => unknown> {
readonly values = new Set<T>();
addListener = (listener: T) => this.values.add(listener);
removeListener = (listener: T) => this.values.delete(listener);
}
function portHarness() {
const onMessage = new PortListeners<(message: unknown) => void>();
const onDisconnect = new PortListeners<() => void>();
const outbound: RecordFramePortMessage[] = [];
const port = {
name: "bsk-record-frame",
onMessage,
onDisconnect,
postMessage(message: RecordFramePortMessage) {
outbound.push(message);
if (message.type === "ready") {
queueMicrotask(() => {
for (const listener of onMessage.values) {
listener({
type: "ready_ack",
requestId: message.requestId,
producerId: message.producerId,
});
}
});
}
},
disconnect: vi.fn(),
} as unknown as chrome.runtime.Port;
return {
port,
outbound,
receive(message: RecordFramePortMessage) {
for (const listener of onMessage.values) listener(message);
},
};
}
describe("RecordFrameAgent", () => {
afterEach(() => {
document.documentElement.removeAttribute(RECORD_DOCUMENT_ATTRIBUTE);
document.body.replaceChildren();
});
it("keeps a failed stop retryable and flushes the final dirty fill", async () => {
const harness = portHarness();
const sendMessage = vi
.fn<(message: { sequence: number }) => Promise<unknown>>()
.mockRejectedValueOnce(new Error("offline"))
.mockRejectedValueOnce(new Error("still offline"))
.mockImplementation(async (message) => ({ ok: true, sequence: message.sequence }));
vi.stubGlobal("chrome", {
runtime: {
connect: vi.fn(() => harness.port),
sendMessage,
},
});
document.body.innerHTML = `<label for="draft">Draft</label><input id="draft" />`;
const agent = new RecordFrameAgent();
await expect(
agent.start({ type: RECORD_FRAME_START, requestId: "rec-1", startedAtMs: 10 }),
).resolves.toEqual({ ok: true });
expect(document.documentElement.hasAttribute(RECORD_DOCUMENT_ATTRIBUTE)).toBe(true);
const input = document.querySelector("input")!;
input.dispatchEvent(new FocusEvent("focusin", { bubbles: true }));
input.value = "final value";
input.dispatchEvent(new Event("input", { bubbles: true }));
harness.receive({ type: "stop", requestId: "rec-1", commandId: "stop-1" });
await vi.waitFor(() =>
expect(harness.outbound).toContainEqual({
type: "stopped",
requestId: "rec-1",
commandId: "stop-1",
ok: false,
error: "failed to deliver one or more recorded steps",
}),
);
expect(document.documentElement.hasAttribute(RECORD_DOCUMENT_ATTRIBUTE)).toBe(true);
harness.receive({ type: "stop", requestId: "rec-1", commandId: "stop-2" });
await vi.waitFor(() =>
expect(harness.outbound).toContainEqual({
type: "stopped",
requestId: "rec-1",
commandId: "stop-2",
ok: true,
}),
);
expect(sendMessage).toHaveBeenCalledTimes(3);
expect(document.documentElement.hasAttribute(RECORD_DOCUMENT_ATTRIBUTE)).toBe(false);
});
});
@@ -0,0 +1,36 @@
import { describe, expect, it, vi } from "vitest";
import type { RecordStepMessage } from "@/lib/record-bridge";
import { RecordStepDelivery } from "../record-step-delivery";
describe("RecordStepDelivery", () => {
it("retries from the first unacknowledged sequence before sending later steps", async () => {
const sent: RecordStepMessage[] = [];
const send = vi.fn(async (message: RecordStepMessage) => {
sent.push(message);
if (sent.length === 1) throw new Error("service worker unavailable");
return { ok: true, sequence: message.sequence };
});
const delivery = new RecordStepDelivery("rec-ordered", send, "document-1");
delivery.enqueue({ op: "click", target: { tag: "button", name: "First" } });
delivery.enqueue({ op: "click", target: { tag: "button", name: "Second" } });
await expect(delivery.flush()).resolves.toBe(true);
expect(sent.map((message) => message.sequence)).toEqual([1, 1, 2]);
expect(sent.map((message) => message.step.target?.name)).toEqual(["First", "First", "Second"]);
});
it("keeps unacknowledged steps pending for a later flush", async () => {
const send = vi
.fn<(message: RecordStepMessage) => Promise<unknown>>()
.mockRejectedValueOnce(new Error("offline"))
.mockRejectedValueOnce(new Error("still offline"))
.mockImplementation(async (message) => ({ ok: true, sequence: message.sequence }));
const delivery = new RecordStepDelivery("rec-retry", send, "document-1");
delivery.enqueue({ op: "click", target: { tag: "button", name: "Save" } });
await expect(delivery.flush()).resolves.toBe(false);
await expect(delivery.flush()).resolves.toBe(true);
expect(send).toHaveBeenCalledTimes(3);
});
});
+76 -150
View File
@@ -1,9 +1,9 @@
import {
type CaptureTargetDescriptor,
describeEventTarget,
describeTarget,
resolveClickableElement,
resolveHoverElement,
type TargetDescriptor,
} from "@/lib/describe-target";
import {
evaluateHoverTrigger,
@@ -12,22 +12,8 @@ import {
hasDirectHoverInteractiveSignal,
hasStrongHoverExpansionSignal,
} from "@/lib/hover-trigger-policy";
import {
isRecordCancelMessage,
isRecordStartMessage,
isRecordStopMessage,
RECORD_CANCEL,
RECORD_START,
RECORD_STEP,
RECORD_STOP,
type RecordCancelMessage,
type RecordStartAck,
type RecordStartMessage,
type RecordStepPayload,
type RecordStopAck,
type RecordStopMessage,
} from "@/lib/record-bridge";
import { shouldRecordPress } from "@/lib/trace-reducer";
import type { RecordStepPayload } from "@/lib/record-bridge";
import { shouldRecordPress } from "@/lib/recording/draft-policy";
import {
closestHoverSurfaceCandidate,
collectHoverSurfaceStates,
@@ -37,42 +23,21 @@ import {
isLikelyHoverSurfaceOwner,
} from "./record-hover-surface";
const pendingStepSends = new Map<string, Set<Promise<boolean>>>();
const failedStepDeliveries = new Set<string>();
const knownRecordRequests = new Set<string>();
function sendRecordStep(requestId: string, step: RecordStepPayload): void {
const payload = { type: RECORD_STEP, requestId, step };
const pending = Promise.resolve(chrome.runtime.sendMessage(payload)).then(
() => true,
() => {
failedStepDeliveries.add(requestId);
return false;
},
);
const sends = pendingStepSends.get(requestId) ?? new Set<Promise<boolean>>();
sends.add(pending);
pendingStepSends.set(requestId, sends);
void pending.then(() => {
sends.delete(pending);
if (sends.size === 0) pendingStepSends.delete(requestId);
});
}
export interface RecordCaptureController {
dispose(): void;
}
interface FillSession {
element: FillableElement;
target: TargetDescriptor;
target: CaptureTargetDescriptor;
baselineValue: string;
lastValue: string;
pendingCommit?: "enter" | "suggestion" | "blur";
}
interface HoverCandidate {
element: Element;
target: TargetDescriptor;
target: CaptureTargetDescriptor;
recordedAt: number;
score: number;
eligible: boolean;
@@ -157,6 +122,24 @@ function nearbyFillableFromSearchChrome(target: Element): FillableElement | null
return null;
}
function captureGeometry(el: Element): RecordStepPayload["geometry"] {
const rect = el.getBoundingClientRect();
return {
rect: { x: rect.x, y: rect.y, w: rect.width, h: rect.height },
tag: el.tagName.toLowerCase(),
};
}
function geometryForEventTarget(
target: EventTarget | null,
): RecordStepPayload["geometry"] | undefined {
if (!(target instanceof Element)) return undefined;
const clickable = target.closest(
'a,button,input,select,textarea,[role="button"],[role="link"],[role="menuitem"],[contenteditable="true"]',
);
return captureGeometry(clickable ?? target);
}
function fillableValue(el: FillableElement): string {
if (el instanceof HTMLInputElement || el instanceof HTMLTextAreaElement) {
return el.value;
@@ -232,7 +215,10 @@ function collectHoverTriggerLabelText(root: Element): string {
return normalizeLabelText(text);
}
function compactHoverTargetName(el: Element, desc: TargetDescriptor): TargetDescriptor {
function compactHoverTargetName(
el: Element,
desc: CaptureTargetDescriptor,
): CaptureTargetDescriptor {
if (!desc.name) return desc;
const fullText = normalizeLabelText(el.textContent ?? "");
const compactName = collectHoverTriggerLabelText(el);
@@ -250,7 +236,7 @@ function compactHoverTargetName(el: Element, desc: TargetDescriptor): TargetDesc
return desc;
}
function isWeakHoverTarget(target: TargetDescriptor): boolean {
function isWeakHoverTarget(target: CaptureTargetDescriptor): boolean {
return !target.role && !target.name && target.tag === "div";
}
@@ -261,9 +247,9 @@ function looksLikeAvatarElement(el: Element): boolean {
function normalizeHoverTarget(
el: Element,
desc: TargetDescriptor,
desc: CaptureTargetDescriptor,
decision: HoverTriggerDecision,
): TargetDescriptor {
): CaptureTargetDescriptor {
if (desc.role === "img" && !desc.name && looksLikeAvatarElement(el)) {
return { ...desc, name: "image" };
}
@@ -277,7 +263,7 @@ function normalizeHoverTarget(
return desc;
}
function hoverTriggerSignals(el: Element, desc: TargetDescriptor) {
function hoverTriggerSignals(el: Element, desc: CaptureTargetDescriptor) {
if (!(el instanceof HTMLElement)) return null;
const style = hoverTriggerStyle(el);
return {
@@ -373,6 +359,7 @@ function scheduleInputCompletionCommit(
export function startRecordCapture(
_requestId: string,
sendStep: (step: RecordStepPayload) => void,
options: { captureNavigation?: boolean } = {},
): RecordCaptureController {
const emitStep = (step: RecordStepPayload) => {
sendStep({ page_url: location.href, ...step });
@@ -411,14 +398,17 @@ export function startRecordCapture(
emitStep({
op: "fill",
target: session.target,
geometry: captureGeometry(session.element),
value,
commit: session.pendingCommit ?? "blur",
...(isPassword ? { redacted: true } : {}),
});
};
const commitFillSession = () => {
const commitFillSession = (commit: "enter" | "suggestion" | "blur" = "blur") => {
if (!fillSession || composing) return;
const session = fillSession;
session.pendingCommit = commit;
fillSession = null;
emitFill(session);
committedValues.set(session.element, session.lastValue);
@@ -624,6 +614,7 @@ export function startRecordCapture(
emitStep({
op: "hover",
target: hover.target,
geometry: captureGeometry(hover.element),
});
emittedHoverElements.add(hover.element);
};
@@ -678,6 +669,7 @@ export function startRecordCapture(
emitStep({
op: "click",
target,
geometry: geometryForEventTarget(eventTarget(event)),
expects_navigation: true,
});
};
@@ -779,7 +771,7 @@ export function startRecordCapture(
scheduleInputCompletionCommit(
sessionElement,
syncFillSessionValue,
commitFillSession,
() => commitFillSession("suggestion"),
(el) => fillSession?.element === el,
);
return;
@@ -839,6 +831,7 @@ export function startRecordCapture(
emitStep({
op: "select",
target: desc,
geometry: captureGeometry(target),
values,
labels,
expects_navigation: true,
@@ -879,7 +872,7 @@ export function startRecordCapture(
};
}
if (fillable) {
commitFillSession();
commitFillSession(event.key === "Enter" ? "enter" : "blur");
}
const desc = describeEventTarget(target);
if (!desc && !event.key) return;
@@ -887,7 +880,7 @@ export function startRecordCapture(
emitStep({
op: "press",
key: event.key,
...(desc ? { target: desc } : {}),
...(desc ? { target: desc, geometry: geometryForEventTarget(target) } : {}),
...(modifiers.length ? { modifiers } : {}),
expects_navigation: event.key === "Enter",
});
@@ -903,23 +896,33 @@ export function startRecordCapture(
document.addEventListener("change", onChange, true);
document.addEventListener("keydown", onKeyDown, true);
const urlObserver = new MutationObserver(() => emitNavigateIfChanged());
urlObserver.observe(document, { subtree: true, childList: true });
const captureNavigation = options.captureNavigation ?? true;
const urlObserver = captureNavigation
? new MutationObserver(() => emitNavigateIfChanged())
: null;
urlObserver?.observe(document, { subtree: true, childList: true });
const onUrlEvent = () => emitNavigateIfChanged();
window.addEventListener("hashchange", onUrlEvent);
window.addEventListener("popstate", onUrlEvent);
window.addEventListener("pagehide", commitFillSession);
// Wrap: passing commitFillSession directly would forward the DOM event as
// the `commit` argument and stamp it onto the recorded fill step.
const onPageHide = () => commitFillSession();
if (captureNavigation) {
window.addEventListener("hashchange", onUrlEvent);
window.addEventListener("popstate", onUrlEvent);
}
window.addEventListener("pagehide", onPageHide);
const originalPushState = history.pushState;
const originalReplaceState = history.replaceState;
history.pushState = function (...args: Parameters<History["pushState"]>) {
originalPushState.apply(this, args);
emitNavigateIfChanged(navigationActionPending ? true : undefined);
};
history.replaceState = function (...args: Parameters<History["replaceState"]>) {
originalReplaceState.apply(this, args);
emitNavigateIfChanged(navigationActionPending ? true : undefined);
};
if (captureNavigation) {
history.pushState = function (...args: Parameters<History["pushState"]>) {
originalPushState.apply(this, args);
emitNavigateIfChanged(navigationActionPending ? true : undefined);
};
history.replaceState = function (...args: Parameters<History["replaceState"]>) {
originalReplaceState.apply(this, args);
emitNavigateIfChanged(navigationActionPending ? true : undefined);
};
}
return {
dispose() {
@@ -933,93 +936,16 @@ export function startRecordCapture(
document.removeEventListener("compositionend", onCompositionEnd, true);
document.removeEventListener("change", onChange, true);
document.removeEventListener("keydown", onKeyDown, true);
urlObserver.disconnect();
window.removeEventListener("hashchange", onUrlEvent);
window.removeEventListener("popstate", onUrlEvent);
window.removeEventListener("pagehide", commitFillSession);
history.pushState = originalPushState;
history.replaceState = originalReplaceState;
urlObserver?.disconnect();
if (captureNavigation) {
window.removeEventListener("hashchange", onUrlEvent);
window.removeEventListener("popstate", onUrlEvent);
}
window.removeEventListener("pagehide", onPageHide);
if (captureNavigation) {
history.pushState = originalPushState;
history.replaceState = originalReplaceState;
}
},
};
}
export type RecordContentMessage = RecordStartMessage | RecordStopMessage | RecordCancelMessage;
export function isRecordContentMessage(msg: unknown): msg is RecordContentMessage {
return isRecordStartMessage(msg) || isRecordStopMessage(msg) || isRecordCancelMessage(msg);
}
export function handleRecordContentMessage(
message: RecordContentMessage,
state: {
activeRequestId: string | null;
capture: RecordCaptureController | null;
setActiveRequestId(id: string | null): void;
setCapture(capture: RecordCaptureController | null): void;
onStart(requestId: string, startedAtMs?: number): void;
onStop(): void;
},
sendResponse?: (response: RecordStartAck | RecordStopAck) => void,
): boolean {
if (isRecordStartMessage(message)) {
if (!knownRecordRequests.has(message.requestId)) {
knownRecordRequests.add(message.requestId);
failedStepDeliveries.delete(message.requestId);
}
state.capture?.dispose();
state.setCapture(
startRecordCapture(message.requestId, (step) => {
sendRecordStep(message.requestId, step);
}),
);
state.setActiveRequestId(message.requestId);
state.onStart(message.requestId, message.startedAtMs);
sendResponse?.({ ok: true });
return sendResponse !== undefined;
}
if (isRecordStopMessage(message) || isRecordCancelMessage(message)) {
// Require an active recording that matches this requestId — otherwise a
// stray STOP/CANCEL (e.g. after teardown) would still run finishStop and
// clear overlay state even though nothing was capturing.
if (!state.activeRequestId || state.activeRequestId !== message.requestId) {
return false;
}
state.capture?.dispose();
state.setCapture(null);
const finishStop = () => {
state.onStop();
state.setActiveRequestId(null);
};
if (isRecordStopMessage(message) && sendResponse) {
const pending = [...(pendingStepSends.get(message.requestId) ?? [])];
void Promise.all(pending).then((delivered) => {
finishStop();
const succeeded = delivered.every(Boolean) && !failedStepDeliveries.has(message.requestId);
if (succeeded) {
failedStepDeliveries.delete(message.requestId);
knownRecordRequests.delete(message.requestId);
}
sendResponse(
succeeded
? { ok: true }
: {
ok: false,
error: "failed to deliver one or more recorded steps",
},
);
});
return true;
}
if (isRecordCancelMessage(message)) {
failedStepDeliveries.delete(message.requestId);
knownRecordRequests.delete(message.requestId);
}
finishStop();
return false;
}
return false;
}
export { RECORD_CANCEL, RECORD_START, RECORD_STEP, RECORD_STOP };
@@ -0,0 +1,65 @@
import {
isAcceptedRecordStepAck,
RECORD_STEP,
type RecordStepMessage,
type RecordStepPayload,
} from "@/lib/record-bridge";
export type SendRecordStepMessage = (message: RecordStepMessage) => Promise<unknown>;
export class RecordStepDelivery {
readonly #requestId: string;
readonly #producerId: string;
readonly #send: SendRecordStepMessage;
readonly #pending = new Map<number, RecordStepPayload>();
#nextSequence = 1;
#tail = Promise.resolve();
constructor(
requestId: string,
send: SendRecordStepMessage = (message) => chrome.runtime.sendMessage(message),
producerId: string = crypto.randomUUID(),
) {
this.#requestId = requestId;
this.#producerId = producerId;
this.#send = send;
}
enqueue(step: RecordStepPayload): void {
const sequence = this.#nextSequence;
this.#nextSequence += 1;
this.#pending.set(sequence, step);
void this.#schedule();
}
async flush(): Promise<boolean> {
await this.#tail;
await this.#schedule();
return this.#pending.size === 0;
}
#schedule(): Promise<void> {
const drain = () => this.#drain();
this.#tail = this.#tail.then(drain, drain);
return this.#tail;
}
async #drain(): Promise<void> {
for (const [sequence, step] of this.#pending) {
const message: RecordStepMessage = {
type: RECORD_STEP,
requestId: this.#requestId,
producerId: this.#producerId,
sequence,
step,
};
try {
const response = await this.#send(message);
if (!isAcceptedRecordStepAck(response, sequence)) return;
this.#pending.delete(sequence);
} catch {
return;
}
}
}
}
@@ -0,0 +1,209 @@
import {
markRecordingDocument,
type RecordingDocumentMarker,
waitForDocumentElement,
} from "@/lib/recording/document-marker";
import {
isRecordFrameStartMessage,
RECORD_FRAME_PORT,
RECORD_FRAME_QUERY,
RECORD_FRAME_START,
type RecordFramePortMessage,
type RecordFrameQueryResponse,
type RecordFrameStartMessage,
} from "@/lib/recording/frame-bridge";
import { type RecordCaptureController, startRecordCapture } from "../record-capture";
import { RecordStepDelivery } from "../record-step-delivery";
interface ActiveFrameRecording {
requestId: string;
producerId: string;
port: chrome.runtime.Port;
delivery: RecordStepDelivery;
capture: RecordCaptureController;
marker: RecordingDocumentMarker;
}
const RECORD_FRAME_REGISTER_TIMEOUT_MS = 5_000;
export class RecordFrameAgent {
#active: ActiveFrameRecording | null = null;
#startPromise: Promise<{ ok: true } | { ok: false; error: string }> | null = null;
#disposed = false;
async start(
message: RecordFrameStartMessage,
): Promise<{ ok: true } | { ok: false; error: string }> {
if (this.#active?.requestId === message.requestId) return { ok: true };
if (this.#disposed) return { ok: false, error: "recording document is disposed" };
if (this.#startPromise) {
await this.#startPromise;
return this.start(message);
}
if (this.#active) this.cancel(this.#active.requestId);
const starting = this.#start(message);
this.#startPromise = starting;
try {
return await starting;
} finally {
if (this.#startPromise === starting) this.#startPromise = null;
}
}
cancel(requestId: string): void {
const active = this.#active;
if (!active || active.requestId !== requestId) return;
active.capture.dispose();
active.marker.restore();
active.port.disconnect();
this.#active = null;
}
dispose(): void {
this.#disposed = true;
if (this.#active) this.cancel(this.#active.requestId);
}
async #start(
message: RecordFrameStartMessage,
): Promise<{ ok: true } | { ok: false; error: string }> {
const producerId = crypto.randomUUID();
const root = await waitForDocumentElement();
if (this.#disposed) return { ok: false, error: "recording document is disposed" };
const marker = markRecordingDocument(producerId, root);
const port = chrome.runtime.connect({ name: RECORD_FRAME_PORT });
try {
await registerPort(port, message.requestId, producerId);
} catch {
marker.restore();
port.disconnect();
return { ok: false, error: "failed to register recording document" };
}
if (this.#disposed) {
marker.restore();
port.disconnect();
return { ok: false, error: "recording document is disposed" };
}
const delivery = new RecordStepDelivery(message.requestId, undefined, producerId);
const capture = startRecordCapture(
message.requestId,
(step) => {
marker.ensure();
delivery.enqueue(step);
},
{ captureNavigation: window.top === window },
);
const active: ActiveFrameRecording = {
requestId: message.requestId,
producerId,
port,
delivery,
capture,
marker,
};
this.#active = active;
port.onMessage.addListener((raw: unknown) => {
const command = raw as Partial<RecordFramePortMessage>;
if (command.type === "stop" && command.requestId === active.requestId) {
void this.#stop(active, command.commandId);
} else if (command.type === "cancel" && command.requestId === active.requestId) {
this.cancel(active.requestId);
}
});
port.onDisconnect.addListener(() => {
if (this.#active !== active) return;
active.capture.dispose();
active.marker.restore();
this.#active = null;
});
return { ok: true };
}
async #stop(active: ActiveFrameRecording, commandId: string | undefined): Promise<void> {
if (!commandId || this.#active !== active) return;
active.capture.dispose();
const succeeded = await active.delivery.flush();
try {
active.port.postMessage({
type: "stopped",
requestId: active.requestId,
commandId,
ok: succeeded,
...(succeeded ? {} : { error: "failed to deliver one or more recorded steps" }),
} satisfies RecordFramePortMessage);
} catch {
return;
}
if (!succeeded || this.#active !== active) return;
active.marker.restore();
this.#active = null;
active.port.disconnect();
}
}
function registerPort(
port: chrome.runtime.Port,
requestId: string,
producerId: string,
): Promise<void> {
return new Promise((resolve, reject) => {
const timer = setTimeout(() => {
cleanup();
reject(new Error("recording document registration timed out"));
}, RECORD_FRAME_REGISTER_TIMEOUT_MS);
const cleanup = () => {
clearTimeout(timer);
port.onMessage.removeListener(onMessage);
port.onDisconnect.removeListener(onDisconnect);
};
const onMessage = (raw: unknown) => {
const message = raw as Partial<RecordFramePortMessage>;
if (
message.type !== "ready_ack" ||
message.requestId !== requestId ||
message.producerId !== producerId
) {
return;
}
cleanup();
resolve();
};
const onDisconnect = () => {
cleanup();
reject(new Error("recording document port disconnected"));
};
port.onMessage.addListener(onMessage);
port.onDisconnect.addListener(onDisconnect);
port.postMessage({ type: "ready", requestId, producerId } satisfies RecordFramePortMessage);
});
}
export function attachRecordFrameAgent(): () => void {
const agent = new RecordFrameAgent();
const onMessage = (
message: unknown,
_sender: chrome.runtime.MessageSender,
sendResponse: (response: { ok: true } | { ok: false; error: string }) => void,
) => {
if (!isRecordFrameStartMessage(message)) return false;
void agent.start(message).then(sendResponse);
return true;
};
chrome.runtime.onMessage.addListener(onMessage);
void chrome.runtime
.sendMessage({ type: RECORD_FRAME_QUERY })
.then((response: RecordFrameQueryResponse | undefined) => {
if (!response?.active || !response.requestId || response.startedAtMs === undefined) return;
return agent.start({
type: RECORD_FRAME_START,
requestId: response.requestId,
startedAtMs: response.startedAtMs,
});
})
.catch(() => {});
return () => {
chrome.runtime.onMessage.removeListener(onMessage);
agent.dispose();
};
}
+25 -21
View File
@@ -21,6 +21,7 @@ import {
type OverlayMode,
} from "@/lib/overlay-bridge";
import { POPUP_PORT_NAME, type PopupInbound, type PopupOutbound } from "@/lib/popup-bridge";
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";
@@ -38,6 +39,7 @@ import {
attachRecordFinishListener,
attachRecordQueryListener,
attachRecordStepListener,
type RecordRuntimeDeps,
} from "@/tools/record";
import { detectBrowserMeta } from "@/transport/handshake";
import type { Transport } from "@/transport/transport";
@@ -145,10 +147,33 @@ export default defineBackground(() => {
void sessionsLive.syncFromManager();
},
});
const recordDeps = {
tabsApi: chrome.tabs,
cdp,
frameCoordinator: recordFrameCoordinator,
sendToTab: (tabId: number, msg: Parameters<typeof chrome.tabs.sendMessage>[1]) =>
chrome.tabs.sendMessage(tabId, msg),
bypassOverlay: async (tabId: number, enabled: boolean) => {
try {
await chrome.tabs.sendMessage(tabId, {
type: OVERLAY_AUTOMATION_BYPASS,
enabled,
});
} catch {
// Content script may be unavailable on restricted pages.
}
},
} satisfies RecordRuntimeDeps;
recordFrameCoordinator.attach();
attachRecordStepListener(recordDeps);
attachRecordFinishListener(recordDeps);
attachRecordQueryListener(recordDeps);
const dispatcher = new ToolDispatcher({
transport,
sessions,
cdp,
recording: recordDeps,
onSessionsChanged: onOverlaySessionStateChanged,
onBrowserControlResumed,
approveBorrow: (ctx) =>
@@ -170,27 +195,6 @@ export default defineBackground(() => {
}),
});
dispatcher.start();
const recordDeps = {
tabsApi: chrome.tabs,
sendToTab: (tabId: number, msg: Parameters<typeof chrome.tabs.sendMessage>[1]) =>
chrome.tabs.sendMessage(tabId, msg),
bypassOverlay: async (tabId: number, enabled: boolean) => {
try {
await chrome.tabs.sendMessage(tabId, {
type: OVERLAY_AUTOMATION_BYPASS,
enabled,
});
} catch {
// Content script may be unavailable on restricted pages.
}
},
};
// Message listeners stay up (cheap; fire only for record message types).
// Tab / webNavigation observation attaches lazily while a recording is
// active — see ensureBrowserObservationListeners in tools/record.ts.
attachRecordStepListener();
attachRecordFinishListener(recordDeps);
attachRecordQueryListener(recordDeps);
if (typeof chrome.notifications?.onClicked?.addListener === "function") {
attachBorrowNotificationClickHandler({
onClicked: chrome.notifications.onClicked,
+34 -45
View File
@@ -11,11 +11,6 @@ import { createHelpRequestData } from "@/content/help-request";
import overlayCss from "@/content/overlay.css?inline";
import { OverlayController, shouldShowAgentControlOverlay } from "@/content/overlay-controller";
import { RecordOverlay } from "@/content/RecordOverlay";
import {
handleRecordContentMessage,
isRecordContentMessage,
type RecordCaptureController,
} from "@/content/record-capture";
import {
type CaptureSuppressAck,
type CaptureSuppressMessage,
@@ -45,11 +40,17 @@ import {
} from "@/lib/overlay-bridge";
import { sendInterrupt } from "@/lib/overlay-interrupt-client";
import {
isRecordCancelMessage,
isRecordStartMessage,
isRecordStopMessage,
RECORD_FINISH,
RECORD_QUERY,
type RecordCancelMessage,
type RecordQueryResponse,
type RecordStartAck,
type RecordStartMessage,
type RecordStopAck,
type RecordStopMessage,
} from "@/lib/record-bridge";
import type {
BorrowCancelMessage,
@@ -69,7 +70,6 @@ export default defineContentScript({
if (window.top !== window) return;
const overlays = new OverlayController();
let recordCapture: RecordCaptureController | null = null;
let activeRecordRequestId: string | null = null;
let reactRoot: ReactDOM.Root | null = null;
let overlayHost: HTMLElement | null = null;
@@ -228,8 +228,6 @@ export default defineContentScript({
if (previousHelp) {
void sendHelpFinish(previousHelp.id, "cancelled");
}
recordCapture?.dispose();
recordCapture = null;
activeRecordRequestId = null;
renderAll();
}
@@ -259,6 +257,9 @@ export default defineContentScript({
| HelpRequestMessage
| HelpCancelMessage
| CaptureSuppressMessage
| RecordStartMessage
| RecordStopMessage
| RecordCancelMessage
| OverlayAgentOverlayResetMessage
| OverlayAgentStateMessage
| OverlayAutomationBypassMessage,
@@ -269,41 +270,32 @@ export default defineContentScript({
return captureSuppress.handleMessage(message, sendResponse);
}
if (isRecordContentMessage(message)) {
const needsAsync = handleRecordContentMessage(
message,
{
activeRequestId: activeRecordRequestId,
capture: recordCapture,
setActiveRequestId: (id) => {
activeRecordRequestId = id;
},
setCapture: (capture) => {
recordCapture = capture;
},
onStart: (requestId, startedAtMs) => {
overlays.setAgentRecordRequest({
id: requestId,
...(typeof startedAtMs === "number" ? { startedAtMs } : {}),
onFinish: () => {
void chrome.runtime.sendMessage({
type: RECORD_FINISH,
requestId,
});
},
});
renderAll();
},
onStop: () => {
overlays.clearAgentRecordRequest(activeRecordRequestId ?? undefined);
renderAll();
},
if (isRecordStartMessage(message)) {
activeRecordRequestId = message.requestId;
overlays.setAgentRecordRequest({
id: message.requestId,
...(typeof message.startedAtMs === "number" ? { startedAtMs: message.startedAtMs } : {}),
onFinish: () => {
void chrome.runtime.sendMessage({
type: RECORD_FINISH,
requestId: message.requestId,
});
},
sendResponse as unknown as
| ((response: RecordStartAck | RecordStopAck) => void)
| undefined,
);
return needsAsync;
});
renderAll();
(sendResponse as unknown as (response: RecordStartAck) => void)({ ok: true });
return false;
}
if (isRecordStopMessage(message) || isRecordCancelMessage(message)) {
if (activeRecordRequestId !== message.requestId) return false;
overlays.clearAgentRecordRequest(message.requestId);
activeRecordRequestId = null;
renderAll();
if (isRecordStopMessage(message)) {
(sendResponse as unknown as (response: RecordStopAck) => void)({ ok: true });
}
return false;
}
if (
@@ -526,9 +518,6 @@ export default defineContentScript({
chrome.runtime.onMessage.removeListener(onMessage);
chrome.storage.onChanged.removeListener(onStorageChange);
window.removeEventListener("pageshow", onPageShow);
// Restore history hooks / remove capture listeners before the CS unloads.
recordCapture?.dispose();
recordCapture = null;
activeRecordRequestId = null;
});
},
@@ -200,6 +200,9 @@ describe("App", () => {
expect(copied).toContain("--browser 03c3e47f");
expect(copied).toContain('--purpose "发布 wiki 文档"');
expect(copied).not.toMatch(/bsk record start[^\n]*--url/);
expect(copied).toContain("./trace");
expect(copied).toContain("trace.json");
expect(copied).toContain("states/");
// Button label stays static; a transient toast confirms the copy.
expect(copyButton.textContent).toContain("复制录制指令");
await waitFor(() => expect(screen.getByRole("status").textContent).toContain("已复制"));
@@ -0,0 +1,14 @@
import { attachRecordFrameAgent } from "@/content/recording/frame-agent";
export default defineContentScript({
matches: ["<all_urls>"],
runAt: "document_start",
allFrames: true,
matchAboutBlank: true,
matchOriginAsFallback: true,
main(ctx) {
const dispose = attachRecordFrameAgent();
ctx.onInvalidated(dispose);
},
});
@@ -26,20 +26,20 @@ function handshake(
}
describe("computeConnectedState (protocol-based compat)", () => {
it("returns connected when protocol strings match", () => {
expect(computeConnectedState(handshake("1.0", "1.0"), MIN_COMPATIBLE_PROTOCOL)).toEqual({
it("returns connected when daemon protocol equals extension protocol", () => {
expect(computeConnectedState(handshake("1.1", "1.0"), MIN_COMPATIBLE_PROTOCOL)).toEqual({
kind: "connected",
});
});
it("returns version_skew when daemon protocol minor is newer", () => {
expect(computeConnectedState(handshake("1.1", "1.0"))).toEqual({
expect(computeConnectedState(handshake("1.2", "1.0"))).toEqual({
kind: "version_skew",
});
});
it("returns version_skew when daemon protocol string differs but floor is satisfied", () => {
expect(computeConnectedState(handshake("1", "1.0"))).toEqual({
expect(computeConnectedState(handshake("1.1.0", "1.0"))).toEqual({
kind: "version_skew",
});
});
@@ -53,7 +53,7 @@ describe("computeConnectedState (protocol-based compat)", () => {
});
it("rejects when extension is below daemon min_compatible_protocol", () => {
const result = computeConnectedState(handshake("1.0", "1.5"));
const result = computeConnectedState(handshake("1.1", "1.5"));
expect(result.kind).toBe("rejected");
if (result.kind === "rejected") {
expect(result.reason).toContain("min_compatible_protocol");
@@ -65,7 +65,7 @@ describe("computeConnectedState (protocol-based compat)", () => {
const result = computeConnectedState({
server: "browser-skill-daemon",
version: "0.1.0",
protocol_version: "1.0",
protocol_version: "1.1",
min_compatible_peer: "0.1.0",
});
expect(result).toEqual({ kind: "connected" });
@@ -80,7 +80,7 @@ describe("computeConnectedState (protocol-based compat)", () => {
});
it("rejects malformed daemon min_compatible_protocol with a daemon-floor reason", () => {
const result = computeConnectedState(handshake("1.0", "not-a-protocol"));
const result = computeConnectedState(handshake("1.1", "not-a-protocol"));
expect(result.kind).toBe("rejected");
if (result.kind === "rejected") {
expect(result.reason).toContain("daemon min_compatible_protocol");
@@ -242,11 +242,11 @@ describe("ConnectionController connectionEnabled", () => {
const second = transport.send.mock.calls[1]?.[0] as { id: string };
expect(second.id).not.toBe(first.id);
transport.emitMessage({ id: first.id, result: handshake("1.0", "1.0") });
transport.emitMessage({ id: first.id, result: handshake("1.1", "1.0") });
await Promise.resolve();
expect(controller.snapshot().state).not.toBe("connected");
transport.emitMessage({ id: second.id, result: handshake("1.0", "1.0") });
transport.emitMessage({ id: second.id, result: handshake("1.1", "1.0") });
await vi.waitFor(() => expect(controller.snapshot().state).toBe("connected"));
});
});
@@ -0,0 +1,101 @@
import { afterEach, describe, expect, it, vi } from "vitest";
import type { CdpRunner } from "@/tools/shared";
import { waitForDocumentSettled } from "../recording/document-settle";
function quietCdp(): { cdp: CdpRunner; send: ReturnType<typeof vi.fn> } {
const send = vi.fn(async () => ({
result: { value: { idleMs: 1_000, readyState: "complete" } },
}));
return {
cdp: { send: send as unknown as CdpRunner["send"] },
send,
};
}
describe("waitForDocumentSettled", () => {
afterEach(() => {
vi.useRealTimers();
});
it("waits through the minimum observation floor before accepting a quiet page", async () => {
vi.useFakeTimers();
const { cdp, send } = quietCdp();
const settled = waitForDocumentSettled(cdp, { target: { tabId: 7 } });
await vi.advanceTimersByTimeAsync(180);
await expect(settled).resolves.toBe("quiet");
expect(send).toHaveBeenCalledTimes(3);
});
it("cancels before probing a superseded observation", async () => {
vi.useFakeTimers();
const { cdp, send } = quietCdp();
const controller = new AbortController();
const settled = waitForDocumentSettled(
cdp,
{ target: { tabId: 7 } },
{
signal: controller.signal,
},
);
controller.abort();
await vi.advanceTimersByTimeAsync(60);
await expect(settled).resolves.toBe("cancelled");
expect(send).not.toHaveBeenCalled();
});
it("probes a same-process iframe through its isolated execution context", async () => {
vi.useFakeTimers();
const send = vi.fn(async (_tabId: number, method: string) => {
if (method === "Page.createIsolatedWorld") return { executionContextId: 91 };
return { result: { value: { idleMs: 1_000, readyState: "complete" } } };
});
const cdp: CdpRunner = { send: send as unknown as CdpRunner["send"] };
const settled = waitForDocumentSettled(cdp, {
target: { tabId: 7 },
frameId: "child-frame",
});
await vi.advanceTimersByTimeAsync(180);
await expect(settled).resolves.toBe("quiet");
expect(send).toHaveBeenCalledWith(
7,
"Page.createIsolatedWorld",
expect.objectContaining({ frameId: "child-frame" }),
);
expect(send).toHaveBeenCalledWith(
7,
"Runtime.evaluate",
expect.objectContaining({ contextId: 91 }),
);
});
it("routes an OOPIF probe through its CDP target session", async () => {
vi.useFakeTimers();
const send = vi.fn();
const sendToTarget = vi.fn(async (_target, method: string) => {
if (method === "Page.createIsolatedWorld") return { executionContextId: 27 };
return { result: { value: { idleMs: 1_000, readyState: "complete" } } };
});
const cdp: CdpRunner = {
send: send as unknown as CdpRunner["send"],
sendToTarget: sendToTarget as unknown as NonNullable<CdpRunner["sendToTarget"]>,
};
const target = { tabId: 7, sessionId: "oopif-session" };
const settled = waitForDocumentSettled(cdp, { target, frameId: "oopif-frame" });
await vi.advanceTimersByTimeAsync(180);
await expect(settled).resolves.toBe("quiet");
expect(send).not.toHaveBeenCalled();
expect(sendToTarget).toHaveBeenCalledWith(
target,
"Runtime.evaluate",
expect.objectContaining({ contextId: 27 }),
);
});
});
@@ -0,0 +1,32 @@
import { describe, expect, it } from "vitest";
import {
isAcceptedRecordStepAck,
isRecordStepMessage,
RECORD_STEP,
type RecordStepMessage,
} from "../record-bridge";
const message: RecordStepMessage = {
type: RECORD_STEP,
requestId: "rec-1",
producerId: "document-1",
sequence: 1,
step: { op: "click", target: { tag: "button", name: "Submit" } },
};
describe("record step delivery protocol", () => {
it("requires a producer and a positive integer sequence", () => {
expect(isRecordStepMessage(message)).toBe(true);
expect(isRecordStepMessage({ ...message, producerId: "" })).toBe(false);
expect(isRecordStepMessage({ ...message, sequence: 0 })).toBe(false);
expect(isRecordStepMessage({ ...message, sequence: 1.5 })).toBe(false);
});
it("accepts only the acknowledgement for the sent sequence", () => {
expect(isAcceptedRecordStepAck({ ok: true, sequence: 2 }, 2)).toBe(true);
expect(isAcceptedRecordStepAck({ ok: true, sequence: 1 }, 2)).toBe(false);
expect(isAcceptedRecordStepAck({ ok: false, expectedSequence: 2, error: "gap" }, 2)).toBe(
false,
);
});
});
@@ -0,0 +1,213 @@
import { beforeEach, describe, expect, it, vi } from "vitest";
import { RECORD_FRAME_PORT, RECORD_FRAME_QUERY } from "../recording/frame-bridge";
import { RecordFrameCoordinator } from "../recording/frame-coordinator";
class ListenerSet<T extends (...args: never[]) => unknown> {
readonly listeners = new Set<T>();
addListener = (listener: T) => this.listeners.add(listener);
removeListener = (listener: T) => this.listeners.delete(listener);
}
function fakePort(sender: chrome.runtime.MessageSender) {
const messages: unknown[] = [];
const onMessage = new ListenerSet<(message: unknown) => void>();
const onDisconnect = new ListenerSet<() => void>();
return {
port: {
name: RECORD_FRAME_PORT,
sender,
onMessage,
onDisconnect,
postMessage: vi.fn((message: unknown) => messages.push(message)),
disconnect: vi.fn(),
} as unknown as chrome.runtime.Port,
messages,
receive(message: unknown) {
for (const listener of onMessage.listeners) listener(message);
},
};
}
describe("RecordFrameCoordinator", () => {
const onConnect = new ListenerSet<(port: chrome.runtime.Port) => void>();
const onMessage = new ListenerSet<
(
message: unknown,
sender: chrome.runtime.MessageSender,
sendResponse: (response: unknown) => void,
) => boolean
>();
beforeEach(() => {
onConnect.listeners.clear();
onMessage.listeners.clear();
vi.stubGlobal("chrome", {
runtime: { onConnect, onMessage },
});
});
it("answers child-document queries for the initial tab before it is armed", () => {
const coordinator = new RecordFrameCoordinator({
getAllFrames: async () => [],
sendToDocument: async () => ({ ok: true }),
});
coordinator.attach();
coordinator.begin("rec-1", 10, 3);
const query = (tabId: number) => {
let response: unknown;
for (const listener of onMessage.listeners) {
listener(
{ type: RECORD_FRAME_QUERY },
{
tab: { id: tabId },
frameId: 7,
documentId: "child-document",
} as chrome.runtime.MessageSender,
(value) => {
response = value;
},
);
}
return response;
};
expect(query(3)).toEqual({ active: true, requestId: "rec-1", startedAtMs: 10 });
expect(query(4)).toEqual({ active: false });
});
it("starts a child Document that appears after the initial frame snapshot", async () => {
let notifyFrameNavigation:
| ((frame: {
tabId: number;
frameId: number;
documentId?: string;
lifecycle: "committed" | "completed";
}) => void)
| undefined;
const sendToDocument = vi.fn(async () => ({ ok: true }));
const coordinator = new RecordFrameCoordinator({
getAllFrames: async () => [{ frameId: 0, documentId: "top-document" }],
sendToDocument,
subscribeFrameNavigation(listener) {
notifyFrameNavigation = listener;
return () => {};
},
});
coordinator.attach();
const onDocumentReady = vi.fn();
coordinator.begin("rec-1", 10, 3, onDocumentReady);
await coordinator.armTab("rec-1", 3);
sendToDocument.mockClear();
notifyFrameNavigation?.({
tabId: 3,
frameId: 7,
documentId: "late-child",
lifecycle: "completed",
});
await vi.waitFor(() => {
expect(sendToDocument).toHaveBeenCalledWith(
3,
{ type: "bsk-record-frame-start", requestId: "rec-1", startedAtMs: 10 },
{ documentId: "late-child" },
);
});
expect(onDocumentReady).not.toHaveBeenCalled();
const child = fakePort({
tab: { id: 3 },
frameId: 7,
documentId: "late-child",
} as chrome.runtime.MessageSender);
for (const listener of onConnect.listeners) listener(child.port);
child.receive({ type: "ready", requestId: "rec-1", producerId: "producer-1" });
notifyFrameNavigation?.({
tabId: 3,
frameId: 7,
documentId: "late-child",
lifecycle: "completed",
});
await vi.waitFor(() => {
expect(onDocumentReady).toHaveBeenCalledWith({
tabId: 3,
documentId: "late-child",
browserFrameId: 7,
producerId: "producer-1",
});
});
});
it("binds a producer to its sender Document and keeps final steps valid while stopping", async () => {
const sendToDocument = vi.fn(async () => ({ ok: true }));
const coordinator = new RecordFrameCoordinator({
getAllFrames: async () => [
{ frameId: 0, documentId: "top-document" },
{ frameId: 7, documentId: "child-document" },
],
sendToDocument,
});
coordinator.attach();
coordinator.begin("rec-1", 10, 3);
await expect(coordinator.armTab("rec-1", 3)).resolves.toBe(true);
const sender = {
tab: { id: 3 },
frameId: 7,
documentId: "child-document",
} as chrome.runtime.MessageSender;
const frame = fakePort(sender);
for (const listener of onConnect.listeners) listener(frame.port);
frame.receive({ type: "ready", requestId: "rec-1", producerId: "producer-1" });
expect(coordinator.sourceFor("rec-1", "producer-1", sender)).toEqual({
tabId: 3,
documentId: "child-document",
browserFrameId: 7,
producerId: "producer-1",
});
const stopping = coordinator.stop("rec-1");
expect(coordinator.sourceFor("rec-1", "producer-1", sender)).not.toBeNull();
const stop = frame.messages.find(
(message): message is { type: "stop"; commandId: string } =>
typeof message === "object" &&
message !== null &&
(message as { type?: string }).type === "stop",
);
expect(stop).toBeDefined();
frame.receive({
type: "stopped",
requestId: "rec-1",
commandId: stop!.commandId,
ok: true,
});
await expect(stopping).resolves.toBe(true);
expect(coordinator.sourceFor("rec-1", "producer-1", sender)).toBeNull();
});
it("rejects a producer used from another Document", async () => {
const coordinator = new RecordFrameCoordinator({
getAllFrames: async () => [{ frameId: 0, documentId: "top-document" }],
sendToDocument: async () => ({ ok: true }),
});
coordinator.attach();
coordinator.begin("rec-1", 10, 3);
await coordinator.armTab("rec-1", 3);
const sender = {
tab: { id: 3 },
frameId: 0,
documentId: "top-document",
} as chrome.runtime.MessageSender;
const frame = fakePort(sender);
for (const listener of onConnect.listeners) listener(frame.port);
frame.receive({ type: "ready", requestId: "rec-1", producerId: "producer-1" });
expect(
coordinator.sourceFor("rec-1", "producer-1", {
...sender,
documentId: "other-document",
}),
).toBeNull();
});
});
@@ -0,0 +1,70 @@
import { describe, expect, it } from "vitest";
import {
RECORD_DOCUMENT_ATTRIBUTE,
recordingDocumentMarkerValue,
} from "@/shared/recording-document-identity";
import { markRecordingDocument } from "../recording/document-marker";
import { ObservationNodeIndex } from "../recording/observation-capture";
describe("recording document marker", () => {
it("rejects a non-random Document identity", () => {
expect(() => markRecordingDocument("user@example.com")).toThrow(
"recording Document identity must be a random UUID",
);
});
it("restores the Document attribute exactly", () => {
document.documentElement.setAttribute(RECORD_DOCUMENT_ATTRIBUTE, "page-value");
const producerId = "123e4567-e89b-42d3-a456-426614174000";
const marker = markRecordingDocument(producerId);
expect(document.documentElement.getAttribute(RECORD_DOCUMENT_ATTRIBUTE)).toBe(
recordingDocumentMarkerValue(producerId),
);
marker.restore();
expect(document.documentElement.getAttribute(RECORD_DOCUMENT_ATTRIBUTE)).toBe("page-value");
document.documentElement.removeAttribute(RECORD_DOCUMENT_ATTRIBUTE);
});
it("indexes the CDP scope from the safe frame identity", () => {
const index = new ObservationNodeIndex({
rootFrameId: "root",
frames: [
{
frameId: "child",
target: { tabId: 3, sessionId: "oopif-session" },
recordingDocumentId: "producer-1",
},
],
matchNodes: [],
refs: [],
});
expect(index.documentScope("producer-1")).toEqual({
frameId: "child",
target: { tabId: 3, sessionId: "oopif-session" },
});
});
it("fails closed when a recording identity maps to multiple Documents", () => {
const index = new ObservationNodeIndex({
rootFrameId: "root",
frames: [
{
frameId: "left",
target: { tabId: 3, sessionId: "left-session" },
recordingDocumentId: "duplicate",
},
{
frameId: "right",
target: { tabId: 3, sessionId: "right-session" },
recordingDocumentId: "duplicate",
},
],
matchNodes: [],
refs: [],
});
expect(index.documentScope("duplicate")).toBeUndefined();
});
});
@@ -0,0 +1,146 @@
import { describe, expect, it } from "vitest";
import { ObservationNodeIndex } from "../recording/observation-capture";
import { RecordingObservationSession } from "../recording/observation-session";
import { RecordingStateRegistry } from "../recording/state-registry";
import { buildTraceV3 } from "../recording/trace-builder-v3";
import type { RecordingDraftStep } from "../recording/types";
const URL = "https://example.com/login";
function sessionWithInput(redactValues = false): RecordingObservationSession {
const session = new RecordingObservationSession({ redactValues });
const state = session.registry.register({
url: URL,
vomText: '@vom 1\ntextbox "Password" value="••••••" [ref=e1]',
});
session.cursor.lastSettled = {
stateId: state.id,
rootFrameId: "root",
index: new ObservationNodeIndex({
rootFrameId: "root",
matchNodes: [
{
frameId: "root",
backendNodeId: 42,
tag: "input",
rect: { x: 20, y: 40, w: 200, h: 30 },
localRect: { x: 20, y: 40, w: 200, h: 30 },
},
],
refs: [{ ref: "e1", backendNodeId: 42, role: "textbox", name: "Password", line: 1 }],
}),
url: URL,
};
return session;
}
function finalizedFillTrace(value: string, redactValues: boolean) {
const session = sessionWithInput(redactValues);
const draft: RecordingDraftStep = {
op: "fill",
captureTarget: { tag: "input", role: "textbox", name: "Password" },
value,
targetHint: {
geometry: { rect: { x: 20, y: 40, w: 200, h: 30 }, tag: "input" },
},
};
session.bindDraft(draft, 1);
draft.postStateId = draft.preStateId;
return buildTraceV3({
registry: session.registry,
drafts: [draft],
annotations: session.annotations,
startedAt: "2026-08-12T00:00:00.000Z",
stoppedBy: "user_finish",
bskVersion: "test",
redactValues,
});
}
function finalizedFillBody(value: string, redactValues: boolean): string {
return finalizedFillTrace(value, redactValues).states[0]!.body;
}
describe("record observation annotations", () => {
it("omits a fill literal when values are redacted", () => {
const secret = "hunter2-private";
const dumped = JSON.stringify(finalizedFillTrace(secret, true));
expect(dumped).toContain("step 1: fill");
expect(dumped).not.toContain(secret);
});
it("keeps ordinary fill details", () => {
expect(finalizedFillBody("ordinary text", false)).toContain('step 1: fill: "ordinary text"');
});
it("encodes title and URL so line breaks cannot corrupt state metadata", () => {
const registry = new RecordingStateRegistry();
const state = registry.register({
url: "https://example.com/a\nb",
title: "hello\nworld",
vomText: "@vom 1",
});
const trace = buildTraceV3({
registry,
drafts: [{ op: "scroll", preStateId: state.id, postStateId: state.id }],
startedAt: "2026-08-12T00:00:00.000Z",
stoppedBy: "user_finish",
bskVersion: "test",
});
expect(trace.states[0]?.body).toContain('url: "https://example.com/a\\nb"');
expect(trace.states[0]?.body).toContain('title: "hello\\nworld"');
});
});
describe("recording state ownership", () => {
it("deduplicates within one recording and isolates ids between recordings", () => {
const first = new RecordingStateRegistry();
const second = new RecordingStateRegistry();
expect(first.register({ url: URL, vomText: "same" }).id).toBe("s1");
expect(first.register({ url: URL, vomText: "same" }).id).toBe("s1");
expect(second.register({ url: URL, vomText: "other" }).id).toBe("s1");
});
it("enriches metadata when a deduplicated observation becomes more complete", () => {
const registry = new RecordingStateRegistry();
registry.register({ url: URL, vomText: "same" });
const state = registry.register({
url: URL,
title: "Login",
vomText: "same",
truncated: true,
});
expect(state).toMatchObject({ id: "s1", title: "Login", truncated: true });
});
});
describe("draft binding", () => {
it("keeps capture semantics when no observation exists", () => {
const session = new RecordingObservationSession();
const draft: RecordingDraftStep = {
op: "hover",
captureTarget: { tag: "button", role: "button", name: "新建" },
};
session.bindDraft(draft, 1);
expect(draft.matchedTarget).toEqual({ role: "button", name: "新建", unmatched: true });
expect(draft.preStateId).toBeUndefined();
});
it("does not bind an unmatched new action to a stale observation", () => {
const session = sessionWithInput();
const draft: RecordingDraftStep = {
op: "click",
captureTarget: { tag: "button", role: "button", name: "Confirm" },
targetHint: {
geometry: { rect: { x: 400, y: 300, w: 80, h: 30 }, tag: "button" },
},
};
session.bindDraft(draft, 2, true);
expect(draft.preStateId).toBeUndefined();
expect("matchedTarget" in draft ? draft.matchedTarget : undefined).toEqual({
role: "button",
name: "Confirm",
unmatched: true,
});
});
});
@@ -0,0 +1,289 @@
import { afterEach, describe, expect, it, vi } from "vitest";
import type { CdpRunner, ChromeTabsApi } from "@/tools/shared";
const captureRecordingObservation = vi.hoisted(() => vi.fn());
const waitForDocumentSettled = vi.hoisted(() => vi.fn(async () => "quiet" as const));
vi.mock("../recording/observation-capture", async (importOriginal) => {
const actual = await importOriginal<typeof import("../recording/observation-capture")>();
return { ...actual, captureRecordingObservation };
});
vi.mock("../recording/document-settle", async (importOriginal) => {
const actual = await importOriginal<typeof import("../recording/document-settle")>();
return { ...actual, waitForDocumentSettled };
});
import { ObservationNodeIndex } from "../recording/observation-capture";
import { RecordingObservationRuntime } from "../recording/recording-runtime";
import type { RecordingDraftStep } from "../recording/types";
function observation(url = "https://example.com/") {
return {
rootFrameId: "root",
index: new ObservationNodeIndex({ rootFrameId: "root", matchNodes: [], refs: [] }),
url,
title: "Example",
vomText: '@vom 1\nRootWebArea "Example"',
truncated: false,
};
}
function runtime(): RecordingObservationRuntime {
return new RecordingObservationRuntime({
cdp: { send: vi.fn() as unknown as CdpRunner["send"] },
tabsApi: { get: vi.fn(), query: vi.fn() } as unknown as ChromeTabsApi,
});
}
describe("RecordingObservationRuntime", () => {
afterEach(() => {
captureRecordingObservation.mockReset();
waitForDocumentSettled.mockClear();
});
it("shares one initial capture and includes it in flush", async () => {
let release!: (value: ReturnType<typeof observation>) => void;
captureRecordingObservation.mockReturnValueOnce(
new Promise((resolve) => {
release = resolve;
}),
);
const recording = runtime();
const first = recording.captureInitial(7);
const second = recording.captureInitial(7);
const flushed = recording.flush();
let flushCompleted = false;
void flushed.then(() => {
flushCompleted = true;
});
await Promise.resolve();
expect(captureRecordingObservation).toHaveBeenCalledTimes(1);
expect(flushCompleted).toBe(false);
release(observation());
await Promise.all([first, second, flushed]);
expect(flushCompleted).toBe(true);
});
it("cancels an in-flight initial capture", async () => {
captureRecordingObservation.mockImplementationOnce(
(input: { signal?: AbortSignal }) =>
new Promise((_, reject) => {
input.signal?.addEventListener(
"abort",
() => reject(new DOMException("observation aborted", "AbortError")),
{ once: true },
);
}),
);
const recording = runtime();
const pending = recording.captureInitial(7);
recording.cancel();
await expect(pending).rejects.toMatchObject({ name: "AbortError" });
});
it("retries a failed initial capture when the recording settles at stop", async () => {
captureRecordingObservation
.mockRejectedValueOnce(new Error("document swapped"))
.mockResolvedValueOnce(observation());
const recording = runtime();
await expect(recording.captureInitial(7)).rejects.toThrow("document swapped");
await recording.settleTrailing(7, []);
const trace = recording.buildTrace({
drafts: [],
startedAt: "2026-08-19T00:00:00.000Z",
stoppedBy: "user_finish",
bskVersion: "0.1.6",
});
expect(captureRecordingObservation).toHaveBeenCalledTimes(2);
expect(trace.states).toHaveLength(1);
});
it("keeps runtime matching data out of the serialized trace", async () => {
const captured = observation();
captured.index = new ObservationNodeIndex({
rootFrameId: "root",
matchNodes: [
{
frameId: "root",
backendNodeId: 42,
tag: "button",
rect: { x: 10, y: 20, w: 100, h: 30 },
localRect: { x: 10, y: 20, w: 100, h: 30 },
},
],
refs: [{ ref: "e1", backendNodeId: 42, role: "button", name: "Submit", line: 1 }],
});
captureRecordingObservation.mockResolvedValueOnce(captured);
const recording = runtime();
await recording.captureInitial(7);
const dumped = JSON.stringify(
recording.buildTrace({
drafts: [],
startedAt: "2026-08-19T00:00:00.000Z",
stoppedBy: "user_finish",
bskVersion: "0.1.6",
}),
);
expect(dumped).toContain("RootWebArea");
expect(dumped).not.toContain("matchNodes");
expect(dumped).not.toContain("localRect");
expect(dumped).not.toContain("backendNodeId");
expect(dumped).not.toContain("ObservationNodeIndex");
});
it("captures a tab transition without sharing observation cursors", async () => {
captureRecordingObservation
.mockResolvedValueOnce(observation("https://example.com/first"))
.mockResolvedValueOnce(observation("https://example.com/second"));
const recording = runtime();
await recording.captureInitial(4);
const transition = await recording.captureTabTransition(4, 5);
await recording.captureInitial(4);
await recording.captureInitial(5);
expect(transition).toEqual({
preStateId: "s1",
postStateId: "s2",
targetUrl: "https://example.com/second",
});
expect(captureRecordingObservation.mock.calls.map(([input]) => input.tabId)).toEqual([4, 5]);
});
it("binds an iframe draft to its marked CDP Document scope", async () => {
captureRecordingObservation.mockResolvedValueOnce({
...observation(),
index: new ObservationNodeIndex({
rootFrameId: "root",
frames: [
{
frameId: "child",
target: { tabId: 7, sessionId: "oopif-session" },
recordingDocumentId: "producer-1",
},
],
matchNodes: [
{
backendNodeId: 42,
frameId: "child",
tag: "button",
rect: { x: 410, y: 20, w: 100, h: 30 },
localRect: { x: 10, y: 20, w: 100, h: 30 },
},
],
refs: [{ ref: "e1", backendNodeId: 42, frameId: "child", line: 1 }],
}),
});
const recording = runtime();
const drafts: RecordingDraftStep[] = [
{
op: "click" as const,
captureTarget: { tag: "button", role: "button", name: "Save" },
targetHint: {
geometry: { rect: { x: 10, y: 20, w: 100, h: 30 }, tag: "button" },
},
},
];
await recording.captureInitial(7);
await recording.processDraft(7, drafts, 0, "producer-1");
recording.cancel();
const draft = drafts[0];
expect(draft?.op).toBe("click");
if (!draft || draft.op !== "click") throw new Error("expected click draft");
expect(draft.targetHint).toMatchObject({
frameId: "child",
geometrySpace: "local",
});
expect(draft.matchedTarget?.ref).toBe("e1");
});
it("refreshes a late iframe before matching its first action", async () => {
const iframeObservation = {
...observation(),
index: new ObservationNodeIndex({
rootFrameId: "root",
frames: [
{
frameId: "child",
target: { tabId: 7, sessionId: "oopif-session" },
recordingDocumentId: "producer-1",
},
],
matchNodes: [
{
backendNodeId: 42,
frameId: "child",
tag: "button",
rect: { x: 410, y: 20, w: 100, h: 30 },
localRect: { x: 10, y: 20, w: 100, h: 30 },
},
],
refs: [
{
ref: "e1",
backendNodeId: 42,
frameId: "child",
role: "button",
name: "表格视图",
line: 1,
},
],
}),
};
captureRecordingObservation
.mockResolvedValueOnce(observation())
.mockResolvedValueOnce(iframeObservation)
.mockResolvedValueOnce(iframeObservation);
const recording = runtime();
await recording.captureInitial(7);
await recording.refreshDocument(7, "producer-1");
const drafts: RecordingDraftStep[] = [
{
op: "click",
captureTarget: { tag: "button", role: "button", name: "表格视图" },
targetHint: {
geometry: { rect: { x: 10, y: 20, w: 100, h: 30 }, tag: "button" },
},
},
];
await recording.processDraft(7, drafts, 0, "producer-1");
recording.cancel();
expect(captureRecordingObservation).toHaveBeenCalledTimes(3);
expect(waitForDocumentSettled).toHaveBeenCalledWith(
expect.anything(),
{ frameId: "child", target: { tabId: 7, sessionId: "oopif-session" } },
{ signal: expect.any(AbortSignal) },
);
const draft = drafts[0];
expect(draft?.op).toBe("click");
if (!draft || draft.op !== "click") throw new Error("expected click draft");
expect(draft.matchedTarget).toMatchObject({ ref: "e1", name: "表格视图" });
});
it("does not poison final flush when a frame readiness refresh fails", async () => {
captureRecordingObservation
.mockResolvedValueOnce(observation())
.mockRejectedValueOnce(new Error("child Document replaced"));
const recording = runtime();
await recording.captureInitial(7);
await expect(recording.refreshDocument(7, "producer-1")).rejects.toThrow(
"child Document replaced",
);
await expect(recording.flush()).resolves.toBeUndefined();
});
});
@@ -1,80 +0,0 @@
import { describe, expect, it } from "vitest";
import { appendRecordedPayload, observeRecordedNavigation } from "../recording-step-buffer";
describe("recording-step-buffer", () => {
it("stores semantic click without summary", () => {
const buffer = { steps: [], pendingNavigation: false };
appendRecordedPayload(buffer, {
op: "click",
target: { tag: "button", role: "button", name: "发布" },
expects_navigation: true,
});
expect(buffer.steps).toEqual([
{
op: "click",
target: { tag: "button", role: "button", name: "发布" },
},
]);
expect(buffer.pendingNavigation).toBe(true);
});
it("annotates navigated_to on action-caused navigation instead of wait_for_navigation", () => {
const buffer = {
steps: [
{
op: "click" as const,
target: { tag: "button", role: "button", name: "发布" },
},
],
currentUrl: "https://example.com/a",
pendingNavigation: true,
pendingNavigationDeadline: Date.now() + 5_000,
};
observeRecordedNavigation(buffer, "https://example.com/b", true);
expect(buffer.steps).toEqual([
{
op: "click",
target: { tag: "button", role: "button", name: "发布" },
navigated_to: "https://example.com/b",
},
]);
expect(JSON.stringify(buffer.steps)).not.toContain("wait_for_navigation");
});
it("annotates navigated_to onto a select that auto-submits a navigation", () => {
const buffer = {
steps: [
{
op: "select" as const,
target: { tag: "select", role: "combobox", name: "分类" },
values: ["tech"],
},
],
currentUrl: "https://example.com/list",
pendingNavigation: true,
pendingNavigationDeadline: Date.now() + 5_000,
};
observeRecordedNavigation(buffer, "https://example.com/list?cat=tech", true);
expect(buffer.steps).toEqual([
{
op: "select",
target: { tag: "select", role: "combobox", name: "分类" },
values: ["tech"],
navigated_to: "https://example.com/list?cat=tech",
},
]);
});
it("emits navigate for uncaused URL changes", () => {
const buffer = {
steps: [],
currentUrl: "https://example.com/a",
pendingNavigation: false,
};
observeRecordedNavigation(buffer, "https://example.com/b", false);
expect(buffer.steps[0]).toMatchObject({
op: "navigate",
url: "https://example.com/b",
});
});
});
@@ -0,0 +1,136 @@
import { afterEach, describe, expect, it, vi } from "vitest";
import type { CdpRunner, ChromeTabsApi } from "@/tools/shared";
import { ObservationNodeIndex, type RegisteredObservation } from "../recording/observation-capture";
import { RecordingObservationSession } from "../recording/observation-session";
import { SettleController } from "../recording/settle-controller";
import type { RecordingDraftStep } from "../recording/types";
const OBSERVATION: RegisteredObservation = {
stateId: "s-next",
rootFrameId: "root",
index: new ObservationNodeIndex({ rootFrameId: "root", matchNodes: [], refs: [] }),
url: "https://example.com/next",
};
describe("SettleController", () => {
afterEach(() => vi.useRealTimers());
it("aborts superseded capture work and lands the prior action on the next origin", async () => {
vi.useFakeTimers();
const cdp: CdpRunner = {
send: vi.fn(async () => ({
result: { value: { idleMs: 1_000, readyState: "complete" } },
})) as unknown as CdpRunner["send"],
};
const tabsApi: ChromeTabsApi = {
get: vi.fn(async () => ({ id: 7, url: OBSERVATION.url }) as chrome.tabs.Tab),
query: vi.fn(async () => []),
};
const session = new RecordingObservationSession();
let firstSignal: AbortSignal | undefined;
vi.spyOn(session, "capture")
.mockImplementationOnce(async (_cdp, _tabs, _tabId, signal) => {
firstSignal = signal;
return new Promise<RegisteredObservation>((_resolve, reject) => {
signal?.addEventListener(
"abort",
() => reject(new DOMException("observation aborted", "AbortError")),
{ once: true },
);
});
})
.mockResolvedValueOnce(OBSERVATION);
const drafts: RecordingDraftStep[] = [
{ op: "click", captureTarget: { tag: "button", name: "First" }, preStateId: "s1" },
];
const controller = new SettleController({ session, cdp, tabsApi, tabId: 7 });
controller.schedule(drafts, 0);
await vi.advanceTimersByTimeAsync(180);
drafts.push({
op: "click",
captureTarget: { tag: "button", name: "Second" },
preStateId: "s-next",
});
controller.schedule(drafts, 1);
expect(firstSignal?.aborted).toBe(true);
expect(drafts[0]?.postStateId).toBe("s-next");
await vi.advanceTimersByTimeAsync(500);
await controller.flush();
expect(drafts[1]?.postStateId).toBe("s-next");
});
it("does not consume a newer redirect while reading the prior landing URL", async () => {
vi.useFakeTimers();
const cdp: CdpRunner = {
send: vi.fn(async () => ({
result: { value: { idleMs: 1_000, readyState: "complete" } },
})) as unknown as CdpRunner["send"],
};
let resolveFirstTab!: (tab: chrome.tabs.Tab) => void;
const firstTab = new Promise<chrome.tabs.Tab>((resolve) => {
resolveFirstTab = resolve;
});
const tabsApi: ChromeTabsApi = {
get: vi
.fn()
.mockImplementationOnce(() => firstTab)
.mockResolvedValue({ id: 7, url: OBSERVATION.url }),
query: vi.fn(async () => []),
};
const session = new RecordingObservationSession();
session.cursor.lastSettled = OBSERVATION;
const drafts: RecordingDraftStep[] = [];
const controller = new SettleController({ session, cdp, tabsApi, tabId: 7 });
controller.scheduleRedirect(drafts, "https://example.com/intermediate");
await vi.advanceTimersByTimeAsync(180);
expect(tabsApi.get).toHaveBeenCalledTimes(1);
controller.scheduleRedirect(drafts, OBSERVATION.url);
resolveFirstTab({ id: 7, url: "https://example.com/intermediate" } as chrome.tabs.Tab);
await vi.advanceTimersByTimeAsync(500);
await controller.flushRedirects();
expect(drafts).toEqual([]);
expect(controller.hasPending).toBe(false);
});
it("settles an action in its own OOPIF document scope", async () => {
vi.useFakeTimers();
const send = vi.fn();
const sendToTarget = vi.fn(async (_target, method: string) => {
if (method === "Page.createIsolatedWorld") return { executionContextId: 39 };
return { result: { value: { idleMs: 1_000, readyState: "complete" } } };
});
const cdp: CdpRunner = {
send: send as unknown as CdpRunner["send"],
sendToTarget: sendToTarget as unknown as NonNullable<CdpRunner["sendToTarget"]>,
};
const tabsApi: ChromeTabsApi = {
get: vi.fn(async () => ({ id: 7, url: OBSERVATION.url }) as chrome.tabs.Tab),
query: vi.fn(async () => []),
};
const session = new RecordingObservationSession();
vi.spyOn(session, "capture").mockResolvedValue(OBSERVATION);
const drafts: RecordingDraftStep[] = [
{ op: "click", captureTarget: { tag: "button", name: "Inside frame" } },
];
const controller = new SettleController({ session, cdp, tabsApi, tabId: 7 });
const target = { tabId: 7, sessionId: "oopif-session" };
controller.schedule(drafts, 0, { target, frameId: "oopif-frame" });
await vi.advanceTimersByTimeAsync(180);
await controller.flush();
expect(send).not.toHaveBeenCalled();
expect(sendToTarget).toHaveBeenCalledWith(
target,
"Runtime.evaluate",
expect.objectContaining({ contextId: 39 }),
);
expect(drafts[0]?.postStateId).toBe("s-next");
});
});
@@ -0,0 +1,164 @@
import { describe, expect, it } from "vitest";
import { appendRecordedPayload, observeRecordedNavigation } from "../recording/step-buffer";
describe("recording-step-buffer", () => {
it("stores semantic click without summary", () => {
const buffer = { steps: [], navigation: { pendingNavigation: false } };
appendRecordedPayload(buffer, {
op: "click",
target: { tag: "button", role: "button", name: "发布" },
expects_navigation: true,
});
expect(buffer.steps).toEqual([
{
op: "click",
captureTarget: { tag: "button", role: "button", name: "发布" },
},
]);
expect(buffer.navigation.pendingNavigation).toBe(true);
});
it("keeps the hovered element description as a capture fallback", () => {
const buffer = { steps: [], navigation: { pendingNavigation: false } };
appendRecordedPayload(
buffer,
{
op: "hover",
target: { tag: "button", role: "button", name: "新建" },
geometry: {
rect: { x: 900, y: 8, w: 60, h: 32 },
tag: "button",
},
},
{
geometry: {
rect: { x: 900, y: 8, w: 60, h: 32 },
tag: "button",
},
},
);
expect(buffer.steps).toEqual([
{
op: "hover",
captureTarget: { tag: "button", role: "button", name: "新建" },
targetHint: {
geometry: {
rect: { x: 900, y: 8, w: 60, h: 32 },
tag: "button",
},
},
},
]);
});
it("annotates navigated_to on action-caused navigation instead of wait_for_navigation", () => {
const buffer = {
steps: [
{
op: "click" as const,
captureTarget: { tag: "button", role: "button", name: "发布" },
},
],
navigation: {
currentUrl: "https://example.com/a",
pendingNavigation: true,
pendingNavigationDeadline: Date.now() + 5_000,
},
};
observeRecordedNavigation(buffer, "https://example.com/b", true);
expect(buffer.steps).toEqual([
{
op: "click",
captureTarget: { tag: "button", role: "button", name: "发布" },
navigatedTo: "https://example.com/b",
},
]);
expect(JSON.stringify(buffer.steps)).not.toContain("wait_for_navigation");
});
it("annotates navigated_to onto a select that auto-submits a navigation", () => {
const buffer = {
steps: [
{
op: "select" as const,
captureTarget: { tag: "select", role: "combobox", name: "分类" },
values: ["tech"],
},
],
navigation: {
currentUrl: "https://example.com/list",
pendingNavigation: true,
pendingNavigationDeadline: Date.now() + 5_000,
},
};
observeRecordedNavigation(buffer, "https://example.com/list?cat=tech", true);
expect(buffer.steps).toEqual([
{
op: "select",
captureTarget: { tag: "select", role: "combobox", name: "分类" },
values: ["tech"],
navigatedTo: "https://example.com/list?cat=tech",
},
]);
});
it("emits navigate for uncaused URL changes", () => {
const buffer = {
steps: [],
navigation: { currentUrl: "https://example.com/a", pendingNavigation: false },
};
const result = observeRecordedNavigation(buffer, "https://example.com/b", false);
expect(result).toEqual({ kind: "appended", index: 0 });
expect(buffer.steps[0]).toMatchObject({
op: "navigate",
url: "https://example.com/b",
});
});
it("asks the recorder to coalesce redirect hops instead of emitting each one", () => {
const buffer = {
steps: [],
navigation: {
currentUrl: "https://passport.example/login",
pendingNavigation: false,
},
};
const hop1 = observeRecordedNavigation(
buffer,
"https://passport.example/callback",
false,
"link",
["server_redirect"],
);
expect(hop1).toEqual({
kind: "coalesce_redirect",
url: "https://passport.example/callback",
});
expect(buffer.steps).toEqual([]);
const hop2 = observeRecordedNavigation(buffer, "https://app.example/dashboard", false, "link", [
"client_redirect",
]);
expect(hop2).toEqual({
kind: "coalesce_redirect",
url: "https://app.example/dashboard",
});
expect(buffer.steps).toEqual([]);
expect(buffer.navigation.currentUrl).toBe("https://app.example/dashboard");
});
it("does not reuse redirect metadata for a later content-observed URL change", () => {
const buffer = {
steps: [],
navigation: {
currentUrl: "https://example.com/redirect",
pendingNavigation: false,
},
};
const result = observeRecordedNavigation(buffer, "https://example.com/spa");
expect(result).toEqual({ kind: "appended", index: 0 });
expect(buffer.steps[0]).toMatchObject({ op: "navigate", url: "https://example.com/spa" });
});
});
@@ -0,0 +1,39 @@
import { describe, expect, it } from "vitest";
import { RecordingTabCoordinator } from "../recording/tab-coordinator";
describe("RecordingTabCoordinator", () => {
it("keeps navigation state isolated per tab", () => {
const tabs = new RecordingTabCoordinator(4, "https://example.com/first");
const first = tabs.navigation(4);
first.pendingNavigation = true;
const second = tabs.navigation(5, "https://example.com/second");
expect(second).toEqual({
currentUrl: "https://example.com/second",
pendingNavigation: false,
});
expect(tabs.navigation(4)).toBe(first);
expect(tabs.navigation(4).pendingNavigation).toBe(true);
});
it("invalidates an earlier activation when a newer tab becomes active", () => {
const tabs = new RecordingTabCoordinator(4);
const second = tabs.noteActivation(5);
const third = tabs.noteActivation(6);
expect(tabs.isLatest(second)).toBe(false);
expect(tabs.isLatest(third)).toBe(true);
tabs.commit(6);
expect(tabs.currentTabId).toBe(6);
});
it("commits the freshly observed URL when returning to an existing tab", () => {
const tabs = new RecordingTabCoordinator(4);
tabs.navigation(5, "https://example.com/old");
tabs.commit(5, "https://example.com/new");
expect(tabs.currentTabId).toBe(5);
expect(tabs.navigation(5).currentUrl).toBe("https://example.com/new");
});
});
@@ -0,0 +1,159 @@
import type { RenderedRef } from "@browser-skill/vom";
import { describe, expect, it } from "vitest";
import type { CaptureVomMatchNode } from "@/tools/capture-vom-observation";
import { ObservationNodeIndex, type RegisteredObservation } from "../recording/observation-capture";
import { matchObservationTarget } from "../recording/target-matcher";
function node(backendNodeId: number, frameId: string, x = 10): CaptureVomMatchNode {
return {
backendNodeId,
frameId,
tag: "button",
rect: { x, y: 20, w: 100, h: 30 },
localRect: { x, y: 20, w: 100, h: 30 },
};
}
function observation(
matchNodes: CaptureVomMatchNode[],
refs: RenderedRef[],
): RegisteredObservation {
return {
stateId: "s1",
rootFrameId: "root",
index: new ObservationNodeIndex({ rootFrameId: "root", matchNodes, refs }),
url: "https://example.com",
};
}
describe("matchObservationTarget", () => {
it("indexes the safe geometry DTO without discarding frame-local geometry", () => {
const geometry = node(42, "child");
geometry.localRect = { x: 5, y: 6, w: 100, h: 30 };
const ref: RenderedRef = {
ref: "e1",
backendNodeId: 42,
frameId: "child",
line: 1,
};
const candidate = observation([geometry], [ref]).index.candidates("child", "button")[0];
expect(candidate).toEqual({ frameId: "child", geometry, ref });
});
it("matches a unique node using canonical top-level viewport geometry", () => {
const target = matchObservationTarget({
observation: observation(
[node(42, "root")],
[{ ref: "e1", backendNodeId: 42, role: "button", name: "发布", line: 1 }],
),
hint: { geometry: { rect: { x: 10, y: 20, w: 100, h: 30 }, tag: "button" } },
});
expect(target).toEqual({ ref: "e1", role: "button", name: "发布" });
});
it("uses frame id with backend node id so sibling frames cannot collide", () => {
const target = matchObservationTarget({
observation: observation(
[node(42, "left"), node(42, "right")],
[
{ ref: "e1", backendNodeId: 42, frameId: "left", line: 1 },
{ ref: "e2", backendNodeId: 42, frameId: "right", line: 2 },
],
),
hint: {
frameId: "right",
geometry: { rect: { x: 10, y: 20, w: 100, h: 30 }, tag: "button" },
},
});
expect(target.ref).toBe("e2");
});
it("matches iframe capture geometry in the child viewport coordinate space", () => {
const child = node(42, "child", 410);
child.localRect = { x: 10, y: 20, w: 100, h: 30 };
const target = matchObservationTarget({
observation: observation(
[child],
[{ ref: "e1", backendNodeId: 42, frameId: "child", line: 1 }],
),
hint: {
frameId: "child",
geometrySpace: "local",
geometry: { rect: { x: 10, y: 20, w: 100, h: 30 }, tag: "button" },
},
});
expect(target.ref).toBe("e1");
});
it("restricts a missing frame hint to the root frame", () => {
const target = matchObservationTarget({
observation: observation(
[node(42, "child")],
[{ ref: "e1", backendNodeId: 42, frameId: "child", line: 1 }],
),
hint: { geometry: { rect: { x: 10, y: 20, w: 100, h: 30 }, tag: "button" } },
fallback: { tag: "button", name: "发布" },
});
expect(target).toEqual({ name: "发布", unmatched: true });
});
it("fails closed when the source Document has no VOM frame mapping", () => {
const target = matchObservationTarget({
observation: observation(
[node(42, "root")],
[{ ref: "e1", backendNodeId: 42, role: "button", name: "发布", line: 1 }],
),
hint: {
frameId: null,
geometry: { rect: { x: 10, y: 20, w: 100, h: 30 }, tag: "button" },
},
fallback: { tag: "button", role: "button", name: "发布" },
});
expect(target).toEqual({ role: "button", name: "发布", unmatched: true });
});
it("returns unmatched for ambiguous geometry", () => {
const target = matchObservationTarget({
observation: observation(
[node(42, "root"), node(43, "root")],
[
{ ref: "e1", backendNodeId: 42, line: 1 },
{ ref: "e2", backendNodeId: 43, line: 2 },
],
),
hint: { geometry: { rect: { x: 10, y: 20, w: 100, h: 30 }, tag: "button" } },
});
expect(target.unmatched).toBe(true);
});
it("uses semantics when geometry is unavailable without crossing frame boundaries", () => {
const target = matchObservationTarget({
observation: observation(
[node(41, "root"), node(42, "child")],
[
{ ref: "e1", backendNodeId: 41, role: "button", name: "保存", line: 1 },
{ ref: "e2", backendNodeId: 42, frameId: "child", role: "button", name: "保存", line: 2 },
],
),
hint: { frameId: "child" },
fallback: { tag: "button", role: "button", name: "保存" },
});
expect(target.ref).toBe("e2");
});
it("uses semantics to disambiguate equal geometry in one frame", () => {
const target = matchObservationTarget({
observation: observation(
[node(42, "root"), node(43, "root")],
[
{ ref: "e1", backendNodeId: 42, role: "button", name: "保存", line: 1 },
{ ref: "e2", backendNodeId: 43, role: "button", name: "取消", line: 2 },
],
),
hint: { geometry: { rect: { x: 10, y: 20, w: 100, h: 30 }, tag: "button" } },
fallback: { tag: "button", role: "button", name: "取消" },
});
expect(target.ref).toBe("e2");
});
});
@@ -1,6 +1,15 @@
import { describe, expect, it } from "vitest";
import type { DraftTraceStep } from "@/transport/types";
import { reduceTraceSteps, resolveTraceStartUrl, shouldRecordPress } from "../trace-reducer";
import type { RecordingDraftStep } from "@/lib/recording/types";
import { shouldRecordPress } from "../recording/draft-policy";
import { buildTraceV2 } from "../recording/trace-reducer-v2";
function reduceTraceSteps(steps: RecordingDraftStep[], startUrl?: string) {
return buildTraceV2({
steps,
startedAt: "2026-01-01T00:00:00.000Z",
...(startUrl ? { startUrl } : {}),
});
}
describe("shouldRecordPress", () => {
it("keeps Enter and Escape", () => {
@@ -18,36 +27,47 @@ describe("shouldRecordPress", () => {
});
describe("reduceTraceSteps", () => {
it("does not expose internal tab transitions in trace v2", () => {
const trace = reduceTraceSteps([
{
op: "switch_tab",
preStateId: "s1",
postStateId: "s2",
},
]);
expect(trace.steps).toEqual([]);
});
it("builds steps with pages dictionary and page id references", () => {
const drafts: DraftTraceStep[] = [
const drafts: RecordingDraftStep[] = [
{
op: "navigate",
url: "https://example.com/search?q=hello&utm_source=x",
page_url: "https://example.com/search?q=hello&utm_source=x",
pageUrl: "https://example.com/search?q=hello&utm_source=x",
},
{
op: "fill",
target: { tag: "input", role: "textbox", name: "搜索", name_attr: "q" },
captureTarget: { tag: "input", role: "textbox", name: "搜索", name_attr: "q" },
value: "browser skill",
page_url: "https://example.com/search?q=hello&utm_source=x",
pageUrl: "https://example.com/search?q=hello&utm_source=x",
},
{
op: "press",
key: "Enter",
target: { tag: "input", role: "textbox", name: "搜索", name_attr: "q" },
navigated_to: "https://example.com/results/42",
page_url: "https://example.com/search?q=hello&utm_source=x",
captureTarget: { tag: "input", role: "textbox", name: "搜索", name_attr: "q" },
navigatedTo: "https://example.com/results/42",
pageUrl: "https://example.com/search?q=hello&utm_source=x",
},
{
op: "click",
target: { tag: "button", role: "button", name: "发布" },
navigated_to: "https://example.com/p/99",
page_url: "https://example.com/results/42",
captureTarget: { tag: "button", role: "button", name: "发布" },
navigatedTo: "https://example.com/p/99",
pageUrl: "https://example.com/results/42",
},
{
op: "press",
key: "a",
page_url: "https://example.com/p/99",
pageUrl: "https://example.com/p/99",
},
];
@@ -109,9 +129,9 @@ describe("reduceTraceSteps", () => {
[
{
op: "fill",
target: { tag: "input", role: "textbox", name: "Search query" },
captureTarget: { tag: "input", role: "textbox", name: "Search query" },
value: "",
page_url: "https://example.com/search",
pageUrl: "https://example.com/search",
},
],
"https://example.com/search",
@@ -125,16 +145,16 @@ describe("reduceTraceSteps", () => {
]);
});
it("maps select navigated_to onto effect.navigated_to (page id)", () => {
it("maps select navigatedTo onto effect.navigatedTo (page id)", () => {
const { pages, steps } = reduceTraceSteps(
[
{
op: "select",
target: { tag: "select", role: "combobox", name: "分类" },
captureTarget: { tag: "select", role: "combobox", name: "分类" },
values: ["tech"],
labels: ["技术"],
navigated_to: "https://example.com/list?cat=tech",
page_url: "https://example.com/list",
navigatedTo: "https://example.com/list?cat=tech",
pageUrl: "https://example.com/list",
},
],
"https://example.com/list",
@@ -151,36 +171,32 @@ describe("reduceTraceSteps", () => {
});
});
it("keeps hover steps before menu clicks", () => {
it("drops hover steps, which only exist in trace v3", () => {
const { steps } = reduceTraceSteps(
[
{
op: "hover",
target: { tag: "span", role: "button", name: "Account" },
page_url: "https://example.com/app",
captureTarget: { tag: "span", role: "button", name: "Account" },
pageUrl: "https://example.com/app",
},
{
op: "click",
target: { tag: "a", role: "link", name: "Profile" },
page_url: "https://example.com/app",
captureTarget: { tag: "a", role: "link", name: "Profile" },
pageUrl: "https://example.com/app",
},
],
"https://example.com/app",
);
expect(steps.map((s) => s.op)).toEqual(["hover", "click"]);
expect(steps[0]).toMatchObject({
op: "hover",
page: "p1",
target: { name: "Account" },
});
expect(steps.map((s) => s.op)).toEqual(["click"]);
});
it("resolveTraceStartUrl prefers explicit start URL", () => {
expect(
resolveTraceStartUrl(
[{ op: "navigate", url: "https://example.com/other" }],
"https://example.com/start",
),
buildTraceV2({
steps: [{ op: "navigate", url: "https://example.com/other" }],
startedAt: "2026-01-01T00:00:00.000Z",
startUrl: "https://example.com/start",
}).entry.start_url,
).toBe("https://example.com/start");
});
});
@@ -0,0 +1,132 @@
import { describe, expect, it } from "vitest";
import { TRACE_VERSION_V3, VOM_FORMAT_VERSION } from "@/transport/types";
import { RecordingStateRegistry } from "../recording/state-registry";
import { buildTraceV3 } from "../recording/trace-builder-v3";
import { reduceTraceStepsV3 } from "../recording/trace-reducer-v3";
import type { RecordingDraftStep } from "../recording/types";
describe("trace reducer v3", () => {
it("removes form literals when value redaction is requested", () => {
const reduced = reduceTraceStepsV3(
[
{
op: "fill",
value: "private@example.com",
preStateId: "s1",
postStateId: "s1",
},
{
op: "select",
values: ["private-account-id"],
labels: ["Private account"],
preStateId: "s1",
postStateId: "s1",
},
],
{ redactValues: true },
);
expect(reduced.steps[0]).toMatchObject({ op: "fill", value: "***", redacted: true });
expect(reduced.steps[1]).toMatchObject({ op: "select" });
expect(reduced.steps[1]).not.toHaveProperty("selection");
expect(JSON.stringify(reduced.steps)).not.toContain("private@example.com");
expect(JSON.stringify(reduced.steps)).not.toContain("private-account-id");
});
it("emits tab transitions only for callers that advertised support", () => {
const drafts: RecordingDraftStep[] = [
{
op: "switch_tab",
preStateId: "s1",
postStateId: "s2",
},
];
expect(reduceTraceStepsV3(drafts).steps).toEqual([]);
expect(reduceTraceStepsV3(drafts, { includeTabSwitches: true }).steps).toEqual([
{ op: "switch_tab", id: 1, state: "s1", result: { state: "s2" } },
]);
});
it("keeps hover steps before menu clicks", () => {
const drafts: RecordingDraftStep[] = [
{
op: "hover",
captureTarget: { tag: "span", role: "button", name: "Account" },
preStateId: "s1",
postStateId: "s1",
},
{
op: "click",
captureTarget: { tag: "a", role: "link", name: "Profile" },
preStateId: "s1",
postStateId: "s1",
},
];
expect(reduceTraceStepsV3(drafts).steps.map((s) => s.op)).toEqual(["hover", "click"]);
});
it("does not export a tab transition without both observation endpoints", () => {
const sourceOnly: RecordingDraftStep = { op: "switch_tab", preStateId: "s1" };
const targetOnly: RecordingDraftStep = { op: "switch_tab", postStateId: "s2" };
expect(
reduceTraceStepsV3([sourceOnly, targetOnly], { includeTabSwitches: true }).steps,
).toEqual([]);
});
it("collapses redirect hops while retaining draft-to-step identity", () => {
const drafts: RecordingDraftStep[] = [
{ op: "navigate", url: "https://example.com/start", preStateId: "s1", postStateId: "s2" },
{
op: "navigate",
url: "https://example.com/final",
transitionQualifiers: ["server_redirect"],
preStateId: "s2",
postStateId: "s3",
},
];
const before = structuredClone(drafts);
const reduced = reduceTraceStepsV3(drafts);
expect(drafts).toEqual(before);
expect(reduced.steps).toEqual([
expect.objectContaining({
id: 1,
op: "navigate",
state: "s1",
to: "https://example.com/final",
result: { state: "s3" },
}),
]);
expect(reduced.stepIdByDraftId.get(1)).toBe(1);
expect(reduced.stepIdByDraftId.get(2)).toBe(1);
});
it("builds the wire model from protocol constants", () => {
const registry = new RecordingStateRegistry();
const state = registry.register({ url: "https://example.com", vomText: "@vom 1" });
const trace = buildTraceV3({
registry,
drafts: [
{
op: "click",
captureTarget: { tag: "button", role: "button", name: "Save" },
preStateId: state.id,
postStateId: state.id,
},
],
startedAt: "2026-08-12T00:00:00.000Z",
stoppedBy: "user_finish",
bskVersion: "test",
});
expect(trace.version).toBe(TRACE_VERSION_V3);
expect(trace.recorder.vom).toBe(VOM_FORMAT_VERSION);
expect(trace.steps[0]).toMatchObject({
op: "click",
target: { role: "button", name: "Save", unmatched: true },
});
});
});
+6 -5
View File
@@ -1,12 +1,13 @@
/**
* Build a semantic TargetDescriptor for an interacted element.
* Build a semantic capture descriptor for an interacted element.
*
* Trace steps are an LLM *textbook*: each click must say what to look for
* on screen (usually a short visible name). Tag-only noise like
* `{ "tag": "div" }` / “点击div” is useless and must not be recorded.
*/
export interface TargetDescriptor {
/** Content-script capture descriptor before VOM geometric matching. */
export interface CaptureTargetDescriptor {
role?: string;
name?: string;
tag: string;
@@ -247,7 +248,7 @@ export function resolveHoverElement(target: Element): Element | null {
* (visible name), or at least a form `name_attr` for checkbox/radio.
* Recording “点击div” with no name fails this bar.
*/
export function isMeaningfulClickTarget(target: TargetDescriptor): boolean {
export function isMeaningfulClickTarget(target: CaptureTargetDescriptor): boolean {
const name = target.name?.trim();
if (name && isActionableLabel(name)) return true;
if (
@@ -323,7 +324,7 @@ function nearbyLabelText(el: Element): string | undefined {
return undefined;
}
export function describeTarget(el: Element): TargetDescriptor {
export function describeTarget(el: Element): CaptureTargetDescriptor {
const tag = el.tagName.toLowerCase();
const role = inferRole(el);
const name = accessibleName(el);
@@ -350,7 +351,7 @@ export function describeTarget(el: Element): TargetDescriptor {
};
}
export function describeEventTarget(target: EventTarget | null): TargetDescriptor | null {
export function describeEventTarget(target: EventTarget | null): CaptureTargetDescriptor | null {
if (!(target instanceof Element)) return null;
const clickable = resolveClickableElement(target);
if (!clickable) return null;
+39 -3
View File
@@ -3,7 +3,7 @@
* service worker and a tab's content script.
*/
import type { TargetDescriptor } from "./describe-target";
import type { CaptureTargetDescriptor } from "./describe-target";
export const RECORD_START = "bsk-record-start";
export const RECORD_STEP = "bsk-record-step";
@@ -16,6 +16,10 @@ export interface RecordStartAck {
ok: true;
}
export type RecordStepAck =
| { ok: true; sequence: number }
| { ok: false; expectedSequence: number; error: string };
export type RecordStopAck = { ok: true } | { ok: false; error: string };
export interface RecordQueryMessage {
@@ -42,7 +46,7 @@ export interface RecordStartMessage {
export interface RecordStepPayload {
op: "click" | "hover" | "fill" | "press" | "select" | "navigate";
target?: TargetDescriptor;
target?: CaptureTargetDescriptor;
value?: string;
key?: string;
modifiers?: Array<"alt" | "ctrl" | "meta" | "shift">;
@@ -50,17 +54,30 @@ export interface RecordStepPayload {
labels?: string[];
url?: string;
redacted?: boolean;
commit?: "enter" | "suggestion" | "blur";
/** Page URL when the step was captured. */
page_url?: string;
/** Event-target bounds in the source frame's viewport coordinate space. */
geometry?: {
rect: { x: number; y: number; w: number; h: number };
tag: string;
};
/** Capture-only hint; never persisted unless converted to navigated_to. */
expects_navigation?: boolean;
/** Whether an observed URL change was synchronously caused by the action. */
navigation_caused_by_action?: boolean;
/** Raw webNavigation transition metadata for cause mapping. */
transitionType?: string;
transitionQualifiers?: string[];
}
export interface RecordStepMessage {
type: typeof RECORD_STEP;
requestId: string;
/** Stable for the lifetime of one content-script document. */
producerId: string;
/** Monotonic within one content-script producer and recording request. */
sequence: number;
step: RecordStepPayload;
}
@@ -112,8 +129,27 @@ export function isRecordQueryMessage(msg: unknown): msg is RecordQueryMessage {
export function isRecordStepMessage(msg: unknown): msg is RecordStepMessage {
if (typeof msg !== "object" || msg === null) return false;
const m = msg as Record<string, unknown>;
if (m.type !== RECORD_STEP || typeof m.requestId !== "string") return false;
if (
m.type !== RECORD_STEP ||
typeof m.requestId !== "string" ||
typeof m.producerId !== "string" ||
m.producerId.length === 0 ||
typeof m.sequence !== "number" ||
!Number.isSafeInteger(m.sequence) ||
m.sequence < 1
) {
return false;
}
const step = m.step;
if (typeof step !== "object" || step === null) return false;
return typeof (step as RecordStepPayload).op === "string";
}
export function isAcceptedRecordStepAck(
value: unknown,
sequence: number,
): value is Extract<RecordStepAck, { ok: true }> {
if (typeof value !== "object" || value === null) return false;
const ack = value as Partial<RecordStepAck>;
return ack.ok === true && ack.sequence === sequence;
}
@@ -1,136 +0,0 @@
import type { DraftTraceStep } from "@/transport/types";
import type { RecordStepPayload } from "./record-bridge";
export interface RecordingStepBuffer {
steps: DraftTraceStep[];
currentUrl?: string;
pendingNavigation: boolean;
pendingNavigationDeadline?: number;
}
const NAVIGATION_TRIGGER_WINDOW_MS = 3_000;
function toDraftStep(payload: RecordStepPayload): DraftTraceStep | null {
const pageUrl = payload.page_url;
switch (payload.op) {
case "click":
return payload.target
? {
op: "click",
target: payload.target,
...(pageUrl ? { page_url: pageUrl } : {}),
}
: null;
case "hover":
return payload.target
? {
op: "hover",
target: payload.target,
...(pageUrl ? { page_url: pageUrl } : {}),
}
: null;
case "fill":
return payload.target
? {
op: "fill",
target: payload.target,
value: payload.value ?? "",
...(payload.redacted ? { redacted: true } : {}),
...(pageUrl ? { page_url: pageUrl } : {}),
}
: null;
case "press":
return payload.key
? {
op: "press",
key: payload.key,
...(payload.target ? { target: payload.target } : {}),
...(payload.modifiers?.length ? { modifiers: payload.modifiers } : {}),
...(pageUrl ? { page_url: pageUrl } : {}),
}
: null;
case "select":
return payload.target && payload.values
? {
op: "select",
target: payload.target,
values: payload.values,
...(payload.labels?.length ? { labels: payload.labels } : {}),
...(pageUrl ? { page_url: pageUrl } : {}),
}
: null;
case "navigate":
return null;
}
}
function annotateLastStepNavigation(buffer: RecordingStepBuffer, url: string): boolean {
for (let i = buffer.steps.length - 1; i >= 0; i -= 1) {
const step = buffer.steps[i];
if (!step) continue;
if (step.op === "click" || step.op === "press" || step.op === "select") {
buffer.steps[i] = { ...step, navigated_to: url };
return true;
}
break;
}
return false;
}
export function observeRecordedNavigation(
buffer: RecordingStepBuffer,
url: string,
causedByAction?: boolean,
): void {
if (!url || url === buffer.currentUrl) return;
buffer.currentUrl = url;
const pendingIsCurrent =
buffer.pendingNavigation &&
(buffer.pendingNavigationDeadline === undefined ||
buffer.pendingNavigationDeadline >= Date.now());
if (causedByAction === true || (causedByAction === undefined && pendingIsCurrent)) {
buffer.pendingNavigation = false;
buffer.pendingNavigationDeadline = undefined;
if (!annotateLastStepNavigation(buffer, url)) {
buffer.steps.push({
op: "navigate",
url,
page_url: url,
});
}
return;
}
buffer.pendingNavigation = false;
buffer.pendingNavigationDeadline = undefined;
buffer.steps.push({
op: "navigate",
url,
page_url: url,
});
}
export function appendRecordedPayload(
buffer: RecordingStepBuffer,
payload: RecordStepPayload,
): void {
if (payload.op === "navigate") {
if (payload.url) {
observeRecordedNavigation(buffer, payload.url, payload.navigation_caused_by_action);
}
return;
}
const step = toDraftStep({
...payload,
page_url: payload.page_url ?? buffer.currentUrl,
});
if (!step) return;
buffer.steps.push(step);
if (step.op === "click" || step.op === "press" || step.op === "select" || step.op === "fill") {
buffer.pendingNavigation = payload.expects_navigation === true;
buffer.pendingNavigationDeadline = buffer.pendingNavigation
? Date.now() + NAVIGATION_TRIGGER_WINDOW_MS
: undefined;
}
}
@@ -0,0 +1,39 @@
import {
RECORD_DOCUMENT_ATTRIBUTE,
recordingDocumentMarkerValue,
} from "@/shared/recording-document-identity";
export interface RecordingDocumentMarker {
restore(): void;
ensure(): void;
}
export function waitForDocumentElement(): Promise<HTMLElement> {
if (document.documentElement) return Promise.resolve(document.documentElement);
return new Promise((resolve) => {
const observer = new MutationObserver(() => {
if (!document.documentElement) return;
observer.disconnect();
resolve(document.documentElement);
});
observer.observe(document, { childList: true });
});
}
export function markRecordingDocument(
producerId: string,
root: HTMLElement = document.documentElement,
): RecordingDocumentMarker {
const previous = root.getAttribute(RECORD_DOCUMENT_ATTRIBUTE);
const markerValue = recordingDocumentMarkerValue(producerId);
const ensure = () => root.setAttribute(RECORD_DOCUMENT_ATTRIBUTE, markerValue);
ensure();
return {
ensure,
restore() {
if (root.getAttribute(RECORD_DOCUMENT_ATTRIBUTE) !== markerValue) return;
if (previous === null) root.removeAttribute(RECORD_DOCUMENT_ATTRIBUTE);
else root.setAttribute(RECORD_DOCUMENT_ATTRIBUTE, previous);
},
};
}
@@ -0,0 +1,134 @@
import type { CdpTarget } from "@/browser-driver/frame-graph";
import { type CdpRunner, sendToCdpTarget } from "@/tools/shared";
const SETTLE_MIN_MS = 150;
const SETTLE_QUIET_MS = 250;
const SETTLE_MAX_MS = 2_000;
const SETTLE_POLL_MS = 60;
const SETTLE_WORLD_NAME = "__bsk_record_settle__";
const QUIET_PROBE = `(() => {
const scope = window;
let probe = scope.__bskRecordQuiet;
if (!probe) {
probe = { changedAt: Date.now() };
const observer = new MutationObserver(() => {
probe.changedAt = Date.now();
});
observer.observe(document.documentElement, {
subtree: true,
childList: true,
attributes: true,
characterData: true,
});
scope.__bskRecordQuiet = probe;
}
return { idleMs: Date.now() - probe.changedAt, readyState: document.readyState };
})()`;
export interface DocumentSettleScope {
target: CdpTarget;
/** Omit only for the target's root document. */
frameId?: string;
}
type SettleOutcome = "quiet" | "timeout" | "cancelled";
interface QuietProbe {
idleMs: number;
readyState: string;
}
interface ProbeContext {
executionContextId?: number;
}
function sleep(ms: number, signal?: AbortSignal): Promise<void> {
return new Promise((resolve) => {
if (signal?.aborted) return resolve();
const timer = setTimeout(resolve, ms);
signal?.addEventListener(
"abort",
() => {
clearTimeout(timer);
resolve();
},
{ once: true },
);
});
}
async function executionContextForScope(
cdp: CdpRunner,
scope: DocumentSettleScope,
): Promise<number | undefined> {
if (!scope.frameId) return undefined;
const result = await sendToCdpTarget<{ executionContextId?: number }>(
cdp,
scope.target,
"Page.createIsolatedWorld",
{
frameId: scope.frameId,
worldName: SETTLE_WORLD_NAME,
grantUniveralAccess: false,
},
);
return result.executionContextId;
}
async function readQuietProbe(
cdp: CdpRunner,
scope: DocumentSettleScope,
context: ProbeContext,
): Promise<QuietProbe | null> {
try {
if (scope.frameId && context.executionContextId === undefined) {
context.executionContextId = await executionContextForScope(cdp, scope);
if (context.executionContextId === undefined) return null;
}
const reply = await sendToCdpTarget<{ result?: { value?: unknown } }>(
cdp,
scope.target,
"Runtime.evaluate",
{
expression: QUIET_PROBE,
returnByValue: true,
...(context.executionContextId !== undefined
? { contextId: context.executionContextId }
: {}),
},
);
const value = reply.result?.value;
if (!value || typeof value !== "object") return null;
const { idleMs, readyState } = value as { idleMs?: unknown; readyState?: unknown };
if (typeof idleMs !== "number" || typeof readyState !== "string") return null;
return { idleMs, readyState };
} catch {
context.executionContextId = undefined;
return null;
}
}
export async function waitForDocumentSettled(
cdp: CdpRunner,
scope: DocumentSettleScope,
options: { signal?: AbortSignal } = {},
): Promise<SettleOutcome> {
const startedAt = Date.now();
const floor = startedAt + SETTLE_MIN_MS;
const deadline = startedAt + SETTLE_MAX_MS;
const context: ProbeContext = {};
for (;;) {
if (options.signal?.aborted) return "cancelled";
await sleep(SETTLE_POLL_MS, options.signal);
if (options.signal?.aborted) return "cancelled";
const probe = await readQuietProbe(cdp, scope, context);
const now = Date.now();
if (now < floor) continue;
if (now >= deadline) return "timeout";
if (!probe || probe.readyState === "loading") continue;
if (probe.idleMs >= SETTLE_QUIET_MS) return "quiet";
}
}
@@ -0,0 +1,34 @@
import type { RecordingDraftStep } from "./types";
const CLIPBOARD_KEYS = new Set(["a", "c", "v", "x", "A", "C", "V", "X"]);
const MODIFIER_ONLY_KEYS = new Set(["Meta", "Control", "Alt", "Shift", "OS", "Hyper", "Super"]);
export function shouldRecordPress(
key: string,
modifiers?: Array<"alt" | "ctrl" | "meta" | "shift">,
): boolean {
if (MODIFIER_ONLY_KEYS.has(key)) return false;
const mods = modifiers ?? [];
const hasCtrlOrMeta = mods.includes("ctrl") || mods.includes("meta");
if (hasCtrlOrMeta && CLIPBOARD_KEYS.has(key)) return false;
if (key === "Enter" || key === "Escape") return true;
if (key.length === 1 && !hasCtrlOrMeta && !mods.includes("alt")) return false;
return false;
}
export function shouldIncludeDraft(step: RecordingDraftStep): boolean {
return step.op !== "press" || shouldRecordPress(step.key, step.modifiers);
}
export function resolveDraftStartUrl(
drafts: RecordingDraftStep[],
explicitStartUrl?: string,
fallbackUrl?: string,
): string {
if (explicitStartUrl) return explicitStartUrl;
const navigation = drafts.find(
(step): step is Extract<RecordingDraftStep, { op: "navigate" }> => step.op === "navigate",
);
if (navigation) return navigation.url;
return drafts.find((step) => step.pageUrl)?.pageUrl ?? fallbackUrl ?? "about:blank";
}
@@ -0,0 +1,65 @@
export const RECORD_FRAME_PORT = "bsk-record-frame";
export const RECORD_FRAME_QUERY = "bsk-record-frame-query";
export const RECORD_FRAME_START = "bsk-record-frame-start";
export interface RecordFrameQueryMessage {
type: typeof RECORD_FRAME_QUERY;
}
export interface RecordFrameQueryResponse {
active: boolean;
requestId?: string;
startedAtMs?: number;
}
export interface RecordFrameStartMessage {
type: typeof RECORD_FRAME_START;
requestId: string;
startedAtMs: number;
}
export type RecordFramePortMessage =
| {
type: "ready";
requestId: string;
producerId: string;
}
| {
type: "ready_ack";
requestId: string;
producerId: string;
}
| {
type: "stop";
requestId: string;
commandId: string;
}
| {
type: "cancel";
requestId: string;
}
| {
type: "stopped";
requestId: string;
commandId: string;
ok: boolean;
error?: string;
};
export function isRecordFrameQueryMessage(value: unknown): value is RecordFrameQueryMessage {
return (
typeof value === "object" &&
value !== null &&
(value as { type?: unknown }).type === RECORD_FRAME_QUERY
);
}
export function isRecordFrameStartMessage(value: unknown): value is RecordFrameStartMessage {
if (typeof value !== "object" || value === null) return false;
const message = value as Partial<RecordFrameStartMessage>;
return (
message.type === RECORD_FRAME_START &&
typeof message.requestId === "string" &&
typeof message.startedAtMs === "number"
);
}
@@ -0,0 +1,399 @@
import {
isRecordFrameQueryMessage,
RECORD_FRAME_PORT,
RECORD_FRAME_START,
type RecordFramePortMessage,
type RecordFrameQueryResponse,
type RecordFrameStartMessage,
} from "./frame-bridge";
export interface RecordingCaptureScope {
tabId: number;
documentId: string;
browserFrameId: number;
producerId: string;
}
interface ArmedRecording {
requestId: string;
startedAtMs: number;
tabIds: Set<number>;
agents: Map<string, FrameAgent>;
onDocumentReady?: (scope: RecordingCaptureScope) => void | Promise<void>;
finishing: boolean;
}
interface FrameAgent extends RecordingCaptureScope {
requestId: string;
port: chrome.runtime.Port;
stopWaiters: Map<string, (ok: boolean) => void>;
}
interface BrowserFrame {
frameId: number;
documentId?: string;
}
interface BrowserFrameNavigation extends BrowserFrame {
tabId: number;
lifecycle: "committed" | "completed";
}
const RECORD_FRAME_STOP_TIMEOUT_MS = 5_000;
export interface RecordFrameCoordinatorDeps {
getAllFrames(tabId: number): Promise<BrowserFrame[]>;
sendToDocument(
tabId: number,
message: RecordFrameStartMessage,
target: { documentId?: string; frameId?: number },
): Promise<unknown>;
subscribeFrameNavigation?(listener: (frame: BrowserFrameNavigation) => void): () => void;
}
function documentKey(tabId: number, documentId: string): string {
return `${tabId}:${documentId}`;
}
function senderAddress(
sender: chrome.runtime.MessageSender,
): { tabId: number; documentId: string; browserFrameId: number } | null {
const tabId = sender.tab?.id;
const documentId = sender.documentId;
const browserFrameId = sender.frameId;
if (
typeof tabId !== "number" ||
typeof documentId !== "string" ||
typeof browserFrameId !== "number"
) {
return null;
}
return { tabId, documentId, browserFrameId };
}
function defaultDeps(): RecordFrameCoordinatorDeps {
return {
async getAllFrames(tabId) {
return (await chrome.webNavigation.getAllFrames({ tabId })) ?? [];
},
sendToDocument(tabId, message, target) {
return chrome.tabs.sendMessage(tabId, message, target);
},
subscribeFrameNavigation(listener) {
const onCommitted = (details: chrome.webNavigation.WebNavigationTransitionCallbackDetails) =>
listener({ ...details, lifecycle: "committed" });
const onCompleted = (details: chrome.webNavigation.WebNavigationFramedCallbackDetails) =>
listener({ ...details, lifecycle: "completed" });
chrome.webNavigation.onCommitted.addListener(onCommitted);
chrome.webNavigation.onCompleted.addListener(onCompleted);
return () => {
chrome.webNavigation.onCommitted.removeListener(onCommitted);
chrome.webNavigation.onCompleted.removeListener(onCompleted);
};
},
};
}
export class RecordFrameCoordinator {
readonly #deps: RecordFrameCoordinatorDeps;
readonly #recordings = new Map<string, ArmedRecording>();
#attached = false;
#detachFrameNavigation: (() => void) | null = null;
constructor(deps: RecordFrameCoordinatorDeps = defaultDeps()) {
this.#deps = deps;
}
attach(): () => void {
if (this.#attached) return () => {};
this.#attached = true;
chrome.runtime.onConnect.addListener(this.#onConnect);
chrome.runtime.onMessage.addListener(this.#onMessage);
this.#detachFrameNavigation =
this.#deps.subscribeFrameNavigation?.(this.#onFrameNavigation) ?? null;
return () => {
if (!this.#attached) return;
this.#attached = false;
chrome.runtime.onConnect.removeListener(this.#onConnect);
chrome.runtime.onMessage.removeListener(this.#onMessage);
this.#detachFrameNavigation?.();
this.#detachFrameNavigation = null;
for (const recording of this.#recordings.values()) this.#cancelAgents(recording);
this.#recordings.clear();
};
}
begin(
requestId: string,
startedAtMs: number,
initialTabId: number,
onDocumentReady?: (scope: RecordingCaptureScope) => void | Promise<void>,
): void {
const previous = this.#recordings.get(requestId);
if (previous) this.#cancelAgents(previous);
this.#recordings.set(requestId, {
requestId,
startedAtMs,
// Register the initial tab before navigation begins. A newly-created
// child document can query at document_start before armTab's frame
// snapshot sees it, especially when Chromium promotes it to an OOPIF.
tabIds: new Set([initialTabId]),
agents: new Map(),
...(onDocumentReady ? { onDocumentReady } : {}),
finishing: false,
});
}
async armTab(requestId: string, tabId: number): Promise<boolean> {
const recording = this.#recordings.get(requestId);
if (!recording || recording.finishing) return false;
recording.tabIds.add(tabId);
let frames: BrowserFrame[];
try {
frames = await this.#deps.getAllFrames(tabId);
} catch {
frames = [{ frameId: 0 }];
}
if (!frames.some((frame) => frame.frameId === 0)) frames.unshift({ frameId: 0 });
const results = await Promise.all(
frames.map(async (frame) => {
const started = await this.#startFrame(recording, tabId, frame);
return { frameId: frame.frameId, started };
}),
);
return results.some((result) => result.frameId === 0 && result.started);
}
sourceFor(
requestId: string,
producerId: string,
sender: chrome.runtime.MessageSender,
): RecordingCaptureScope | null {
const address = senderAddress(sender);
const recording = this.#recordings.get(requestId);
if (!address || !recording) return null;
const agent = recording.agents.get(documentKey(address.tabId, address.documentId));
if (!agent || agent.producerId !== producerId) return null;
return {
tabId: agent.tabId,
documentId: agent.documentId,
browserFrameId: agent.browserFrameId,
producerId: agent.producerId,
};
}
async stop(requestId: string): Promise<boolean> {
const recording = this.#recordings.get(requestId);
if (!recording) return true;
recording.finishing = true;
const commandId = crypto.randomUUID();
const results = await Promise.all(
[...recording.agents.values()].map((agent) => this.#stopAgent(agent, requestId, commandId)),
);
const succeeded = results.every(Boolean);
if (succeeded) this.#recordings.delete(requestId);
else recording.finishing = false;
return succeeded;
}
cancel(requestId: string): void {
const recording = this.#recordings.get(requestId);
if (!recording) return;
this.#cancelAgents(recording);
this.#recordings.delete(requestId);
}
readonly #onMessage = (
message: unknown,
sender: chrome.runtime.MessageSender,
sendResponse: (response: RecordFrameQueryResponse) => void,
): boolean => {
if (!isRecordFrameQueryMessage(message)) return false;
const tabId = sender.tab?.id;
const recording =
typeof tabId === "number"
? [...this.#recordings.values()].find(
(candidate) => !candidate.finishing && candidate.tabIds.has(tabId),
)
: undefined;
sendResponse(
recording
? {
active: true,
requestId: recording.requestId,
startedAtMs: recording.startedAtMs,
}
: { active: false },
);
return false;
};
readonly #onFrameNavigation = (frame: BrowserFrameNavigation): void => {
if (frame.frameId === 0) return;
for (const recording of this.#recordings.values()) {
if (recording.finishing || !recording.tabIds.has(frame.tabId)) continue;
// onCommitted can race document_start, while onCompleted runs after the
// content script is available. Starting twice is safe and closes both
// the early-query and late-SPA-frame gaps.
void (async () => {
const started = await this.#startFrame(recording, frame.tabId, frame);
if (!started || frame.lifecycle !== "completed" || recording.finishing) return;
const agent = frame.documentId
? recording.agents.get(documentKey(frame.tabId, frame.documentId))
: [...recording.agents.values()].find(
(candidate) =>
candidate.tabId === frame.tabId && candidate.browserFrameId === frame.frameId,
);
if (!agent) return;
await recording.onDocumentReady?.({
tabId: agent.tabId,
documentId: agent.documentId,
browserFrameId: agent.browserFrameId,
producerId: agent.producerId,
});
})();
}
};
readonly #onConnect = (port: chrome.runtime.Port): void => {
if (port.name !== RECORD_FRAME_PORT) return;
const address = senderAddress(port.sender ?? {});
if (!address) {
port.disconnect();
return;
}
let registered: FrameAgent | null = null;
port.onMessage.addListener((raw: unknown) => {
const message = raw as Partial<RecordFramePortMessage>;
if (message.type === "ready") {
if (typeof message.requestId !== "string" || typeof message.producerId !== "string") {
port.disconnect();
return;
}
const requestId = message.requestId;
const producerId = message.producerId;
if (registered) {
if (registered.requestId === requestId && registered.producerId === producerId) {
port.postMessage({
type: "ready_ack",
requestId,
producerId,
} satisfies RecordFramePortMessage);
} else {
port.disconnect();
}
return;
}
const recording = this.#recordings.get(requestId);
if (
!recording ||
recording.finishing ||
!recording.tabIds.has(address.tabId) ||
producerId.length === 0
) {
port.disconnect();
return;
}
const agent: FrameAgent = {
...address,
requestId,
producerId,
port,
stopWaiters: new Map(),
};
registered = agent;
const key = documentKey(address.tabId, address.documentId);
const previous = recording.agents.get(key);
if (previous && previous !== agent) previous.port.disconnect();
recording.agents.set(key, agent);
port.postMessage({
type: "ready_ack",
requestId,
producerId,
} satisfies RecordFramePortMessage);
return;
}
if (
message.type === "stopped" &&
registered &&
message.requestId === registered.requestId &&
typeof message.commandId === "string"
) {
registered.stopWaiters.get(message.commandId)?.(message.ok === true);
registered.stopWaiters.delete(message.commandId);
}
});
port.onDisconnect.addListener(() => {
if (!registered) return;
for (const resolve of registered.stopWaiters.values()) resolve(true);
const recording = this.#recordings.get(registered.requestId);
const key = documentKey(registered.tabId, registered.documentId);
if (recording?.agents.get(key) === registered) recording.agents.delete(key);
});
};
#stopAgent(agent: FrameAgent, requestId: string, commandId: string): Promise<boolean> {
return new Promise((resolve) => {
const timer = setTimeout(() => {
agent.stopWaiters.delete(commandId);
resolve(false);
}, RECORD_FRAME_STOP_TIMEOUT_MS);
agent.stopWaiters.set(commandId, (ok) => {
clearTimeout(timer);
resolve(ok);
});
try {
agent.port.postMessage({
type: "stop",
requestId,
commandId,
} satisfies RecordFramePortMessage);
} catch {
clearTimeout(timer);
agent.stopWaiters.delete(commandId);
resolve(true);
}
});
}
async #startFrame(
recording: ArmedRecording,
tabId: number,
frame: BrowserFrame,
): Promise<boolean> {
const message: RecordFrameStartMessage = {
type: RECORD_FRAME_START,
requestId: recording.requestId,
startedAtMs: recording.startedAtMs,
};
try {
const response = await this.#deps.sendToDocument(tabId, message, {
...(frame.documentId ? { documentId: frame.documentId } : { frameId: frame.frameId }),
});
return isStarted(response);
} catch {
return false;
}
}
#cancelAgents(recording: ArmedRecording): void {
for (const agent of recording.agents.values()) {
try {
agent.port.postMessage({
type: "cancel",
requestId: recording.requestId,
} satisfies RecordFramePortMessage);
} catch {
// A destroyed Document has no remaining capture state to cancel.
}
}
}
}
function isStarted(value: unknown): boolean {
return typeof value === "object" && value !== null && (value as { ok?: unknown }).ok === true;
}
export const recordFrameCoordinator = new RecordFrameCoordinator();
@@ -0,0 +1,5 @@
const REDIRECT_QUALIFIERS = new Set(["client_redirect", "server_redirect"]);
export function hasRedirectQualifier(qualifiers?: readonly string[]): boolean {
return (qualifiers ?? []).some((qualifier) => REDIRECT_QUALIFIERS.has(qualifier));
}
@@ -0,0 +1,140 @@
import type { RenderedRef } from "@browser-skill/vom";
import type { CdpTarget } from "@/browser-driver/frame-graph";
import {
type CaptureVomMatchNode,
type CaptureVomObservationResult,
captureVomObservation,
} from "@/tools/capture-vom-observation";
import type { CdpRunner, ChromeTabsApi } from "@/tools/shared";
export interface IndexedObservationNode {
frameId: string;
geometry: CaptureVomMatchNode;
ref?: RenderedRef;
}
export interface CapturedRecordingObservation {
rootFrameId: string;
index: ObservationNodeIndex;
url: string;
title?: string;
vomText: string;
truncated: boolean;
}
export interface RegisteredObservation {
stateId: string;
rootFrameId: string;
index: ObservationNodeIndex;
url: string;
}
export interface RecordingDocumentScope {
frameId: string;
target: CdpTarget;
}
function nodeKey(frameId: string, backendNodeId: number): string {
return `${frameId}:${backendNodeId}`;
}
function frameTagKey(frameId: string, tag: string): string {
return `${frameId}:${tag.toLowerCase()}`;
}
export class ObservationNodeIndex {
readonly #nodesByFrameTag = new Map<string, IndexedObservationNode[]>();
readonly #refById = new Map<string, RenderedRef>();
readonly #refsByFrame = new Map<string, RenderedRef[]>();
readonly #scopeByProducer = new Map<string, RecordingDocumentScope | null>();
constructor(
input: Pick<CaptureVomObservationResult, "rootFrameId" | "matchNodes" | "refs"> &
Partial<Pick<CaptureVomObservationResult, "frames">>,
) {
const refByNode = new Map<string, RenderedRef>();
for (const ref of input.refs) {
const frameId = ref.frameId ?? input.rootFrameId;
refByNode.set(nodeKey(frameId, ref.backendNodeId), ref);
this.#refById.set(ref.ref, ref);
const frameRefs = this.#refsByFrame.get(frameId) ?? [];
frameRefs.push(ref);
this.#refsByFrame.set(frameId, frameRefs);
}
for (const frame of input.frames ?? []) {
const documentId = frame.recordingDocumentId;
if (!documentId) continue;
this.#scopeByProducer.set(
documentId,
this.#scopeByProducer.has(documentId)
? null
: { frameId: frame.frameId, target: frame.target },
);
}
for (const geometry of input.matchNodes) {
const { frameId } = geometry;
const entry = {
frameId,
geometry,
ref: refByNode.get(nodeKey(frameId, geometry.backendNodeId)),
};
const key = frameTagKey(frameId, geometry.tag);
const bucket = this.#nodesByFrameTag.get(key) ?? [];
bucket.push(entry);
this.#nodesByFrameTag.set(key, bucket);
}
}
candidates(frameId: string, tag: string): readonly IndexedObservationNode[] {
return this.#nodesByFrameTag.get(frameTagKey(frameId, tag)) ?? [];
}
ref(refId: string): RenderedRef | undefined {
return this.#refById.get(refId);
}
refs(frameId: string): readonly RenderedRef[] {
return this.#refsByFrame.get(frameId) ?? [];
}
documentScope(producerId: string): RecordingDocumentScope | undefined {
return this.#scopeByProducer.get(producerId) ?? undefined;
}
}
async function readTabMeta(
tabsApi: ChromeTabsApi,
tabId: number,
): Promise<{ url: string; title?: string }> {
try {
const tab = await tabsApi.get(tabId);
return { url: tab.url ?? "about:blank", title: tab.title };
} catch {
return { url: "about:blank" };
}
}
export async function captureRecordingObservation(input: {
cdp: CdpRunner;
tabsApi: ChromeTabsApi;
tabId: number;
maxTokens: number;
redactValues: boolean;
signal?: AbortSignal;
}): Promise<CapturedRecordingObservation> {
const { url, title } = await readTabMeta(input.tabsApi, input.tabId);
const captured = await captureVomObservation(input.cdp, input.tabId, url, {
maxTokens: input.maxTokens,
redactValues: input.redactValues,
conditionalSurfaceProbe: false,
signal: input.signal,
});
return {
rootFrameId: captured.rootFrameId,
index: new ObservationNodeIndex(captured),
url,
title,
vomText: captured.text,
truncated: captured.truncated,
};
}
@@ -0,0 +1,128 @@
import type { CdpRunner, ChromeTabsApi } from "@/tools/shared";
import { captureRecordingObservation, type RegisteredObservation } from "./observation-capture";
import { RecordingStateRegistry } from "./state-registry";
import { matchObservationTarget, unmatchedTarget } from "./target-matcher";
import type { RecordingDraftStep, StepAnnotation, TargetedRecordingDraft } from "./types";
const DEFAULT_MAX_PAGE_TOKENS = 3_000;
const MIN_CAPTURE_INTERVAL_MS = 200;
export interface TabObservationCursor {
lastSettled: RegisteredObservation | null;
lastCaptureAt: number;
}
function abortableDelay(ms: number, signal?: AbortSignal): Promise<void> {
return new Promise((resolve) => {
if (signal?.aborted) return resolve();
const timer = setTimeout(resolve, ms);
signal?.addEventListener(
"abort",
() => {
clearTimeout(timer);
resolve();
},
{ once: true },
);
});
}
function isTargeted(draft: RecordingDraftStep): draft is TargetedRecordingDraft {
return (
draft.op === "click" ||
draft.op === "hover" ||
draft.op === "fill" ||
draft.op === "press" ||
draft.op === "select"
);
}
export class RecordingObservationSession {
readonly registry: RecordingStateRegistry;
readonly cursor: TabObservationCursor;
readonly annotations: StepAnnotation[];
readonly #maxTokens: number;
readonly #redactValues: boolean;
constructor(
options: {
registry?: RecordingStateRegistry;
cursor?: TabObservationCursor;
annotations?: StepAnnotation[];
maxTokens?: number;
redactValues?: boolean;
} = {},
) {
this.registry = options.registry ?? new RecordingStateRegistry();
this.cursor = options.cursor ?? { lastSettled: null, lastCaptureAt: 0 };
this.annotations = options.annotations ?? [];
this.#maxTokens = options.maxTokens ?? DEFAULT_MAX_PAGE_TOKENS;
this.#redactValues = options.redactValues ?? false;
}
async capture(
cdp: CdpRunner,
tabsApi: ChromeTabsApi,
tabId: number,
signal?: AbortSignal,
): Promise<RegisteredObservation> {
const waitMs = Math.max(0, MIN_CAPTURE_INTERVAL_MS - (Date.now() - this.cursor.lastCaptureAt));
if (waitMs > 0) await abortableDelay(waitMs, signal);
if (signal?.aborted) throw new DOMException("observation aborted", "AbortError");
const captured = await captureRecordingObservation({
cdp,
tabsApi,
tabId,
maxTokens: this.#maxTokens,
redactValues: this.#redactValues,
signal,
});
const state = this.registry.register({
url: captured.url,
title: captured.title,
vomText: captured.vomText,
truncated: captured.truncated,
});
const observation: RegisteredObservation = {
stateId: state.id,
rootFrameId: captured.rootFrameId,
index: captured.index,
url: captured.url,
};
this.cursor.lastSettled = observation;
this.cursor.lastCaptureAt = Date.now();
return observation;
}
bindDraft(draft: RecordingDraftStep, draftId: number, previousActionPending = false): void {
const observation = this.cursor.lastSettled;
if (isTargeted(draft)) {
draft.matchedTarget = observation
? matchObservationTarget({
observation,
hint: draft.targetHint,
fallback: draft.captureTarget,
})
: unmatchedTarget(draft.captureTarget);
}
if (!observation) return;
const unmatched = isTargeted(draft) && draft.matchedTarget?.unmatched === true;
if (previousActionPending && unmatched) return;
draft.preStateId = observation.stateId;
this.registry.markStep(observation.stateId, draftId);
if (!isTargeted(draft) || !draft.matchedTarget?.ref) return;
const ref = observation.index.ref(draft.matchedTarget.ref);
if (!ref) return;
this.annotations.push({
draftId,
op: draft.op,
line: ref.line,
stateId: observation.stateId,
...(draft.op === "fill" && !this.#redactValues
? { detail: JSON.stringify(draft.value) }
: {}),
});
}
}
@@ -0,0 +1,277 @@
import type { CdpRunner, ChromeTabsApi } from "@/tools/shared";
import type { StopReason, TraceV3 } from "@/transport/types";
import { type DocumentSettleScope, waitForDocumentSettled } from "./document-settle";
import type { RegisteredObservation } from "./observation-capture";
import { RecordingObservationSession } from "./observation-session";
import { inferMissingPostStates, SettleController } from "./settle-controller";
import { RecordingStateRegistry } from "./state-registry";
import { buildTraceV3 } from "./trace-builder-v3";
import type { RecordingDraftStep, StepAnnotation } from "./types";
interface TabRecordingContext {
session: RecordingObservationSession;
settle: SettleController;
pendingCapture: PendingCapture | null;
frameRefresh: Promise<void>;
frameRefreshAbort: AbortController | null;
}
interface PendingCapture {
promise: Promise<RegisteredObservation>;
abort: AbortController;
}
export class RecordingObservationRuntime {
readonly #cdp: CdpRunner;
readonly #tabsApi: ChromeTabsApi;
readonly #registry = new RecordingStateRegistry();
readonly #annotations: StepAnnotation[] = [];
readonly #contexts = new Map<number, TabRecordingContext>();
readonly #maxTokens?: number;
readonly #redactValues: boolean;
constructor(input: {
cdp: CdpRunner;
tabsApi: ChromeTabsApi;
maxTokens?: number;
redactValues?: boolean;
}) {
this.#cdp = input.cdp;
this.#tabsApi = input.tabsApi;
this.#maxTokens = input.maxTokens;
this.#redactValues = input.redactValues ?? false;
}
#context(tabId: number): TabRecordingContext {
const existing = this.#contexts.get(tabId);
if (existing) return existing;
const session = new RecordingObservationSession({
registry: this.#registry,
annotations: this.#annotations,
maxTokens: this.#maxTokens,
redactValues: this.#redactValues,
});
const context: TabRecordingContext = {
session,
settle: new SettleController({
session,
cdp: this.#cdp,
tabsApi: this.#tabsApi,
tabId,
}),
pendingCapture: null,
frameRefresh: Promise.resolve(),
frameRefreshAbort: null,
};
this.#contexts.set(tabId, context);
return context;
}
async #capture(tabId: number): Promise<RegisteredObservation> {
const context = this.#context(tabId);
if (context.pendingCapture) return context.pendingCapture.promise;
const abort = new AbortController();
let pending!: PendingCapture;
const promise = context.session.capture(this.#cdp, this.#tabsApi, tabId, abort.signal);
pending = {
abort,
promise: promise.finally(() => {
if (context.pendingCapture === pending) context.pendingCapture = null;
}),
};
context.pendingCapture = pending;
return pending.promise;
}
async captureInitial(tabId: number): Promise<void> {
const context = this.#context(tabId);
if (context.session.cursor.lastSettled) return;
await this.#capture(tabId);
}
/** Refresh the safe observation after a late child Document becomes usable. */
async refreshDocument(tabId: number, producerId: string): Promise<void> {
const context = this.#context(tabId);
context.frameRefreshAbort?.abort();
const abort = new AbortController();
context.frameRefreshAbort = abort;
const previous = context.frameRefresh;
const refresh = (async () => {
await previous;
if (abort.signal.aborted) return;
if (context.pendingCapture) await Promise.allSettled([context.pendingCapture.promise]);
if (abort.signal.aborted) return;
// The first capture discovers the safe producer -> CDP Document scope.
await this.#capture(tabId);
if (abort.signal.aborted) return;
const scope = context.session.cursor.lastSettled?.index.documentScope(producerId);
if (scope) {
await waitForDocumentSettled(this.#cdp, scope, { signal: abort.signal });
}
if (abort.signal.aborted) return;
// Capture again after the child SPA is quiet so the next action can be
// matched against refs that actually exist in its pre-action state.
await this.#capture(tabId);
})();
const tracked = refresh.catch(() => {});
context.frameRefresh = tracked;
try {
await refresh;
} finally {
if (context.frameRefresh === tracked) context.frameRefreshAbort = null;
}
}
async captureTabTransition(
fromTabId: number,
toTabId: number,
): Promise<{ preStateId?: string; postStateId?: string; targetUrl?: string }> {
const from = this.#context(fromTabId);
await this.#flushContext(from);
if (!from.session.cursor.lastSettled) {
try {
await this.captureInitial(fromTabId);
} catch {
// A target-tab observation can still preserve the forward transition.
}
}
const to = this.#context(toTabId);
let target = to.session.cursor.lastSettled;
try {
target = await this.#capture(toTabId);
} catch {
// Reuse the last settled target state when a refresh races navigation.
}
const source = from.session.cursor.lastSettled;
return {
...(source ? { preStateId: source.stateId } : {}),
...(target ? { postStateId: target.stateId, targetUrl: target.url } : {}),
};
}
bindTabTransition(
draft: Extract<RecordingDraftStep, { op: "switch_tab" }>,
draftId: number,
): void {
if (draft.preStateId) this.#registry.markStep(draft.preStateId, draftId);
}
async processDraft(
tabId: number,
drafts: RecordingDraftStep[],
draftIndex: number,
producerId?: string,
): Promise<void> {
const draft = drafts[draftIndex];
if (!draft) return;
const context = this.#context(tabId);
if (!context.session.cursor.lastSettled) {
try {
await this.captureInitial(tabId);
} catch {
// Post-action settle can still provide a usable state.
}
}
const scope = producerId
? context.session.cursor.lastSettled?.index.documentScope(producerId)
: undefined;
if (
producerId &&
(draft.op === "click" ||
draft.op === "hover" ||
draft.op === "fill" ||
draft.op === "press" ||
draft.op === "select")
) {
draft.targetHint = {
...(draft.targetHint ?? {}),
frameId: scope?.frameId ?? null,
geometrySpace: "local",
};
}
context.session.bindDraft(draft, draftIndex + 1, context.settle.hasPending);
context.settle.schedule(drafts, draftIndex, scope);
}
scheduleSettle(
tabId: number,
drafts: RecordingDraftStep[],
draftIndex: number,
scope?: DocumentSettleScope,
): void {
this.#context(tabId).settle.schedule(drafts, draftIndex, scope);
}
clearRedirect(tabId: number): void {
this.#contexts.get(tabId)?.settle.clearRedirect();
}
scheduleRedirect(tabId: number, drafts: RecordingDraftStep[], url: string): void {
this.#context(tabId).settle.scheduleRedirect(drafts, url);
}
async #flushContext(context: TabRecordingContext): Promise<void> {
await context.frameRefresh;
if (context.pendingCapture) await Promise.allSettled([context.pendingCapture.promise]);
await context.settle.flushRedirects();
await context.settle.flush();
}
async flushRedirects(tabId: number): Promise<void> {
await this.#contexts.get(tabId)?.settle.flushRedirects();
}
async flush(): Promise<void> {
for (const context of this.#contexts.values()) await this.#flushContext(context);
}
async settleTrailing(tabId: number, drafts: RecordingDraftStep[]): Promise<void> {
const context = this.#context(tabId);
if (context.pendingCapture) await Promise.allSettled([context.pendingCapture.promise]);
if (!context.session.cursor.lastSettled) {
try {
await this.captureInitial(tabId);
} catch {
// The trace may remain observation-free when CDP is unavailable.
}
}
await context.settle.settleTrailing(drafts);
inferMissingPostStates(drafts);
}
cancel(): void {
for (const context of this.#contexts.values()) {
context.frameRefreshAbort?.abort();
context.pendingCapture?.abort.abort();
context.settle.cancel();
}
}
buildTrace(input: {
drafts: RecordingDraftStep[];
startedAt: string;
purpose?: string;
startUrl?: string;
stoppedBy: StopReason;
bskVersion: string;
includeTabSwitches?: boolean;
}): TraceV3 {
return buildTraceV3({
registry: this.#registry,
drafts: input.drafts,
annotations: this.#annotations,
startedAt: input.startedAt,
purpose: input.purpose,
startUrl: input.startUrl,
stoppedBy: input.stoppedBy,
bskVersion: input.bskVersion,
redactValues: this.#redactValues,
includeTabSwitches: input.includeTabSwitches,
});
}
}
@@ -0,0 +1,233 @@
import type { CdpRunner, ChromeTabsApi } from "@/tools/shared";
import { type DocumentSettleScope, waitForDocumentSettled } from "./document-settle";
import { RecordingObservationSession } from "./observation-session";
import type { RecordingDraftStep } from "./types";
interface PendingSettle {
abort: AbortController;
scope: DocumentSettleScope;
}
interface PendingRedirect {
url: string;
abort: AbortController;
}
const CAPTURE_RETRY_DELAY_MS = 250;
function delay(ms: number, signal?: AbortSignal): Promise<void> {
return new Promise((resolve) => {
if (signal?.aborted) return resolve();
const timer = setTimeout(resolve, ms);
signal?.addEventListener(
"abort",
() => {
clearTimeout(timer);
resolve();
},
{ once: true },
);
});
}
function isAbortError(error: unknown): boolean {
return (
typeof error === "object" &&
error !== null &&
(error as { name?: string }).name === "AbortError"
);
}
export function inferMissingPostStates(drafts: RecordingDraftStep[]): void {
for (let index = 0; index < drafts.length - 1; index += 1) {
const draft = drafts[index];
const next = drafts[index + 1];
if (draft && next && !draft.postStateId && next.preStateId) {
draft.postStateId = next.preStateId;
}
}
}
export class SettleController {
readonly #session: RecordingObservationSession;
readonly #cdp: CdpRunner;
readonly #tabsApi: ChromeTabsApi;
readonly #tabId: number;
readonly #rootScope: DocumentSettleScope;
readonly #pending = new Map<number, PendingSettle>();
#queue = Promise.resolve();
#pendingRedirect: PendingRedirect | null = null;
#redirectQueue = Promise.resolve();
constructor(input: {
session: RecordingObservationSession;
cdp: CdpRunner;
tabsApi: ChromeTabsApi;
tabId: number;
}) {
this.#session = input.session;
this.#cdp = input.cdp;
this.#tabsApi = input.tabsApi;
this.#tabId = input.tabId;
this.#rootScope = { target: { tabId: input.tabId } };
}
get hasPending(): boolean {
return this.#pending.size > 0 || this.#pendingRedirect !== null;
}
async #captureWithRetry(signal?: AbortSignal) {
try {
return await this.#session.capture(this.#cdp, this.#tabsApi, this.#tabId, signal);
} catch (error) {
if (signal?.aborted) throw error;
await delay(CAPTURE_RETRY_DELAY_MS, signal);
return this.#session.capture(this.#cdp, this.#tabsApi, this.#tabId, signal);
}
}
schedule(
drafts: RecordingDraftStep[],
draftIndex: number,
scope: DocumentSettleScope = this.#rootScope,
): void {
if (scope.target.tabId !== this.#tabId) {
throw new Error(
`settle scope tab ${scope.target.tabId} does not belong to tab ${this.#tabId}`,
);
}
const landing = drafts[draftIndex]?.preStateId;
for (const [index, pending] of this.#pending) {
if (index >= draftIndex) continue;
pending.abort.abort();
this.#pending.delete(index);
if (landing && drafts[index] && !drafts[index].postStateId) {
drafts[index].postStateId = landing;
}
}
this.#pending.get(draftIndex)?.abort.abort();
const pending = { abort: new AbortController(), scope };
this.#pending.set(draftIndex, pending);
const task = async () => {
if (pending.abort.signal.aborted) return;
try {
const outcome = await waitForDocumentSettled(this.#cdp, pending.scope, {
signal: pending.abort.signal,
});
if (outcome === "cancelled") return;
const observation = await this.#captureWithRetry(pending.abort.signal);
if (!pending.abort.signal.aborted && drafts[draftIndex]) {
drafts[draftIndex].postStateId = observation.stateId;
}
} catch (error) {
if (!isAbortError(error)) {
console.warn(
`[bsk record] post-action observation failed for step ${draftIndex + 1}`,
error,
);
}
} finally {
if (this.#pending.get(draftIndex) === pending) this.#pending.delete(draftIndex);
}
};
this.#queue = this.#queue.then(task, task).catch(() => {});
}
cancel(): void {
for (const pending of this.#pending.values()) pending.abort.abort();
this.#pending.clear();
this.clearRedirect();
}
async flush(): Promise<void> {
for (let round = 0; round < 10; round += 1) {
const current = this.#queue;
await current;
if (current === this.#queue) return;
}
console.warn("[bsk record] settle queue kept growing; using completed observations");
}
async settleTrailing(drafts: RecordingDraftStep[]): Promise<void> {
const trailing: RecordingDraftStep[] = [];
for (let index = drafts.length - 1; index >= 0; index -= 1) {
const draft = drafts[index];
if (!draft || draft.postStateId) break;
trailing.push(draft);
}
if (trailing.length === 0) return;
try {
const observation = await this.#captureWithRetry();
for (const draft of trailing) draft.postStateId = observation.stateId;
} catch (error) {
console.warn("[bsk record] final observation at stop failed", error);
}
}
clearRedirect(): void {
this.#pendingRedirect?.abort.abort();
this.#pendingRedirect = null;
}
scheduleRedirect(drafts: RecordingDraftStep[], url: string): void {
this.#pendingRedirect?.abort.abort();
const pending: PendingRedirect = { url, abort: new AbortController() };
this.#pendingRedirect = pending;
const task = () => this.#settleRedirect(drafts, pending);
this.#redirectQueue = this.#redirectQueue.then(task, task).catch(() => {});
}
async #settleRedirect(drafts: RecordingDraftStep[], pending: PendingRedirect): Promise<void> {
const outcome = await waitForDocumentSettled(this.#cdp, this.#rootScope, {
signal: pending.abort.signal,
});
if (outcome === "cancelled" || this.#pendingRedirect !== pending) return;
let finalUrl = pending.url;
try {
finalUrl = (await this.#tabsApi.get(this.#tabId)).url || finalUrl;
} catch {
// The navigation event URL remains the best available destination.
}
if (this.#pendingRedirect !== pending) return;
this.#pendingRedirect = null;
if (
!finalUrl ||
finalUrl === "about:blank" ||
this.#session.cursor.lastSettled?.url === finalUrl
) {
return;
}
const last = drafts[drafts.length - 1];
if (last?.op === "navigate" && last.url === finalUrl) {
if (!last.postStateId) this.schedule(drafts, drafts.length - 1);
await this.flush();
return;
}
const draft: RecordingDraftStep = {
op: "navigate",
url: finalUrl,
pageUrl: finalUrl,
cause: "browser",
transitionQualifiers: ["server_redirect"],
preStateId: this.#session.cursor.lastSettled?.stateId,
};
drafts.push(draft);
const draftIndex = drafts.length - 1;
if (draft.preStateId) this.#session.registry.markStep(draft.preStateId, draftIndex + 1);
this.schedule(drafts, draftIndex);
await this.flush();
}
async flushRedirects(): Promise<void> {
for (let round = 0; round < 10; round += 1) {
const current = this.#redirectQueue;
await current;
if (current === this.#redirectQueue) return;
}
console.warn("[bsk record] redirect queue kept growing; using the final completed landing");
}
}
@@ -0,0 +1,57 @@
export interface RecordedStateEntry {
id: string;
url: string;
title?: string;
/** Rendered VOM text; capture indexes and source DOM/AX data never enter the registry. */
vomText: string;
truncated: boolean;
stepsHere: number[];
}
function stateIdentity(url: string, body: string): string {
return `${url}\0${body}`;
}
export class RecordingStateRegistry {
readonly #entriesById = new Map<string, RecordedStateEntry>();
readonly #idByIdentity = new Map<string, string>();
#nextId = 1;
register(input: {
url: string;
title?: string;
vomText: string;
truncated?: boolean;
}): RecordedStateEntry {
const identity = stateIdentity(input.url, input.vomText);
const existingId = this.#idByIdentity.get(identity);
if (existingId) {
const existing = this.#entriesById.get(existingId)!;
if (!existing.title && input.title) existing.title = input.title;
if (input.truncated) existing.truncated = true;
return existing;
}
const entry: RecordedStateEntry = {
id: `s${this.#nextId}`,
url: input.url,
...(input.title ? { title: input.title } : {}),
vomText: input.vomText,
truncated: input.truncated ?? false,
stepsHere: [],
};
this.#nextId += 1;
this.#entriesById.set(entry.id, entry);
this.#idByIdentity.set(identity, entry.id);
return entry;
}
values(): RecordedStateEntry[] {
return [...this.#entriesById.values()];
}
markStep(stateId: string, draftId: number): void {
const entry = this.#entriesById.get(stateId);
if (entry && !entry.stepsHere.includes(draftId)) entry.stepsHere.push(draftId);
}
}
@@ -0,0 +1,155 @@
import type { RecordStepPayload } from "../record-bridge";
import { hasRedirectQualifier } from "./navigation-policy";
import type { RecordingDraftStep, TargetMatchHint } from "./types";
export interface RecordingStepBuffer {
steps: RecordingDraftStep[];
navigation: RecordingNavigationCursor;
}
export interface RecordingNavigationCursor {
currentUrl?: string;
pendingNavigation: boolean;
pendingNavigationDeadline?: number;
}
const NAVIGATION_TRIGGER_WINDOW_MS = 3_000;
function toDraftStep(
payload: RecordStepPayload,
targetHint?: TargetMatchHint,
): RecordingDraftStep | null {
const pageUrl = payload.page_url;
const common = {
...(pageUrl ? { pageUrl } : {}),
...(targetHint ? { targetHint } : {}),
};
switch (payload.op) {
case "click":
return payload.target ? { op: "click", captureTarget: payload.target, ...common } : null;
case "hover":
return payload.target ? { op: "hover", captureTarget: payload.target, ...common } : null;
case "fill":
return payload.target
? {
op: "fill",
captureTarget: payload.target,
value: payload.value ?? "",
...(payload.commit ? { commit: payload.commit } : {}),
...(payload.redacted ? { redacted: true } : {}),
...common,
}
: null;
case "press":
return payload.key
? {
op: "press",
key: payload.key,
...(payload.target ? { captureTarget: payload.target } : {}),
...(payload.modifiers?.length ? { modifiers: payload.modifiers } : {}),
...common,
}
: null;
case "select":
return payload.target && payload.values
? {
op: "select",
captureTarget: payload.target,
values: payload.values,
...(payload.labels?.length ? { labels: payload.labels } : {}),
...common,
}
: null;
case "navigate":
return null;
}
}
function annotateLastStepNavigation(buffer: RecordingStepBuffer, url: string): number {
for (let index = buffer.steps.length - 1; index >= 0; index -= 1) {
const step = buffer.steps[index];
if (!step) continue;
if (step.op === "click" || step.op === "press" || step.op === "select" || step.op === "fill") {
step.navigatedTo = url;
return index;
}
break;
}
return -1;
}
export type NavigationObserveResult =
| { kind: "noop" }
| { kind: "annotated"; index: number }
| { kind: "appended"; index: number }
| { kind: "coalesce_redirect"; url: string };
export function observeRecordedNavigation(
buffer: RecordingStepBuffer,
url: string,
causedByAction?: boolean,
transitionType?: string,
transitionQualifiers?: string[],
): NavigationObserveResult {
const navigation = buffer.navigation;
if (!url || url === navigation.currentUrl) return { kind: "noop" };
navigation.currentUrl = url;
const pendingIsCurrent =
navigation.pendingNavigation &&
(navigation.pendingNavigationDeadline === undefined ||
navigation.pendingNavigationDeadline >= Date.now());
if (causedByAction === true || (causedByAction === undefined && pendingIsCurrent)) {
navigation.pendingNavigation = false;
navigation.pendingNavigationDeadline = undefined;
const annotatedIndex = annotateLastStepNavigation(buffer, url);
if (annotatedIndex >= 0) return { kind: "annotated", index: annotatedIndex };
} else {
navigation.pendingNavigation = false;
navigation.pendingNavigationDeadline = undefined;
}
if (hasRedirectQualifier(transitionQualifiers)) {
return { kind: "coalesce_redirect", url };
}
buffer.steps.push({
op: "navigate",
url,
pageUrl: url,
transitionType,
transitionQualifiers,
});
return { kind: "appended", index: buffer.steps.length - 1 };
}
export function appendRecordedPayload(
buffer: RecordingStepBuffer,
payload: RecordStepPayload,
targetHint?: TargetMatchHint,
): number | null {
if (payload.op === "navigate") {
if (!payload.url) return null;
const result = observeRecordedNavigation(
buffer,
payload.url,
payload.navigation_caused_by_action,
payload.transitionType,
payload.transitionQualifiers,
);
return result.kind === "appended" ? result.index : null;
}
const step = toDraftStep(
{ ...payload, page_url: payload.page_url ?? buffer.navigation.currentUrl },
targetHint,
);
if (!step) return null;
buffer.steps.push(step);
if (step.op === "click" || step.op === "press" || step.op === "select" || step.op === "fill") {
buffer.navigation.pendingNavigation = payload.expects_navigation === true;
buffer.navigation.pendingNavigationDeadline = buffer.navigation.pendingNavigation
? Date.now() + NAVIGATION_TRIGGER_WINDOW_MS
: undefined;
}
return buffer.steps.length - 1;
}
@@ -0,0 +1,61 @@
import type { RecordingNavigationCursor } from "./step-buffer";
export interface TabActivation {
tabId: number;
revision: number;
}
export class RecordingTabCoordinator {
readonly #navigationByTab = new Map<number, RecordingNavigationCursor>();
#activeTabId: number;
#currentTabId: number;
#activationRevision = 0;
constructor(initialTabId: number, initialUrl?: string) {
this.#activeTabId = initialTabId;
this.#currentTabId = initialTabId;
this.#navigationByTab.set(initialTabId, {
currentUrl: initialUrl,
pendingNavigation: false,
});
}
get activeTabId(): number {
return this.#activeTabId;
}
get currentTabId(): number {
return this.#currentTabId;
}
noteActivation(tabId: number): TabActivation {
this.#activeTabId = tabId;
this.#activationRevision += 1;
return { tabId, revision: this.#activationRevision };
}
isLatest(activation: TabActivation): boolean {
return (
activation.tabId === this.#activeTabId && activation.revision === this.#activationRevision
);
}
commit(tabId: number, currentUrl?: string): void {
if (currentUrl !== undefined) this.navigation(tabId).currentUrl = currentUrl;
this.#currentTabId = tabId;
}
navigation(tabId: number, fallbackUrl?: string): RecordingNavigationCursor {
const existing = this.#navigationByTab.get(tabId);
if (existing) {
if (!existing.currentUrl && fallbackUrl) existing.currentUrl = fallbackUrl;
return existing;
}
const created: RecordingNavigationCursor = {
currentUrl: fallbackUrl,
pendingNavigation: false,
};
this.#navigationByTab.set(tabId, created);
return created;
}
}
@@ -0,0 +1,86 @@
import type { RenderedRef } from "@browser-skill/vom";
import type { TargetDescriptorV3 } from "@/transport/types";
import type { CaptureTargetDescriptor } from "../describe-target";
import type { IndexedObservationNode, RegisteredObservation } from "./observation-capture";
import type { TargetGeometry, TargetMatchHint } from "./types";
const MATCH_TOLERANCE_PX = 2;
function close(a: number, b: number): boolean {
return Math.abs(a - b) <= MATCH_TOLERANCE_PX;
}
function rectMatches(a: TargetGeometry["rect"], b: TargetGeometry["rect"]): boolean {
return close(a.x, b.x) && close(a.y, b.y) && close(a.w, b.w) && close(a.h, b.h);
}
function candidateMatches(
candidate: IndexedObservationNode,
geometry: TargetGeometry,
geometrySpace: "top" | "local",
): boolean {
const rect = geometrySpace === "local" ? candidate.geometry.localRect : candidate.geometry.rect;
return rect != null && rectMatches(geometry.rect, rect);
}
export function unmatchedTarget(fallback?: CaptureTargetDescriptor): TargetDescriptorV3 {
return {
...(fallback?.role ? { role: fallback.role } : {}),
...(fallback?.name ? { name: fallback.name } : {}),
unmatched: true,
};
}
function normalized(value: string | undefined): string {
return value?.replace(/\s+/g, " ").trim().toLowerCase() ?? "";
}
function matchesSemantics(ref: RenderedRef, fallback?: CaptureTargetDescriptor): boolean {
const role = normalized(fallback?.role);
const name = normalized(fallback?.name);
if (!role && !name) return false;
if (role && normalized(ref.role) !== role) return false;
if (name && normalized(ref.name) !== name) return false;
return true;
}
function descriptor(ref: RenderedRef): TargetDescriptorV3 {
return {
ref: ref.ref,
...(ref.role ? { role: ref.role } : {}),
...(ref.name ? { name: ref.name } : {}),
...(ref.ctx ? { ctx: ref.ctx } : {}),
};
}
export function matchObservationTarget(input: {
observation: RegisteredObservation;
hint?: TargetMatchHint;
fallback?: CaptureTargetDescriptor;
}): TargetDescriptorV3 {
if (input.hint?.frameId === null) return unmatchedTarget(input.fallback);
const frameId = input.hint?.frameId ?? input.observation.rootFrameId;
const geometry = input.hint?.geometry;
if (geometry) {
const matches = input.observation.index
.candidates(frameId, geometry.tag)
.filter(
(candidate) =>
candidate.ref &&
candidateMatches(candidate, geometry, input.hint?.geometrySpace ?? "top"),
);
if (matches.length === 1) return descriptor(matches[0]!.ref!);
const semanticMatches = matches.filter(
(candidate) => candidate.ref && matchesSemantics(candidate.ref, input.fallback),
);
if (semanticMatches.length === 1) return descriptor(semanticMatches[0]!.ref!);
return unmatchedTarget(input.fallback);
}
const semanticMatches = input.observation.index
.refs(frameId)
.filter((ref) => matchesSemantics(ref, input.fallback));
return semanticMatches.length === 1
? descriptor(semanticMatches[0]!)
: unmatchedTarget(input.fallback);
}
@@ -0,0 +1,92 @@
import {
type StepV3,
type StopReason,
TRACE_VERSION_V3,
type TraceStateV3,
type TraceV3,
VOM_FORMAT_VERSION,
} from "@/transport/types";
import { resolveDraftStartUrl } from "./draft-policy";
import type { RecordedStateEntry, RecordingStateRegistry } from "./state-registry";
import { reduceTraceStepsV3 } from "./trace-reducer-v3";
import { formatTraceStateBody } from "./trace-state-body";
import type { RecordingDraftStep, StepAnnotation } from "./types";
function publishedEntries(registry: RecordingStateRegistry, steps: StepV3[]): RecordedStateEntry[] {
const entries = registry.values();
if (steps.length === 0) return entries.slice(0, 1);
const referenced = new Set(steps.flatMap((step) => [step.state, step.result.state]));
return entries.filter((entry) => referenced.has(entry.id));
}
function remapDraftIds(draftIds: number[], stepIdByDraftId: Map<number, number>): number[] {
return [
...new Set(
draftIds.flatMap((id) => {
const stepId = stepIdByDraftId.get(id);
return stepId === undefined ? [] : [stepId];
}),
),
].sort((a, b) => a - b);
}
export function buildTraceV3(input: {
registry: RecordingStateRegistry;
drafts: RecordingDraftStep[];
annotations?: StepAnnotation[];
startedAt: string;
purpose?: string;
startUrl?: string;
stoppedBy: StopReason;
bskVersion: string;
redactValues?: boolean;
includeTabSwitches?: boolean;
}): TraceV3 {
const reduced = reduceTraceStepsV3(input.drafts, {
includeTabSwitches: input.includeTabSwitches,
redactValues: input.redactValues,
});
const entries = publishedEntries(input.registry, reduced.steps);
const publishedId = new Map(entries.map((entry, index) => [entry.id, `s${index + 1}`]));
const annotationsByState = new Map<string, StepAnnotation[]>();
for (const annotation of input.annotations ?? []) {
const bucket = annotationsByState.get(annotation.stateId) ?? [];
bucket.push(annotation);
annotationsByState.set(annotation.stateId, bucket);
}
const steps = reduced.steps.map((step) => ({
...step,
state: publishedId.get(step.state) ?? step.state,
result: { state: publishedId.get(step.result.state) ?? step.result.state },
}));
const states: TraceStateV3[] = entries.map((entry) => {
const id = publishedId.get(entry.id) ?? entry.id;
return {
id,
url: entry.url,
...(entry.title ? { title: entry.title } : {}),
body: formatTraceStateBody({
stateId: id,
url: entry.url,
title: entry.title,
stepIds: remapDraftIds(entry.stepsHere, reduced.stepIdByDraftId),
vomText: entry.vomText,
annotations: annotationsByState.get(entry.id) ?? [],
stepIdByDraftId: reduced.stepIdByDraftId,
}),
...(entry.truncated ? { truncated: true } : {}),
};
});
return {
version: TRACE_VERSION_V3,
...(input.purpose ? { purpose: input.purpose } : {}),
recorded_at: new Date().toISOString(),
started_at: input.startedAt,
stopped_by: input.stoppedBy,
entry: { start_url: resolveDraftStartUrl(input.drafts, input.startUrl, states[0]?.url) },
recorder: { bsk: input.bskVersion, vom: VOM_FORMAT_VERSION },
states,
steps,
};
}
@@ -1,30 +1,10 @@
import type { DraftTraceStep, PageRefV2, SelectedOptionV2, StepV2 } from "@/transport/types";
const CLIPBOARD_KEYS = new Set(["a", "c", "v", "x", "A", "C", "V", "X"]);
const MODIFIER_ONLY_KEYS = new Set(["Meta", "Control", "Alt", "Shift", "OS", "Hyper", "Super"]);
export function shouldRecordPress(
key: string,
modifiers?: Array<"alt" | "ctrl" | "meta" | "shift">,
): boolean {
if (MODIFIER_ONLY_KEYS.has(key)) return false;
const mods = modifiers ?? [];
const hasCtrlOrMeta = mods.includes("ctrl") || mods.includes("meta");
if (hasCtrlOrMeta && CLIPBOARD_KEYS.has(key)) return false;
if (key === "Enter" || key === "Escape") return true;
// Drop bare character typing — FillSession already records the value.
if (key.length === 1 && !hasCtrlOrMeta && !mods.includes("alt")) return false;
return false;
}
function shouldIncludeDraft(step: DraftTraceStep): boolean {
if (step.op === "press" && !shouldRecordPress(step.key, step.modifiers)) return false;
return true;
}
import type { PageRefV2, SelectedOptionV2, StepV2, TraceV2 } from "@/transport/types";
import { resolveDraftStartUrl, shouldIncludeDraft } from "./draft-policy";
import type { RecordingDraftStep } from "./types";
/** Collapse consecutive navigations to the last hop. */
function collapseNavigations(steps: DraftTraceStep[]): DraftTraceStep[] {
const out: DraftTraceStep[] = [];
function collapseNavigations(steps: RecordingDraftStep[]): RecordingDraftStep[] {
const out: RecordingDraftStep[] = [];
for (const step of steps) {
const prev = out[out.length - 1];
if (step.op === "navigate" && prev?.op === "navigate") {
@@ -36,7 +16,7 @@ function collapseNavigations(steps: DraftTraceStep[]): DraftTraceStep[] {
return out;
}
function collectUrls(steps: DraftTraceStep[], startUrl?: string): string[] {
function collectUrls(steps: RecordingDraftStep[], startUrl?: string): string[] {
const urls: string[] = [];
if (startUrl) urls.push(startUrl);
for (const step of steps) {
@@ -44,8 +24,8 @@ function collectUrls(steps: DraftTraceStep[], startUrl?: string): string[] {
urls.push(step.url);
continue;
}
if ("page_url" in step && step.page_url) urls.push(step.page_url);
if ("navigated_to" in step && step.navigated_to) urls.push(step.navigated_to);
if (step.pageUrl) urls.push(step.pageUrl);
if ("navigatedTo" in step && step.navigatedTo) urls.push(step.navigatedTo);
}
const seen = new Set<string>();
const unique: string[] = [];
@@ -58,7 +38,7 @@ function collectUrls(steps: DraftTraceStep[], startUrl?: string): string[] {
}
function buildPageRegistry(
steps: DraftTraceStep[],
steps: RecordingDraftStep[],
startUrl?: string,
): { pages: PageRefV2[]; urlToId: Map<string, string> } {
const urls = collectUrls(steps, startUrl);
@@ -81,9 +61,9 @@ function pageIdFor(
return urlToId.values().next().value ?? "p1";
}
function pageUrlForDraft(step: DraftTraceStep, fallbackUrl?: string): string | undefined {
if (step.op === "navigate") return step.page_url ?? step.url;
if ("page_url" in step && step.page_url) return step.page_url;
function pageUrlForDraft(step: RecordingDraftStep, fallbackUrl?: string): string | undefined {
if (step.op === "navigate") return step.pageUrl ?? step.url;
if (step.pageUrl) return step.pageUrl;
return fallbackUrl;
}
@@ -110,7 +90,7 @@ function toSelection(values: string[], labels?: string[]): SelectedOptionV2[] {
}
function toV2Step(
step: DraftTraceStep,
step: RecordingDraftStep,
id: number,
urlToId: Map<string, string>,
fallbackUrl?: string,
@@ -129,28 +109,23 @@ function toV2Step(
to: step.url,
};
case "click":
if (!step.captureTarget) return null;
return withEffect(
{
op: "click",
id,
page,
target: step.target,
target: step.captureTarget,
},
effectForNavigation(step.navigated_to, urlToId),
effectForNavigation(step.navigatedTo, urlToId),
);
case "hover":
return {
op: "hover",
id,
page,
target: step.target,
};
case "fill":
if (!step.captureTarget) return null;
return {
op: "fill",
id,
page,
target: step.target,
target: step.captureTarget,
value: step.value,
...(step.redacted ? { redacted: true } : {}),
};
@@ -161,26 +136,33 @@ function toV2Step(
id,
page,
key: step.key,
...(step.target ? { target: step.target } : {}),
...(step.captureTarget ? { target: step.captureTarget } : {}),
...(step.modifiers?.length ? { modifiers: step.modifiers } : {}),
},
effectForNavigation(step.navigated_to, urlToId),
effectForNavigation(step.navigatedTo, urlToId),
);
case "select":
if (!step.captureTarget) return null;
return withEffect(
{
op: "select",
id,
page,
target: step.target,
target: step.captureTarget,
selection: toSelection(step.values, step.labels),
},
effectForNavigation(step.navigated_to, urlToId),
effectForNavigation(step.navigatedTo, urlToId),
);
// `hover` only exists in Trace v3. Peers that negotiate v2 predate the
// step variant and fail to decode the whole result if we emit it.
case "hover":
case "scroll":
case "switch_tab":
return null;
}
}
export interface ReducedTrace {
interface ReducedTrace {
pages: PageRefV2[];
steps: StepV2[];
}
@@ -189,7 +171,7 @@ export interface ReducedTrace {
* Compile capture drafts into record-only trace v2 steps.
* Variable inputs are NOT classified here — executing agents infer that at run time.
*/
export function reduceTraceSteps(steps: DraftTraceStep[], startUrl?: string): ReducedTrace {
function reduceTraceSteps(steps: RecordingDraftStep[], startUrl?: string): ReducedTrace {
const collapsed = collapseNavigations(steps);
const { pages, urlToId } = buildPageRegistry(collapsed, startUrl);
const out: StepV2[] = [];
@@ -197,8 +179,8 @@ export function reduceTraceSteps(steps: DraftTraceStep[], startUrl?: string): Re
let lastUrl = startUrl;
for (const draft of collapsed) {
if (draft.op === "navigate") lastUrl = draft.url;
else if ("navigated_to" in draft && draft.navigated_to) lastUrl = draft.navigated_to;
else if ("page_url" in draft && draft.page_url) lastUrl = draft.page_url;
else if ("navigatedTo" in draft && draft.navigatedTo) lastUrl = draft.navigatedTo;
else if (draft.pageUrl) lastUrl = draft.pageUrl;
const step = toV2Step(draft, id, urlToId, lastUrl);
if (!step) continue;
out.push(step);
@@ -207,18 +189,27 @@ export function reduceTraceSteps(steps: DraftTraceStep[], startUrl?: string): Re
return { pages, steps: out };
}
export function resolveTraceStartUrl(
drafts: DraftTraceStep[],
function resolveTraceStartUrl(
drafts: RecordingDraftStep[],
startUrl?: string,
pages?: PageRefV2[],
): string {
if (startUrl) return startUrl;
const navigate = drafts.find((step): step is Extract<DraftTraceStep, { op: "navigate" }> => {
return step.op === "navigate";
});
if (navigate) return navigate.url;
for (const draft of drafts) {
if ("page_url" in draft && draft.page_url) return draft.page_url;
}
return pages?.[0]?.url ?? "about:blank";
return resolveDraftStartUrl(drafts, startUrl, pages?.[0]?.url);
}
export function buildTraceV2(input: {
steps: RecordingDraftStep[];
startedAt: string;
startUrl?: string;
purpose?: string;
}): TraceV2 {
const { pages, steps } = reduceTraceSteps(input.steps, input.startUrl);
return {
recorded_at: new Date().toISOString(),
started_at: input.startedAt,
...(input.purpose ? { purpose: input.purpose } : {}),
entry: { start_url: resolveTraceStartUrl(input.steps, input.startUrl, pages) },
pages,
steps,
};
}
@@ -0,0 +1,154 @@
import type { NavigationCause, StepV3 } from "@/transport/types";
import { shouldIncludeDraft } from "./draft-policy";
import { hasRedirectQualifier } from "./navigation-policy";
import { unmatchedTarget } from "./target-matcher";
import type { RecordingDraftStep } from "./types";
interface CollapsedDraft {
draft: RecordingDraftStep;
draftIds: number[];
}
const TRANSITION_CAUSES: Record<string, NavigationCause> = {
typed: "user_typed",
generated: "user_typed",
keyword: "user_typed",
keyword_generated: "user_typed",
link: "link",
form_submit: "form_submit",
reload: "reload",
auto_bookmark: "browser",
start_page: "browser",
};
function isRedirect(step: Extract<RecordingDraftStep, { op: "navigate" }>): boolean {
return hasRedirectQualifier(step.transitionQualifiers);
}
function collapseRedirects(steps: RecordingDraftStep[]): CollapsedDraft[] {
const output: CollapsedDraft[] = [];
steps.forEach((step, index) => {
const previous = output[output.length - 1];
if (step.op === "navigate" && previous?.draft.op === "navigate" && isRedirect(step)) {
previous.draft = {
...previous.draft,
url: step.url,
postStateId: step.postStateId ?? previous.draft.postStateId,
};
previous.draftIds.push(index + 1);
return;
}
output.push({ draft: { ...step }, draftIds: [index + 1] });
});
return output;
}
function navigationCause(step: Extract<RecordingDraftStep, { op: "navigate" }>): NavigationCause {
if (step.cause) return step.cause;
const qualifiers = step.transitionQualifiers ?? [];
if (qualifiers.includes("forward_back")) return "history";
if (qualifiers.includes("from_address_bar")) return "user_typed";
return TRANSITION_CAUSES[step.transitionType ?? ""] ?? "browser";
}
function selection(values: string[], labels?: string[]): Array<{ value: string; label?: string }> {
return values.map((value, index) => ({
value,
...(labels?.[index] ? { label: labels[index] } : {}),
}));
}
interface ReduceDraftOptions {
includeTabSwitches: boolean;
redactValues: boolean;
}
function reduceDraft(
draft: RecordingDraftStep,
id: number,
options: ReduceDraftOptions,
): StepV3 | null {
if (!shouldIncludeDraft(draft)) return null;
if (draft.op === "switch_tab") {
if (!options.includeTabSwitches || !draft.preStateId || !draft.postStateId) return null;
return {
op: "switch_tab",
id,
state: draft.preStateId,
result: { state: draft.postStateId },
};
}
const state = draft.preStateId ?? draft.postStateId;
const resultState = draft.postStateId ?? draft.preStateId;
if (!state || !resultState) return null;
const common = { id, state, result: { state: resultState } };
switch (draft.op) {
case "navigate":
return { op: "navigate", ...common, to: draft.url, cause: navigationCause(draft) };
case "click":
return {
op: "click",
...common,
target: draft.matchedTarget ?? unmatchedTarget(draft.captureTarget),
};
case "hover":
return {
op: "hover",
...common,
target: draft.matchedTarget ?? unmatchedTarget(draft.captureTarget),
};
case "fill":
const fillIsRedacted = options.redactValues || draft.redacted === true;
return {
op: "fill",
...common,
target: draft.matchedTarget ?? unmatchedTarget(draft.captureTarget),
value: fillIsRedacted ? "***" : draft.value,
commit: draft.commit ?? "blur",
...(fillIsRedacted ? { redacted: true } : {}),
};
case "press":
return {
op: "press",
...common,
key: draft.key,
...(draft.captureTarget || draft.matchedTarget
? { target: draft.matchedTarget ?? unmatchedTarget(draft.captureTarget) }
: {}),
...(draft.modifiers?.length ? { modifiers: draft.modifiers } : {}),
};
case "select":
return {
op: "select",
...common,
target: draft.matchedTarget ?? unmatchedTarget(draft.captureTarget),
...(!options.redactValues ? { selection: selection(draft.values, draft.labels) } : {}),
};
case "scroll":
return { op: "scroll", ...common };
}
}
export interface ReducedTraceV3 {
steps: StepV3[];
stepIdByDraftId: Map<number, number>;
}
export function reduceTraceStepsV3(
steps: RecordingDraftStep[],
options: { includeTabSwitches?: boolean; redactValues?: boolean } = {},
): ReducedTraceV3 {
const output: StepV3[] = [];
const stepIdByDraftId = new Map<number, number>();
for (const { draft, draftIds } of collapseRedirects(steps)) {
const step = reduceDraft(draft, output.length + 1, {
includeTabSwitches: options.includeTabSwitches === true,
redactValues: options.redactValues === true,
});
if (!step) continue;
output.push(step);
for (const draftId of draftIds) stepIdByDraftId.set(draftId, step.id);
}
return { steps: output, stepIdByDraftId };
}
@@ -0,0 +1,39 @@
import type { StepAnnotation } from "./types";
function annotationText(annotation: StepAnnotation, stepId: number): string {
return ` ⟵ step ${stepId}: ${annotation.op}${annotation.detail ? `: ${annotation.detail}` : ""}`;
}
export function formatTraceStateBody(input: {
stateId: string;
url: string;
title?: string;
stepIds: number[];
vomText: string;
annotations: StepAnnotation[];
stepIdByDraftId: Map<number, number>;
}): string {
const lines = ["# bsk-observation 1", `state: ${JSON.stringify(input.stateId)}`];
lines.push(`url: ${JSON.stringify(input.url)}`);
if (input.title) lines.push(`title: ${JSON.stringify(input.title)}`);
if (input.stepIds.length > 0) lines.push(`steps_here: [${input.stepIds.join(", ")}]`);
lines.push("---");
const byLine = new Map<number, Array<{ annotation: StepAnnotation; stepId: number }>>();
for (const annotation of input.annotations) {
const stepId = input.stepIdByDraftId.get(annotation.draftId);
if (stepId === undefined) continue;
const bucket = byLine.get(annotation.line) ?? [];
bucket.push({ annotation, stepId });
byLine.set(annotation.line, bucket);
}
input.vomText.split("\n").forEach((bodyLine, lineIndex) => {
let line = bodyLine;
for (const item of byLine.get(lineIndex) ?? []) {
line += annotationText(item.annotation, item.stepId);
}
lines.push(line);
});
return `${lines.join("\n")}\n`;
}
+85
View File
@@ -0,0 +1,85 @@
import type { CaptureTargetDescriptor } from "@/lib/describe-target";
import type {
FillCommit,
KeyModifier,
NavigationCause,
StepV3,
TargetDescriptorV3,
} from "@/transport/types";
export interface TargetGeometry {
/** Top-level viewport-relative CSS pixels, as defined by the geometry module. */
rect: { x: number; y: number; w: number; h: number };
tag: string;
}
export interface TargetMatchHint {
geometry?: TargetGeometry;
/** Missing means top frame; null means the source Document could not be resolved. */
frameId?: string | null;
geometrySpace?: "top" | "local";
}
export interface StepAnnotation {
draftId: number;
op: StepV3["op"];
line: number;
stateId: string;
detail?: string;
}
interface DraftStateLink {
pageUrl?: string;
preStateId?: string;
postStateId?: string;
}
interface DraftTarget {
captureTarget?: CaptureTargetDescriptor;
targetHint?: TargetMatchHint;
matchedTarget?: TargetDescriptorV3;
}
interface DraftNavigationEffect {
navigatedTo?: string;
}
export type RecordingDraftStep =
| ({ op: "click" } & DraftStateLink & DraftTarget & DraftNavigationEffect)
| ({ op: "hover" } & DraftStateLink & DraftTarget)
| ({
op: "fill";
value: string;
commit?: FillCommit;
redacted?: boolean;
} & DraftStateLink &
DraftTarget &
DraftNavigationEffect)
| ({
op: "press";
key: string;
modifiers?: KeyModifier[];
} & DraftStateLink &
DraftTarget &
DraftNavigationEffect)
| ({
op: "select";
values: string[];
labels?: string[];
} & DraftStateLink &
DraftTarget &
DraftNavigationEffect)
| ({ op: "scroll" } & DraftStateLink)
| ({ op: "switch_tab" } & DraftStateLink)
| ({
op: "navigate";
url: string;
cause?: NavigationCause;
transitionType?: string;
transitionQualifiers?: string[];
} & DraftStateLink);
export type TargetedRecordingDraft = Extract<
RecordingDraftStep,
{ op: "click" | "hover" | "fill" | "press" | "select" }
>;
@@ -0,0 +1,22 @@
export const RECORD_DOCUMENT_ATTRIBUTE = "data-bsk-record-document";
const RECORDING_DOCUMENT_ID_PATTERN =
/^[0-9a-f]{8}-[0-9a-f]{4}-4[0-9a-f]{3}-[89ab][0-9a-f]{3}-[0-9a-f]{12}$/i;
const RECORDING_DOCUMENT_MARKER_PREFIX = "bsk:";
export function recordingDocumentMarkerValue(identity: string): string {
if (!RECORDING_DOCUMENT_ID_PATTERN.test(identity)) {
throw new TypeError("recording Document identity must be a random UUID");
}
return `${RECORDING_DOCUMENT_MARKER_PREFIX}${identity}`;
}
/** Read BrowserSkill's opaque per-Document recording identity from captured attributes. */
export function readRecordingDocumentIdentity(
attrs: Readonly<Record<string, string>>,
): string | undefined {
const marker = attrs[RECORD_DOCUMENT_ATTRIBUTE];
if (!marker?.startsWith(RECORDING_DOCUMENT_MARKER_PREFIX)) return undefined;
const identity = marker.slice(RECORDING_DOCUMENT_MARKER_PREFIX.length);
return identity && RECORDING_DOCUMENT_ID_PATTERN.test(identity) ? identity : undefined;
}
@@ -8,6 +8,7 @@ import type {
RequestFrame,
} from "@/transport/types";
import { ToolDispatcher } from "../dispatcher";
import { resetBrowserObservationForTests, setBrowserObservationAttachForTests } from "../record";
type TestDispatcherCdp = NonNullable<ConstructorParameters<typeof ToolDispatcher>[0]["cdp"]>;
@@ -44,9 +45,72 @@ function makeRequest(method: string, params: unknown): RequestFrame {
describe("ToolDispatcher", () => {
afterEach(() => {
resetBrowserObservationForTests();
vi.unstubAllGlobals();
});
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 () => {}),
},
});
await sessions.start("aa11");
const tab = {
id: 7,
windowId: 4242,
active: true,
status: "complete",
url: "https://example.com/start",
} as chrome.tabs.Tab;
const frameCoordinator = {
begin: vi.fn(),
armTab: vi.fn(async () => true),
sourceFor: vi.fn(() => null),
stop: vi.fn(async () => true),
cancel: vi.fn(),
};
const sendToTab = vi.fn(async () => ({ ok: true }));
setBrowserObservationAttachForTests(
() => () => {},
() => () => {},
);
const dispatcher = new ToolDispatcher({
transport,
sessions,
recording: {
tabsApi: {
get: vi.fn(async () => tab),
query: vi.fn(async () => [tab]),
},
frameCoordinator,
sendToTab,
},
});
dispatcher.start();
deliver(
makeRequest("tool.record_start", {
session_id: "aa11",
url: "https://example.com/start",
}),
);
await vi.waitFor(() => expect(sent).toHaveLength(1));
expect(sent[0]).toMatchObject({ result: { tab_id: 7, recording: true } });
expect(frameCoordinator.begin).toHaveBeenCalledOnce();
expect(frameCoordinator.armTab).toHaveBeenCalledWith(expect.any(String), 7);
deliver({ id: "r-2", method: "tool.record_stop", params: { session_id: "aa11" } });
await vi.waitFor(() => expect(sent).toHaveLength(2));
expect(frameCoordinator.stop).toHaveBeenCalledOnce();
expect(sent[1]).toMatchObject({ id: "r-2", result: { trace: { steps: [] } } });
});
it("routes tool.session_start to the SessionManager and replies with the window id", async () => {
const { transport, sent, deliver } = fakeTransport();
const sessions = new SessionManager({
@@ -1656,7 +1656,7 @@ describe("buildVomScene", () => {
const scene = buildVomScene(axNodes, captured);
expect(scene.nodes.find((n) => n.id === 20)).toEqual(
expect.objectContaining({ id: 20, role: "generic", cursor: "pointer" }),
expect.objectContaining({ id: 20, role: "button", name: "close", cursor: "pointer" }),
);
const rendered = renderVom(scene);
expect(rendered.text).toContain('@e1 button "close"');
@@ -1732,7 +1732,7 @@ describe("buildVomScene", () => {
};
const scene = buildVomScene(axNodes, captured);
expect(scene.nodes.find((n) => n.id === 20)?.role).toBe("generic");
expect(scene.nodes.find((n) => n.id === 20)).toBeUndefined();
expect(scene.nodes.find((n) => n.id === 30)?.role).toBe("link");
});
@@ -1801,9 +1801,9 @@ describe("buildVomScene", () => {
const scene = buildVomScene(axNodes, captured);
expect(scene.nodes.find((n) => n.id === 20)).toEqual(
expect.objectContaining({ id: 20, role: "generic", attrs: { "aria-label": "收藏" } }),
expect.objectContaining({ id: 20, role: "button", attrs: { "aria-label": "收藏" } }),
);
expect(scene.nodes.find((n) => n.id === 30)?.role).toBe("generic");
expect(scene.nodes.find((n) => n.id === 30)).toBeUndefined();
const rendered = renderVom(scene);
expect(rendered.text).toContain('@e1 button "收藏"');
expect(rendered.refs.map(({ ref, backendNodeId }) => ({ ref, backendNodeId }))).toEqual([
@@ -2131,7 +2131,7 @@ describe("buildVomScene", () => {
});
expect(scene.surfaces).toEqual([
{ triggerId: 21, triggerAction: "hover", subItems: ["My profile", "Sign out"] },
{ triggerId: 20, triggerAction: "hover", subItems: ["My profile", "Sign out"] },
]);
expect(renderVom(scene).text).toContain(
'@e1 button "image" [hover first: My profile | Sign out]',
@@ -3155,6 +3155,98 @@ describe("handleSnapshot", () => {
});
});
it("decorates live observations with active controlled content", async () => {
const strings = [
"html",
"body",
"button",
"div",
"#text",
"role",
"tab",
"aria-selected",
"true",
"aria-controls",
"reviews-panel",
"id",
"static",
"auto",
"visible",
"1",
"A detailed review",
];
const index = (value: string) => strings.indexOf(value);
const style = [index("static"), index("auto"), index("auto"), index("visible"), index("1")];
const snapshot = {
strings,
documents: [
{
nodes: {
parentIndex: [-1, 0, 1, 1, 3],
nodeName: [index("html"), index("body"), index("button"), index("div"), index("#text")],
backendNodeId: [10, 11, 12, 13, 14],
attributes: [
[],
[],
[
index("role"),
index("tab"),
index("aria-selected"),
index("true"),
index("aria-controls"),
index("reviews-panel"),
],
[index("id"), index("reviews-panel")],
[],
],
nodeValue: [-1, -1, -1, -1, index("A detailed review")],
},
layout: {
nodeIndex: [0, 1, 2, 3, 4],
styles: [style, style, style, style, style],
bounds: [
[0, 0, 1000, 800],
[0, 0, 1000, 800],
[20, 20, 120, 40],
[20, 80, 500, 200],
[20, 80, 200, 20],
],
paintOrders: [0, 0, 1, 1, 1],
},
},
],
};
const ax: CdpAxNode[] = [
{
nodeId: "root",
backendDOMNodeId: 11,
role: { type: "role", value: "RootWebArea" },
childIds: ["reviews"],
},
{
nodeId: "reviews",
parentId: "root",
backendDOMNodeId: 12,
role: { type: "role", value: "tab" },
name: { type: "computedString", value: "Reviews" },
properties: [
{ name: "selected", value: { value: true } },
{ name: "controls", value: { value: "reviews-panel" } },
],
},
];
const result = await captureVomObservation(
makeOverlayDeps(ax, snapshot, VP_METRICS).cdp,
4,
"https://example.com",
);
expect(result.text).toContain('@e1 tab "Reviews"');
expect(result.text).toContain("[§ active: Reviews]");
expect(result.text).toContain("A detailed review");
});
it("does not leak form secrets through the record-safe observation payload", async () => {
const secrets = {
email: "user@example.com",
File diff suppressed because it is too large Load Diff
+34 -37
View File
@@ -51,7 +51,12 @@ import {
handleScreenshot,
handleSnapshot,
} from "./observation";
import { handleRecordAwait, handleRecordStart, handleRecordStop } from "./record";
import {
handleRecordAwait,
handleRecordStart,
handleRecordStop,
type RecordRuntimeDeps,
} from "./record";
import {
handleSessionStart,
handleSessionStop,
@@ -99,6 +104,7 @@ export interface DispatcherDeps {
transport: Transport;
sessions: SessionManager;
cdp?: DispatcherCdpRunner;
recording?: RecordRuntimeDeps;
/**
* Invoked whenever a dispatched RPC may have changed the live
* session set (currently `tool.session_start` and
@@ -134,6 +140,7 @@ export class ToolDispatcher {
private readonly transport: Transport;
private readonly sessions: SessionManager;
private readonly cdp?: DispatcherCdpRunner;
private readonly recording?: RecordRuntimeDeps;
private readonly onSessionsChanged?: () => void;
private readonly onBrowserControlResumed?: (sessionId: string) => void;
private readonly approveBorrow?: BorrowConfirmationApprover;
@@ -153,6 +160,7 @@ export class ToolDispatcher {
this.transport = deps.transport;
this.sessions = deps.sessions;
this.cdp = deps.cdp;
this.recording = deps.recording;
this.onSessionsChanged = deps.onSessionsChanged;
this.onBrowserControlResumed = deps.onBrowserControlResumed;
this.approveBorrow = deps.approveBorrow;
@@ -531,44 +539,26 @@ export class ToolDispatcher {
signal,
});
case "tool.record_start":
return handleRecordStart(this.sessions, req.params as RecordStartParams, {
tabsApi: chromeTabsApi,
sendToTab: (tabId, msg) => chrome.tabs.sendMessage(tabId, msg),
bypassOverlay: async (tabId, enabled) => {
try {
await chrome.tabs.sendMessage(tabId, {
type: OVERLAY_AUTOMATION_BYPASS,
enabled,
});
} catch {
// Content script may be unavailable on restricted pages.
}
},
...(this.cdp ? { cdp: this.cdp } : {}),
signal,
});
return this.recording
? handleRecordStart(this.sessions, req.params as RecordStartParams, {
...this.recording,
signal,
})
: recordingRuntimeUnavailable();
case "tool.record_stop":
return handleRecordStop(this.sessions, req.params as RecordStopParams, {
tabsApi: chromeTabsApi,
sendToTab: (tabId, msg) => chrome.tabs.sendMessage(tabId, msg),
bypassOverlay: async (tabId, enabled) => {
try {
await chrome.tabs.sendMessage(tabId, {
type: OVERLAY_AUTOMATION_BYPASS,
enabled,
});
} catch {
// Content script may be unavailable on restricted pages.
}
},
signal,
});
return this.recording
? handleRecordStop(this.sessions, req.params as RecordStopParams, {
...this.recording,
signal,
})
: recordingRuntimeUnavailable();
case "tool.record_await":
return handleRecordAwait(this.sessions, req.params as RecordAwaitParams, {
tabsApi: chromeTabsApi,
sendToTab: (tabId, msg) => chrome.tabs.sendMessage(tabId, msg),
signal,
});
return this.recording
? handleRecordAwait(this.sessions, req.params as RecordAwaitParams, {
...this.recording,
signal,
})
: recordingRuntimeUnavailable();
default:
return {
code: "unknown_method",
@@ -704,6 +694,13 @@ function isRpcError(v: unknown): v is RpcError {
);
}
function recordingRuntimeUnavailable(): RpcError {
return {
code: "protocol_error",
message: "recording runtime is unavailable",
};
}
function sessionIdForBrowserControlMethod(req: RequestFrame): string | null {
switch (req.method) {
case "tool.tab_create":
File diff suppressed because it is too large Load Diff
+391 -74
View File
@@ -15,22 +15,37 @@ import {
type RecordQueryResponse,
type RecordStartAck,
type RecordStartMessage,
type RecordStepAck,
type RecordStopMessage,
} from "@/lib/record-bridge";
import { appendRecordedPayload, observeRecordedNavigation } from "@/lib/recording-step-buffer";
import { reduceTraceSteps, resolveTraceStartUrl } from "@/lib/trace-reducer";
import {
type RecordFrameCoordinator,
type RecordingCaptureScope,
recordFrameCoordinator,
} from "@/lib/recording/frame-coordinator";
import { RecordingObservationRuntime } from "@/lib/recording/recording-runtime";
import {
appendRecordedPayload,
observeRecordedNavigation,
type RecordingStepBuffer,
} from "@/lib/recording/step-buffer";
import { RecordingTabCoordinator, type TabActivation } from "@/lib/recording/tab-coordinator";
import { buildTraceV2 } from "@/lib/recording/trace-reducer-v2";
import type { RecordingDraftStep } from "@/lib/recording/types";
import type { SessionManager } from "@/session-manager/manager";
import { EXTENSION_VERSION } from "@/transport/handshake";
import type {
DraftTraceStep,
RecordAwaitParams,
RecordAwaitResult,
RecordedTrace,
RecordStartParams,
RecordStartResult,
RecordStopParams,
RecordStopResult,
RpcError,
TraceV2,
StopReason,
} from "@/transport/types";
import { TRACE_VERSION_V3 } from "@/transport/types";
import { handleNavigate } from "./navigation";
import {
type CdpRunner,
@@ -43,21 +58,46 @@ import {
interface ActiveRecording {
requestId: string;
tabId: number;
tabs: RecordingTabCoordinator;
agentWindowId: number;
startUrl?: string;
purpose?: string;
steps: DraftTraceStep[];
steps: RecordingDraftStep[];
startedAt: string;
startedAtMs: number;
finishPromise: Promise<TraceV2>;
resolveFinish: (trace: TraceV2) => void;
traceVersion: 2 | 3;
supportsTabSwitchSteps: boolean;
finishPromise: Promise<RecordedTrace>;
resolveFinish: (trace: RecordedTrace) => void;
rejectFinish: (err: Error) => void;
settled: boolean;
finishing: boolean;
currentUrl?: string;
pendingNavigation: boolean;
pendingNavigationDeadline?: number;
finishAttempt: Promise<RecordedTrace | null> | null;
observation: RecordingObservationRuntime | null;
stoppedBy: StopReason;
/** Navigation callbacks tracked from event receipt through action enqueue. */
navigationCallbacks: Set<Promise<void>>;
/** Synchronous intake gate closed only after finish drains to stability. */
acceptingNavigation: boolean;
/**
* Serializes step appends with navigation observation so a click is always
* in `steps` before a same-turn `webNavigation` tries to annotate it.
*/
actionQueue: Promise<void>;
/** Last accepted sequence for each content-script document producer. */
lastStepSequenceByProducer: Map<string, number>;
}
function enqueueRecordingAction(
recording: ActiveRecording,
task: () => Promise<void>,
): Promise<void> {
const queued = recording.actionQueue.then(task, task);
recording.actionQueue = queued.catch(() => {});
return queued;
}
function isRecordingFinishing(recording: ActiveRecording): boolean {
return recording.settled || recording.finishAttempt !== null;
}
const recordings = new Map<string, ActiveRecording>();
@@ -78,6 +118,9 @@ function sleep(ms: number): Promise<void> {
return new Promise((resolve) => setTimeout(resolve, ms));
}
/** Recording producer version mirrored into trace.recorder.bsk. */
export const BSK_TRACE_VERSION = EXTENSION_VERSION;
/** Injectable http(s) landing page when `tool.record_start` omits `url`. */
export const RECORD_DEFAULT_START_URL = "https://example.com/";
@@ -137,9 +180,11 @@ async function sendRecordStartWithAck(
tabId: number,
msg: RecordStartMessage,
sendToTab: RecordDeps["sendToTab"],
cancelled?: () => boolean,
): Promise<void> {
let lastError: unknown;
for (let attempt = 0; attempt < RECORD_START_RETRIES; attempt += 1) {
if (cancelled?.()) throw new Error("record start cancelled");
try {
const response = await sendToTab(tabId, msg);
if (isRecordStartAck(response)) return;
@@ -147,6 +192,7 @@ async function sendRecordStartWithAck(
} catch (err) {
lastError = err;
}
if (cancelled?.()) throw new Error("record start cancelled");
if (attempt + 1 < RECORD_START_RETRIES) {
await sleep(RECORD_START_RETRY_DELAY_MS);
}
@@ -154,19 +200,83 @@ async function sendRecordStartWithAck(
throw lastError ?? new Error("failed to start recording in content script");
}
function buildTrace(recording: ActiveRecording): TraceV2 {
const { pages, steps } = reduceTraceSteps(recording.steps, recording.startUrl);
const startUrl = resolveTraceStartUrl(recording.steps, recording.startUrl, pages);
function negotiatedTraceVersion(params: RecordStartParams): 2 | 3 | RpcError {
if (params.trace_version === undefined) return 2;
if (params.trace_version === TRACE_VERSION_V3) return 3;
return {
...(recording.purpose ? { purpose: recording.purpose } : {}),
recorded_at: new Date().toISOString(),
started_at: recording.startedAt,
entry: { start_url: startUrl },
pages,
steps,
code: "invalid_params",
message: `unsupported trace_version ${params.trace_version}; supported values are omitted (v2) or ${TRACE_VERSION_V3} (v3)`,
};
}
function buildTrace(recording: ActiveRecording): RecordedTrace {
if (recording.traceVersion === 3 && recording.observation) {
return recording.observation.buildTrace({
drafts: recording.steps,
startedAt: recording.startedAt,
purpose: recording.purpose,
startUrl: recording.startUrl,
stoppedBy: recording.stoppedBy,
bskVersion: BSK_TRACE_VERSION,
includeTabSwitches: recording.supportsTabSwitchSteps,
});
}
if (recording.traceVersion === 3) {
throw new Error("trace v3 observation runtime is unavailable");
}
return buildTraceV2({
steps: recording.steps,
startedAt: recording.startedAt,
...(recording.startUrl ? { startUrl: recording.startUrl } : {}),
...(recording.purpose ? { purpose: recording.purpose } : {}),
});
}
function stepBufferFor(
recording: ActiveRecording,
tabId: number,
fallbackUrl?: string,
): RecordingStepBuffer {
return {
steps: recording.steps,
navigation: recording.tabs.navigation(tabId, fallbackUrl),
};
}
async function activateRecordingTab(
recording: ActiveRecording,
targetTabId: number,
): Promise<void> {
const previousTabId = recording.tabs.currentTabId;
if (targetTabId === previousTabId) return;
const transition = recording.observation
? await recording.observation.captureTabTransition(previousTabId, targetTabId)
: {};
const { targetUrl, ...stateLinks } = transition;
const draft: Extract<RecordingDraftStep, { op: "switch_tab" }> = {
op: "switch_tab",
...stateLinks,
};
recording.steps.push(draft);
recording.observation?.bindTabTransition(draft, recording.steps.length);
recording.tabs.commit(targetTabId, targetUrl);
}
async function processRecordedStep(
recording: ActiveRecording,
draftIndex: number,
tabId: number,
producerId?: string,
): Promise<void> {
if (!recording.observation) return;
try {
await recording.observation.processDraft(tabId, recording.steps, draftIndex, producerId);
} catch (err) {
console.warn(`[bsk record] observation failed for step ${draftIndex + 1}`, err);
}
}
export interface RecordDeps {
tabsApi: ChromeTabsApi;
sendToTab(
@@ -174,16 +284,25 @@ export interface RecordDeps {
msg: RecordStartMessage | RecordStopMessage | RecordCancelMessage,
): Promise<unknown>;
bypassOverlay?: (tabId: number, enabled: boolean) => Promise<void>;
frameCoordinator?: Pick<
RecordFrameCoordinator,
"begin" | "armTab" | "sourceFor" | "stop" | "cancel"
>;
cdp?: CdpRunner;
signal?: AbortSignal;
}
export type RecordRuntimeDeps = Omit<RecordDeps, "frameCoordinator" | "signal"> & {
frameCoordinator: NonNullable<RecordDeps["frameCoordinator"]>;
};
let defaultDeps: RecordDeps | null = null;
function getDefaultDeps(): RecordDeps {
if (!defaultDeps) {
defaultDeps = {
tabsApi: chromeTabsApi,
sendToTab: (tabId, msg) => chrome.tabs.sendMessage(tabId, msg),
frameCoordinator: recordFrameCoordinator,
};
}
return defaultDeps;
@@ -241,16 +360,58 @@ export function releaseBrowserObservationListenersIfIdle(): void {
detachBrowserObservation = null;
}
export function attachRecordStepListener(): () => void {
export function attachRecordStepListener(deps: RecordDeps = getDefaultDeps()): () => void {
const listener = (
message: unknown,
_sender: chrome.runtime.MessageSender,
_sendResponse: () => void,
sender: chrome.runtime.MessageSender,
sendResponse: (response: RecordStepAck) => void,
) => {
if (!isRecordStepMessage(message)) return false;
for (const recording of recordings.values()) {
if (recording.requestId !== message.requestId) continue;
appendRecordedPayload(recording, message.step);
const source: RecordingCaptureScope | null | undefined = deps.frameCoordinator
? deps.frameCoordinator.sourceFor(message.requestId, message.producerId, sender)
: undefined;
if (deps.frameCoordinator && !source) return false;
const sourceTabId = source?.tabId ?? sender.tab?.id ?? recording.tabs.currentTabId;
const sourceWasActive = sender.tab?.active ?? sourceTabId === recording.tabs.activeTabId;
const producerKey = source
? `${source.tabId}:${source.documentId}:${source.producerId}`
: `${sourceTabId}:${message.producerId}`;
const expectedSequence = (recording.lastStepSequenceByProducer.get(producerKey) ?? 0) + 1;
if (message.sequence < expectedSequence) {
sendResponse({ ok: true, sequence: message.sequence });
return false;
}
if (message.sequence > expectedSequence) {
sendResponse({
ok: false,
expectedSequence,
error: `out-of-order recorded step ${message.sequence}`,
});
return false;
}
recording.lastStepSequenceByProducer.set(producerKey, message.sequence);
enqueueRecordingAction(recording, async () => {
if (sourceTabId !== recording.tabs.currentTabId) {
if (!sourceWasActive) return;
if (sourceTabId !== recording.tabs.activeTabId)
recording.tabs.noteActivation(sourceTabId);
await activateRecordingTab(recording, sourceTabId);
}
await recording.observation?.flushRedirects(sourceTabId);
const targetHint = message.step.geometry ? { geometry: message.step.geometry } : undefined;
const draftIndex = appendRecordedPayload(
stepBufferFor(recording, sourceTabId, message.step.page_url),
message.step,
targetHint,
);
if (draftIndex !== null) {
await processRecordedStep(recording, draftIndex, sourceTabId, source?.producerId);
}
});
sendResponse({ ok: true, sequence: message.sequence });
return false;
}
return false;
@@ -277,7 +438,7 @@ export function attachRecordFinishListener(deps: RecordDeps = getDefaultDeps()):
function findRecordingByTabId(tabId: number): ActiveRecording | null {
for (const recording of recordings.values()) {
if (recording.tabId === tabId && !recording.settled) return recording;
if (recording.tabs.currentTabId === tabId && !recording.settled) return recording;
}
return null;
}
@@ -314,7 +475,7 @@ async function clearRearmTimersForRecording(
recording: ActiveRecording,
deps: RecordDeps,
): Promise<void> {
clearRearmTimer(recording.tabId);
clearRearmTimer(recording.tabs.currentTabId);
try {
const tabs = await deps.tabsApi.query({ windowId: recording.agentWindowId });
for (const tab of tabs) {
@@ -329,13 +490,16 @@ async function stopRecordingOnAllAgentTabs(
recording: ActiveRecording,
deps: RecordDeps,
): Promise<void> {
if (deps.frameCoordinator && !(await deps.frameCoordinator.stop(recording.requestId))) {
throw new Error("failed to flush one or more recording documents");
}
const stopMsg: RecordStopMessage = { type: RECORD_STOP, requestId: recording.requestId };
let tabIds = [recording.tabId];
let tabIds = [recording.tabs.currentTabId];
try {
const tabs = await deps.tabsApi.query({ windowId: recording.agentWindowId });
tabIds = [
...new Set([
recording.tabId,
recording.tabs.currentTabId,
...tabs.flatMap((tab) => (typeof tab.id === "number" ? [tab.id] : [])),
]),
];
@@ -346,11 +510,11 @@ async function stopRecordingOnAllAgentTabs(
for (const tabId of tabIds) {
try {
const response = await deps.sendToTab(tabId, stopMsg);
if (tabId === recording.tabId && !isRecordStartAck(response)) {
if (tabId === recording.tabs.currentTabId && !isRecordStartAck(response)) {
throw new Error("content script did not confirm recorded steps");
}
} catch {
if (tabId === recording.tabId) {
if (tabId === recording.tabs.currentTabId) {
throw new Error("failed to flush recorded steps");
}
}
@@ -368,22 +532,39 @@ async function rearmRecording(
recording: ActiveRecording,
targetTabId: number,
deps: RecordDeps,
activation?: TabActivation,
): Promise<boolean> {
// Do NOT toggle automation-bypass here: each retry used to increment the
// content-script counter, and a single stop decrement left the ControlOverlay
// stuck with pointer-events:none (page usable, Interrupt dead). RecordOverlay
// already hides the control chrome while activeRecord is set.
const isFinishing = () => isRecordingFinishing(recording);
for (let attempt = 0; attempt < RECORD_REARM_MAX_ATTEMPTS; attempt += 1) {
if (isFinishing()) return false;
const startMsg: RecordStartMessage = {
type: RECORD_START,
requestId: recording.requestId,
startedAtMs: recording.startedAtMs,
};
try {
await sendRecordStartWithAck(targetTabId, startMsg, deps.sendToTab);
recording.tabId = targetTabId;
if (deps.frameCoordinator) {
const frameStarted = await deps.frameCoordinator.armTab(recording.requestId, targetTabId);
if (!frameStarted) throw new Error("recording document did not start");
}
await sendRecordStartWithAck(targetTabId, startMsg, deps.sendToTab, isFinishing);
if (isFinishing()) return false;
if (activation) {
if (!recording.tabs.isLatest(activation)) return true;
await enqueueRecordingAction(recording, async () => {
if (isFinishing() || !recording.tabs.isLatest(activation)) {
return;
}
await activateRecordingTab(recording, targetTabId);
});
}
return true;
} catch {
if (isFinishing()) return false;
if (attempt + 1 < RECORD_REARM_MAX_ATTEMPTS) {
await sleep(RECORD_REARM_RETRY_DELAY_MS);
}
@@ -392,7 +573,7 @@ async function rearmRecording(
return false;
}
function scheduleRearmForTab(tabId: number, deps: RecordDeps): void {
function scheduleRearmForTab(tabId: number, deps: RecordDeps, activation?: TabActivation): void {
const existing = rearmTimers.get(tabId);
if (existing) clearTimeout(existing);
rearmTimers.set(
@@ -401,7 +582,7 @@ function scheduleRearmForTab(tabId: number, deps: RecordDeps): void {
rearmTimers.delete(tabId);
void (async () => {
const current = await findRecordingForTab(tabId, deps);
if (current) await rearmRecording(current, tabId, deps);
if (current) await rearmRecording(current, tabId, deps, activation);
})();
}, RECORD_REARM_DEBOUNCE_MS),
);
@@ -413,14 +594,20 @@ export function attachRecordTabListener(deps: RecordDeps = getDefaultDeps()): ()
const windowId = tab.windowId;
if (tabId === undefined || windowId === undefined) return;
for (const recording of recordings.values()) {
if (recording.settled || recording.agentWindowId !== windowId) continue;
if (isRecordingFinishing(recording) || recording.agentWindowId !== windowId) continue;
scheduleRearmForTab(tabId, deps);
return;
}
};
const onActivated = (activeInfo: chrome.tabs.TabActiveInfo) => {
scheduleRearmForTab(activeInfo.tabId, deps);
for (const recording of recordings.values()) {
if (isRecordingFinishing(recording) || recording.agentWindowId !== activeInfo.windowId)
continue;
const activation = recording.tabs.noteActivation(activeInfo.tabId);
scheduleRearmForTab(activeInfo.tabId, deps, activation);
return;
}
};
chrome.tabs.onCreated.addListener(onCreated);
@@ -432,13 +619,47 @@ export function attachRecordTabListener(deps: RecordDeps = getDefaultDeps()): ()
}
export function attachRecordNavigationListener(deps: RecordDeps = getDefaultDeps()): () => void {
const observeMainFrame = (tabId: number, url?: string, causedByAction?: boolean) => {
void (async () => {
const recording = await findRecordingForTab(tabId, deps);
if (recording && url) {
observeRecordedNavigation(recording, url, causedByAction);
}
})();
const observeMainFrame = (
tabId: number,
url?: string,
causedByAction?: boolean,
transitionType?: string,
transitionQualifiers?: string[],
) => {
if (!url) return;
const candidates = [...recordings.values()].filter(
(recording) =>
!recording.settled && recording.acceptingNavigation && recording.tabs.activeTabId === tabId,
);
for (const recording of candidates) {
const queued = enqueueRecordingAction(recording, async () => {
if (tabId !== recording.tabs.currentTabId) {
await activateRecordingTab(recording, tabId);
}
const result = observeRecordedNavigation(
stepBufferFor(recording, tabId, url),
url,
causedByAction,
transitionType,
transitionQualifiers,
);
if (result.kind === "coalesce_redirect") {
recording.observation?.scheduleRedirect(tabId, recording.steps, result.url);
return;
}
if (result.kind === "noop") return;
recording.observation?.clearRedirect(tabId);
if (result.kind === "appended") {
await processRecordedStep(recording, result.index, tabId);
} else {
recording.observation?.scheduleSettle(tabId, recording.steps, result.index);
}
});
const tracked = queued.catch(() => {});
recording.navigationCallbacks.add(tracked);
void tracked.finally(() => recording.navigationCallbacks.delete(tracked));
}
};
const onMainFrameComplete = (tabId: number, url?: string) => {
void (async () => {
@@ -458,18 +679,13 @@ export function attachRecordNavigationListener(deps: RecordDeps = getDefaultDeps
details: chrome.webNavigation.WebNavigationTransitionCallbackDetails,
) => {
if (details.frameId !== 0) return;
const causedByAction =
details.transitionType === "link" || details.transitionType === "form_submit"
? true
: details.transitionType === "typed" ||
details.transitionType === "auto_bookmark" ||
details.transitionType === "generated" ||
details.transitionType === "keyword" ||
details.transitionType === "keyword_generated" ||
details.transitionType === "reload"
? false
: undefined;
observeMainFrame(details.tabId, details.url, causedByAction);
observeMainFrame(
details.tabId,
details.url,
undefined,
details.transitionType,
details.transitionQualifiers,
);
};
chrome.webNavigation.onCompleted.addListener(completedListener);
chrome.webNavigation.onCommitted?.addListener(committedListener);
@@ -487,6 +703,26 @@ export function attachRecordNavigationListener(deps: RecordDeps = getDefaultDeps
return () => chrome.tabs.onUpdated.removeListener(listener);
}
const MAX_FINISH_DRAIN_ROUNDS = 10;
async function drainRecordingToStability(recording: ActiveRecording): Promise<boolean> {
for (let round = 0; round < MAX_FINISH_DRAIN_ROUNDS; round += 1) {
await Promise.all([...recording.navigationCallbacks]);
const actionTail = recording.actionQueue;
await actionTail;
await recording.observation?.flush();
if (recording.navigationCallbacks.size === 0 && recording.actionQueue === actionTail) {
// No event or promise continuation can interleave with this synchronous
// check-and-close, so work accepted before the cutoff is fully drained.
recording.acceptingNavigation = false;
return true;
}
}
console.warn("[bsk record] navigation/action queues did not stabilize at stop");
return false;
}
export function attachRecordQueryListener(deps: RecordDeps = getDefaultDeps()): () => void {
const listener = (
message: unknown,
@@ -505,7 +741,8 @@ export function attachRecordQueryListener(deps: RecordDeps = getDefaultDeps()):
sendResponse({ active: false });
return;
}
await rearmRecording(recording, tabId, deps);
const activation = sender.tab?.active ? recording.tabs.noteActivation(tabId) : undefined;
await rearmRecording(recording, tabId, deps, activation);
sendResponse({
active: true,
requestId: recording.requestId,
@@ -527,23 +764,50 @@ async function finishRecordingByRequest(
if (recording.requestId !== requestId || recording.settled) continue;
const match = await findRecordingForTab(tabId, deps);
if (match !== recording) continue;
await finishRecording(sessionId, deps);
await finishRecording(sessionId, deps, "user_finish");
return;
}
}
async function finishRecording(sessionId: string, deps: RecordDeps): Promise<TraceV2 | null> {
async function finishRecording(
sessionId: string,
deps: RecordDeps,
stoppedBy: StopReason,
): Promise<RecordedTrace | null> {
const recording = recordings.get(sessionId);
if (!recording || recording.settled || recording.finishing) return null;
recording.finishing = true;
if (!recording || recording.settled) return null;
if (recording.finishAttempt) return recording.finishAttempt;
recording.stoppedBy = stoppedBy;
const attempt = finishRecordingAttempt(sessionId, recording, deps);
recording.finishAttempt = attempt;
try {
return await attempt;
} finally {
if (!recording.settled) {
recording.finishAttempt = null;
}
}
}
async function finishRecordingAttempt(
sessionId: string,
recording: ActiveRecording,
deps: RecordDeps,
): Promise<RecordedTrace | null> {
await clearRearmTimersForRecording(recording, deps);
try {
// Disposing content capture may commit a final dirty fill. Flush content
// first so all resulting RECORD_STEP messages enter actionQueue before it
// and the observation queues are drained.
await stopRecordingOnAllAgentTabs(recording, deps);
} catch {
recording.finishing = false;
return null;
}
if (!(await drainRecordingToStability(recording))) {
return null;
}
await recording.observation?.settleTrailing(recording.tabs.currentTabId, recording.steps);
recording.settled = true;
recordings.delete(sessionId);
@@ -558,6 +822,15 @@ export async function handleRecordStart(
params: RecordStartParams,
deps: RecordDeps = getDefaultDeps(),
): Promise<RecordStartResult | RpcError> {
const traceVersionOrErr = negotiatedTraceVersion(params);
if (typeof traceVersionOrErr !== "number") return traceVersionOrErr;
if (traceVersionOrErr === 3 && !deps.cdp) {
return {
code: "protocol_error",
message: "trace v3 recording requires an active CDP connection",
};
}
const ctxOrErr = lookupSession(manager, params, "record_start");
if (isRpcError(ctxOrErr)) return ctxOrErr;
const ctx = ctxOrErr;
@@ -574,37 +847,64 @@ export async function handleRecordStart(
// on the destination page can RECORD_QUERY → rearm → show RecordOverlay
// instead of flashing ControlOverlay ("Agent 正在控制").
const requestId = makeRequestId(target.tabId);
let resolveFinish!: (trace: TraceV2) => void;
let resolveFinish!: (trace: RecordedTrace) => void;
let rejectFinish!: (err: Error) => void;
const finishPromise = new Promise<TraceV2>((resolve, reject) => {
const finishPromise = new Promise<RecordedTrace>((resolve, reject) => {
resolveFinish = resolve;
rejectFinish = reject;
});
const navigateUrl = params.url ?? RECORD_DEFAULT_START_URL;
const startedAtMs = Date.now();
const maxPageTokens = params.max_page_tokens;
const redactValues = params.redact_values ?? false;
recordings.set(params.session_id, {
requestId,
tabId: target.tabId,
tabs: new RecordingTabCoordinator(target.tabId, navigateUrl),
agentWindowId: ctx.agentWindowId,
startUrl: navigateUrl,
...(params.purpose ? { purpose: params.purpose } : {}),
steps: [],
startedAt: new Date(startedAtMs).toISOString(),
startedAtMs,
traceVersion: traceVersionOrErr,
supportsTabSwitchSteps: params.supports_tab_switch_steps === true,
finishPromise,
resolveFinish,
rejectFinish,
settled: false,
finishing: false,
currentUrl: navigateUrl,
pendingNavigation: false,
pendingNavigationDeadline: undefined,
finishAttempt: null,
observation:
traceVersionOrErr === 3 && deps.cdp
? new RecordingObservationRuntime({
cdp: deps.cdp,
tabsApi: deps.tabsApi,
maxTokens: maxPageTokens,
redactValues,
})
: null,
stoppedBy: "user_finish",
navigationCallbacks: new Set(),
acceptingNavigation: true,
actionQueue: Promise.resolve(),
lastStepSequenceByProducer: new Map(),
});
deps.frameCoordinator?.begin(requestId, startedAtMs, target.tabId, async (scope) => {
const recording = recordings.get(params.session_id);
if (!recording || recording.requestId !== requestId || recording.settled) return;
try {
await recording.observation?.refreshDocument(scope.tabId, scope.producerId);
} catch {
// The normal action-time capture remains available if a child Document
// is replaced while its readiness refresh is in flight.
}
});
// Observe navigations for the whole recording lifetime; attach before
// optional navigate so the destination load can rearm capture.
ensureBrowserObservationListeners(deps);
const abortPending = async (notifyContent: boolean) => {
recordings.get(params.session_id)?.observation?.cancel();
deps.frameCoordinator?.cancel(requestId);
recordings.delete(params.session_id);
releaseBrowserObservationListenersIfIdle();
if (notifyContent) {
@@ -686,7 +986,7 @@ export async function handleRecordStart(
return cancelledError();
}
active.startUrl = startUrl;
active.currentUrl = startUrl;
active.tabs.navigation(target.tabId, startUrl).currentUrl = startUrl;
if (isContentScriptRestrictedUrl(startUrl)) {
await abortPending(false);
@@ -694,7 +994,7 @@ export async function handleRecordStart(
code: "invalid_params",
message: params.url
? `cannot record on restricted URL (${startUrl}); use an http(s) page`
: `cannot record on restricted URL (${startUrl}); default start page must be injectable http(s)`,
: `cannot record on restricted URL (${startUrl}); default start page https://example.com/ did not load — pass --url with a page you can open`,
};
}
@@ -721,6 +1021,10 @@ export async function handleRecordStart(
const startMsg: RecordStartMessage = { type: RECORD_START, requestId, startedAtMs };
try {
if (deps.frameCoordinator) {
const frameStarted = await deps.frameCoordinator.armTab(requestId, target.tabId);
if (!frameStarted) throw new Error("top recording document did not start");
}
await sendRecordStartWithAck(target.tabId, startMsg, deps.sendToTab);
} catch {
await abortPending(true);
@@ -731,6 +1035,17 @@ export async function handleRecordStart(
};
}
if (active.observation) {
const activeRecording = recordings.get(params.session_id);
if (activeRecording) {
try {
await activeRecording.observation?.captureInitial(target.tabId);
} catch {
// Proceed without initial observation; steps may be dropped by reducer.
}
}
}
{
const cancelled = await abortIfCancelled(true);
if (cancelled) return cancelled;
@@ -755,7 +1070,7 @@ export async function handleRecordStop(
};
}
const trace = await finishRecording(params.session_id, deps);
const trace = await finishRecording(params.session_id, deps, "cli_stop");
if (!trace) {
return {
code: "protocol_error",
@@ -785,9 +1100,9 @@ export async function handleRecordAwait(
return { code: "cancelled", message: "record_await aborted" };
}
const outcome = await new Promise<{ trace: TraceV2 } | { error: RpcError }>((resolve) => {
const outcome = await new Promise<{ trace: RecordedTrace } | { error: RpcError }>((resolve) => {
let settled = false;
const finish = (result: { trace: TraceV2 } | { error: RpcError }) => {
const finish = (result: { trace: RecordedTrace } | { error: RpcError }) => {
if (settled) return;
settled = true;
if (timer) clearTimeout(timer);
@@ -829,6 +1144,8 @@ export function clearRecordingForSession(sessionId: string): void {
}
void clearRearmTimersForRecording(recording, getDefaultDeps());
if (!recording.settled) {
getDefaultDeps().frameCoordinator?.cancel(recording.requestId);
recording.observation?.cancel();
recording.settled = true;
recording.rejectFinish(new Error("recording cleared"));
}
@@ -36,7 +36,7 @@ function fakeSnapshotReply() {
},
layout: {
nodeIndex: [1, 2, 3],
// styles columns follow requested computedStyles order: [position, pointer-events]
// Legacy fixture: omitted style columns fall back to visible defaults.
styles: [
[i("static"), i("auto")],
[i("fixed"), i("auto")],
@@ -682,6 +682,46 @@ describe("captureViewModel", () => {
expect(nodes.find((n) => n.backendNodeId === 11)?.cursor).toBe("auto");
});
it("captures painted visibility as DOM fallback evidence", async () => {
const S = ["html", "body", "button", "static", "auto", "visible", "hidden", "1", "0"];
const i = (value: string) => S.indexOf(value);
const snapshot = {
strings: S,
documents: [
{
nodes: {
parentIndex: [-1, 0, 1, 1, 1],
nodeName: [i("html"), i("body"), i("button"), i("button"), i("button")],
backendNodeId: [10, 11, 12, 13, 14],
attributes: [[], [], [], [], []],
},
layout: {
nodeIndex: [1, 2, 3, 4],
styles: [
[i("static"), i("auto"), i("auto"), i("visible"), i("1")],
[i("static"), i("auto"), i("auto"), i("visible"), i("1")],
[i("static"), i("auto"), i("auto"), i("hidden"), i("1")],
[i("static"), i("auto"), i("auto"), i("visible"), i("0")],
],
bounds: [
[0, 0, 1000, 800],
[10, 10, 100, 30],
[10, 50, 100, 30],
[10, 90, 100, 30],
],
paintOrders: [0, 1, 2, 3],
},
},
],
};
const { nodes } = await captureViewModel(makeCdp(snapshot), 4);
expect(nodes.find((node) => node.backendNodeId === 12)?.rendered).toBe(true);
expect(nodes.find((node) => node.backendNodeId === 13)?.rendered).toBe(false);
expect(nodes.find((node) => node.backendNodeId === 14)?.rendered).toBe(false);
});
it("subtracts scroll offset to make rects viewport-relative", async () => {
const cdp = {
send: vi.fn(async (_t: number, method: string) => {
@@ -132,4 +132,67 @@ describe("captureFrameData", () => {
]),
);
});
it("keeps OOPIF semantics when viewport projection is unavailable", async () => {
const owner = { ...ownerNode(10, 50), rect: null, localRect: null };
const captured: CapturedViewModel = {
nodes: [owner],
viewport: { width: 1000, height: 800 },
iframeNodes: new Map(),
frameNodes: new Map([["main", [owner]]]),
rootFrameId: "main",
excludedBackendNodeIds: new Set(),
};
const cdp: CdpRunner = {
send: vi.fn(async (_tabId, method) => {
if (method === "Accessibility.enable") return {};
if (method === "Accessibility.getFullAXTree") return { nodes: [] };
if (method === "DOM.getBoxModel") throw new Error("owner geometry unavailable");
if (method === "Page.getLayoutMetrics") {
return { cssLayoutViewport: { clientWidth: 1000, clientHeight: 800 } };
}
throw new Error(`unexpected root ${method}`);
}) as CdpRunner["send"],
sendToTarget: vi.fn(async (_target, method) => {
if (method === "Page.getLayoutMetrics") throw new Error("viewport unavailable");
if (method === "DOMSnapshot.enable" || method === "Accessibility.enable") return {};
if (method === "DOMSnapshot.captureSnapshot") return childSnapshot("child", 101);
if (method === "Accessibility.getFullAXTree") {
return {
nodes: [
{
nodeId: "button",
backendDOMNodeId: 101,
role: { type: "role", value: "button" },
name: { type: "computedString", value: "Continue" },
},
],
};
}
throw new Error(`unexpected ${method}`);
}) as unknown as NonNullable<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 documents = await captureFrameData(cdp, 4, captured);
const childDocument = documents.find((document) => document.frameId === "child");
expect(childDocument?.axNodes).toEqual([
expect.objectContaining({ backendDOMNodeId: 101, frameId: "child" }),
]);
expect(childDocument?.domNodes.find((node) => node.backendNodeId === 101)).toEqual(
expect.objectContaining({ rect: null, localRect: { x: 10, y: 20, w: 100, h: 40 } }),
);
});
});
@@ -0,0 +1,183 @@
import { describe, expect, it, vi } from "vitest";
import type { CdpTarget } from "@/browser-driver/frame-graph";
import type { CdpRunner } from "../../shared";
import type { CapturedNode } from "../capture";
import type { FrameDocument, FrameOwnedAxNode } from "../frame-document";
import { probeTooltipNames } from "../name-enrichment";
import { buildSemanticGraph, resolveSemanticGraph } from "../semantic-graph";
import { stableIdentifierName } from "../semantic-graph/name-evidence";
import { frameBackendKey } from "../semantic-graph/types";
function document(
target: CdpTarget,
backendNodeId: number,
accessibleName?: string,
frameId = "child",
x = 10,
): FrameDocument<FrameOwnedAxNode> {
const node: CapturedNode = {
backendNodeId,
parentBackendNodeId: null,
frameId,
tag: "button",
attrs: {
id: "toolInsertRecord",
"data-test-id": "toolInsertRecord",
...(accessibleName ? { "aria-label": accessibleName } : {}),
},
rect: { x, y: 20, w: 40, h: 30 },
localRect: { x: 2, y: 3, w: 40, h: 30 },
paintOrder: 1,
position: "static",
pointerEvents: "auto",
cursor: "pointer",
};
return {
frameId,
parentFrameId: "main",
contextScopeId: "child",
target,
axNodes: [],
domNodes: [node],
url: "https://example.com/frame",
};
}
function semantics(documents: FrameDocument<FrameOwnedAxNode>[]) {
return resolveSemanticGraph(
buildSemanticGraph({
documents,
viewport: { width: 800, height: 600 },
rootFrameId: documents[0]?.frameId,
}),
{ identifierFallback: false },
);
}
describe("tooltip name enrichment", () => {
it("uses the owning frame target and upgrades an identifier name from a tooltip", async () => {
const target = { tabId: 7, sessionId: "oopif-session" };
let hovered = false;
const cdp: CdpRunner = {
send: vi.fn(async (_tabId, method, params) => {
if (method === "Input.dispatchMouseEvent") {
hovered = (params as { x?: number }).x === 30;
}
return {};
}) as CdpRunner["send"],
sendToTarget: vi.fn(async (_target, method) => {
if (method === "Page.createIsolatedWorld") return { executionContextId: 11 };
if (method === "DOM.resolveNode") return { object: { objectId: "button-501" } };
if (method === "Runtime.callFunctionOn") {
return { result: { value: hovered ? ["插入行"] : [] } };
}
return {};
}) as CdpRunner["sendToTarget"],
};
const documents = [document(target, 501)];
const graph = semantics(documents);
const names = await probeTooltipNames(cdp, 7, documents, graph);
expect(names.get(frameBackendKey("child", 501))).toBe("插入行");
expect(cdp.sendToTarget).toHaveBeenCalledWith(
target,
"Page.createIsolatedWorld",
expect.objectContaining({ frameId: "child" }),
);
const sendsBeforeCache = vi.mocked(cdp.send).mock.calls.length;
const targetSendsBeforeCache = vi.mocked(
cdp.sendToTarget as NonNullable<CdpRunner["sendToTarget"]>,
).mock.calls.length;
const cached = await probeTooltipNames(cdp, 7, documents, graph);
expect(cached.get(frameBackendKey("child", 501))).toBe("插入行");
expect(cdp.send).toHaveBeenCalledTimes(sendsBeforeCache);
expect(cdp.sendToTarget).toHaveBeenCalledTimes(targetSendsBeforeCache);
});
it("does not probe controls that already have a higher-confidence name", async () => {
const cdp: CdpRunner = {
send: vi.fn(async () => ({})) as CdpRunner["send"],
sendToTarget: vi.fn(async () => ({ executionContextId: 12 })) as CdpRunner["sendToTarget"],
};
const documents = [document({ tabId: 8 }, 601, "Add a new record")];
const names = await probeTooltipNames(cdp, 8, documents, semantics(documents));
expect(names.size).toBe(0);
expect(cdp.send).not.toHaveBeenCalled();
expect(cdp.sendToTarget).not.toHaveBeenCalled();
});
it("does not collect supplemental names from form controls", async () => {
const cdp: CdpRunner = {
send: vi.fn(async () => ({})) as CdpRunner["send"],
sendToTarget: vi.fn(async () => ({ executionContextId: 12 })) as CdpRunner["sendToTarget"],
};
const inputDocument = document({ tabId: 8 }, 602);
inputDocument.domNodes[0] = {
...inputDocument.domNodes[0],
tag: "input",
attrs: { type: "email" },
formValue: "user@example.com",
formDefaultValue: "",
formState: "filled",
};
const names = await probeTooltipNames(cdp, 8, [inputDocument], semantics([inputDocument]));
expect(names.size).toBe(0);
expect(cdp.send).not.toHaveBeenCalled();
expect(cdp.sendToTarget).not.toHaveBeenCalled();
});
it("keeps tooltip evidence isolated across sibling frame sessions", async () => {
let hoveredX = -10;
const cdp: CdpRunner = {
send: vi.fn(async (_tabId, method, params) => {
if (method === "Input.dispatchMouseEvent") {
hoveredX = (params as { x: number }).x;
}
return {};
}) as CdpRunner["send"],
sendToTarget: vi.fn(async (target, method, params) => {
if (method === "Page.createIsolatedWorld") {
return { executionContextId: target.sessionId === "left-session" ? 21 : 22 };
}
if (method === "DOM.resolveNode") {
return { object: { objectId: target.sessionId === "left-session" ? "left" : "right" } };
}
if (method === "Runtime.callFunctionOn") {
const objectId = (params as { objectId: string }).objectId;
const value =
objectId === "left" && hoveredX === 30
? ["Left action"]
: objectId === "right" && hoveredX === 130
? ["Right action"]
: [];
return { result: { value } };
}
return {};
}) as CdpRunner["sendToTarget"],
};
const documents = [
document({ tabId: 9, sessionId: "left-session" }, 701, undefined, "left", 10),
document({ tabId: 9, sessionId: "right-session" }, 701, undefined, "right", 110),
];
const names = await probeTooltipNames(cdp, 9, documents, semantics(documents));
expect(names.get(frameBackendKey("left", 701))).toBe("Left action");
expect(names.get(frameBackendKey("right", 701))).toBe("Right action");
});
});
describe("stable identifier evidence", () => {
it("accepts readable action identifiers and rejects framework-generated ids", () => {
expect(stableIdentifierName({ id: "toolInsertRecord" })).toBe("Insert record");
expect(stableIdentifierName({ id: "reactivateAccount" })).toBe("Reactivate account");
expect(stableIdentifierName({ id: "headlessui-menu-button-r1" })).toBeUndefined();
expect(stableIdentifierName({ id: "7f15a3bc91" })).toBeUndefined();
});
});
@@ -0,0 +1,82 @@
import { describe, expect, it } from "vitest";
import {
RECORD_DOCUMENT_ATTRIBUTE,
recordingDocumentMarkerValue,
} from "@/shared/recording-document-identity";
import { projectRecordSafeObservation } from "../record-safe-observation";
describe("projectRecordSafeObservation", () => {
it("projects the opaque recording identity without exposing captured attributes", () => {
const recordingDocumentId = "123e4567-e89b-42d3-a456-426614174000";
const result = projectRecordSafeObservation({
rootFrameId: "child",
frameDocuments: [
{
frameId: "child",
target: { tabId: 4, sessionId: "oopif-session" },
contextScopeId: "scope-1",
axNodes: [],
domNodes: [
{
backendNodeId: 1,
parentBackendNodeId: null,
tag: "html",
attrs: {
[RECORD_DOCUMENT_ATTRIBUTE]: recordingDocumentMarkerValue(recordingDocumentId),
value: "user@example.com",
},
rect: null,
paintOrder: 0,
position: "static",
pointerEvents: "auto",
},
],
},
],
rendered: { text: '@vom 1\nRootWebArea "Example"', refs: [], truncated: false },
});
expect(result.frames).toEqual([
{
frameId: "child",
target: { tabId: 4, sessionId: "oopif-session" },
recordingDocumentId,
},
]);
const dumped = JSON.stringify(result);
expect(dumped).not.toContain("user@example.com");
expect(dumped).not.toContain("attrs");
expect(dumped).not.toContain("domNodes");
expect(dumped).not.toContain("axNodes");
});
it("does not project a page-controlled marker value as recording identity", () => {
const result = projectRecordSafeObservation({
rootFrameId: "root",
frameDocuments: [
{
frameId: "root",
target: { tabId: 4 },
contextScopeId: "scope-1",
axNodes: [],
domNodes: [
{
backendNodeId: 1,
parentBackendNodeId: null,
tag: "html",
attrs: { [RECORD_DOCUMENT_ATTRIBUTE]: "bsk:user@example.com" },
rect: null,
paintOrder: 0,
position: "static",
pointerEvents: "auto",
},
],
},
],
rendered: { text: '@vom 1\nRootWebArea "Example"', refs: [], truncated: false },
});
expect(result.frames).toEqual([{ frameId: "root", target: { tabId: 4 } }]);
expect(JSON.stringify(result)).not.toContain("user@example.com");
});
});
@@ -0,0 +1,782 @@
import { isVomReferenceNode, renderVom } from "@browser-skill/vom";
import { describe, expect, it } from "vitest";
import type { CapturedNode } from "../capture";
import type { FrameDocument } from "../frame-document";
import { buildSemanticVomScene, type SemanticAxNode } from "../semantic-graph";
function dom(
backendNodeId: number,
parentBackendNodeId: number | null,
tag: string,
attrs: Record<string, string> = {},
textContent?: string,
): CapturedNode {
return {
backendNodeId,
parentBackendNodeId,
tag,
attrs,
rect: { x: 0, y: 0, w: 100, h: 30 },
paintOrder: backendNodeId,
position: "static",
pointerEvents: "auto",
...(textContent ? { textContent } : {}),
};
}
function document(
frameId: string,
axNodes: SemanticAxNode[],
domNodes: CapturedNode[],
frame?: Pick<FrameDocument<SemanticAxNode>, "parentFrameId" | "ownerBackendNodeId">,
): FrameDocument<SemanticAxNode> {
return {
frameId,
contextScopeId: frameId,
target: { tabId: 1 },
axNodes,
domNodes,
...frame,
};
}
describe("semantic VOM graph", () => {
it("preserves backend-less AX structure without making it referenceable", () => {
const scene = buildSemanticVomScene({
viewport: { width: 800, height: 600 },
rootFrameId: "main",
documents: [
document(
"main",
[
{
nodeId: "root",
backendDOMNodeId: 1,
role: { type: "role", value: "RootWebArea" },
childIds: ["toolbar"],
},
{
nodeId: "toolbar",
parentId: "root",
role: { type: "role", value: "toolbar" },
name: { type: "computedString", value: "Formatting" },
childIds: ["bold", "italic"],
},
{
nodeId: "bold",
parentId: "toolbar",
backendDOMNodeId: 2,
role: { type: "role", value: "button" },
name: { type: "computedString", value: "Bold" },
},
{
nodeId: "italic",
parentId: "toolbar",
backendDOMNodeId: 3,
role: { type: "role", value: "button" },
name: { type: "computedString", value: "Italic" },
},
],
[dom(1, null, "body"), dom(2, 1, "button"), dom(3, 1, "button")],
),
],
});
const toolbar = scene.nodes.find((node) => node.role === "toolbar");
const buttons = scene.nodes.filter((node) => node.role === "button");
expect(toolbar).toEqual(expect.objectContaining({ name: "Formatting", referenceable: false }));
expect(buttons.map((node) => node.parentId)).toEqual([toolbar?.id, toolbar?.id]);
expect(toolbar && isVomReferenceNode(toolbar)).toBe(false);
expect(renderVom(scene).refs.map((ref) => ref.backendNodeId)).toEqual([2, 3]);
});
it("resolves missing AX names through one DOM naming pipeline", () => {
const scene = buildSemanticVomScene({
viewport: { width: 800, height: 600 },
rootFrameId: "main",
documents: [
document(
"main",
[
{
nodeId: "root",
backendDOMNodeId: 1,
role: { type: "role", value: "RootWebArea" },
childIds: ["labelled", "titled", "text"],
},
{
nodeId: "labelled",
parentId: "root",
backendDOMNodeId: 2,
role: { type: "role", value: "button" },
},
{
nodeId: "titled",
parentId: "root",
backendDOMNodeId: 3,
role: { type: "role", value: "button" },
},
{
nodeId: "text",
parentId: "root",
backendDOMNodeId: 4,
role: { type: "role", value: "button" },
},
],
[
dom(1, null, "body"),
dom(2, 1, "button", { "aria-label": "Add row" }),
dom(3, 1, "button", { title: "Settings" }),
dom(4, 1, "button"),
dom(5, 4, "span", {}, "Save"),
],
),
],
});
expect(scene.nodes.filter((node) => node.role === "button").map((node) => node.name)).toEqual([
"Add row",
"Settings",
"Save",
]);
});
it("aggregates AX text at the semantic text frontier", () => {
const scene = buildSemanticVomScene({
viewport: { width: 800, height: 600 },
rootFrameId: "main",
documents: [
document(
"main",
[
{
nodeId: "root",
backendDOMNodeId: 1,
role: { type: "role", value: "RootWebArea" },
childIds: ["covered", "inline-only", "repeated", "labelled"],
},
{
nodeId: "covered",
parentId: "root",
backendDOMNodeId: 2,
role: { type: "role", value: "paragraph" },
childIds: ["covered-static"],
},
{
nodeId: "covered-static",
parentId: "covered",
role: { type: "role", value: "StaticText" },
name: { type: "computedString", value: "Covered text" },
childIds: ["covered-inline"],
},
{
nodeId: "covered-inline",
parentId: "covered-static",
role: { type: "role", value: "InlineTextBox" },
name: { type: "computedString", value: "Covered text" },
},
{
nodeId: "inline-only",
parentId: "root",
backendDOMNodeId: 3,
role: { type: "role", value: "paragraph" },
childIds: ["inline-leaf"],
},
{
nodeId: "inline-leaf",
parentId: "inline-only",
role: { type: "role", value: "InlineTextBox" },
name: { type: "computedString", value: "Inline fallback" },
},
{
nodeId: "repeated",
parentId: "root",
backendDOMNodeId: 4,
role: { type: "role", value: "paragraph" },
childIds: ["first-go", "second-go"],
},
{
nodeId: "first-go",
parentId: "repeated",
role: { type: "role", value: "StaticText" },
name: { type: "computedString", value: "Go" },
},
{
nodeId: "second-go",
parentId: "repeated",
role: { type: "role", value: "StaticText" },
name: { type: "computedString", value: "Go" },
},
{
nodeId: "labelled",
parentId: "root",
backendDOMNodeId: 5,
role: { type: "role", value: "button" },
childIds: ["visible-label"],
},
{
nodeId: "visible-label",
parentId: "labelled",
role: { type: "role", value: "StaticText" },
name: { type: "computedString", value: "Visible text" },
},
],
[
dom(1, null, "body"),
dom(2, 1, "p"),
dom(3, 1, "p"),
dom(4, 1, "p"),
dom(5, 1, "button", { "aria-label": "Explicit action" }),
],
),
],
});
expect(scene.nodes.find((node) => node.backendNodeId === 2)?.name).toBe("Covered text");
expect(scene.nodes.find((node) => node.backendNodeId === 3)?.name).toBe("Inline fallback");
expect(scene.nodes.find((node) => node.backendNodeId === 4)?.name).toBe("Go Go");
expect(scene.nodes.find((node) => node.backendNodeId === 5)?.name).toBe("Explicit action");
expect(renderVom(scene).text).not.toContain("Covered text Covered text");
});
it("uses stable identifiers only as a final name fallback", () => {
const base = document(
"main",
[
{
nodeId: "root",
backendDOMNodeId: 1,
role: { type: "role", value: "RootWebArea" },
childIds: ["fallback", "labelled", "container"],
},
{
nodeId: "fallback",
parentId: "root",
backendDOMNodeId: 2,
role: { type: "role", value: "button" },
},
{
nodeId: "labelled",
parentId: "root",
backendDOMNodeId: 3,
role: { type: "role", value: "button" },
name: { type: "computedString", value: "Localized name" },
},
{
nodeId: "container",
parentId: "root",
backendDOMNodeId: 4,
role: { type: "role", value: "generic" },
},
],
[
dom(1, null, "body"),
dom(2, 1, "button", { "data-test-id": "toolInsertRecord" }),
dom(3, 1, "button", { id: "toolIgnoredFallback" }),
dom(4, 1, "div", { id: "toolbarContainer" }),
],
);
const staticScene = buildSemanticVomScene({
viewport: { width: 800, height: 600 },
rootFrameId: "main",
documents: [base],
});
const enrichedScene = buildSemanticVomScene({
viewport: { width: 800, height: 600 },
rootFrameId: "main",
documents: [base],
supplementalNames: new Map([["main\u00002", "插入行"]]),
});
expect(staticScene.nodes.find((node) => node.backendNodeId === 2)?.name).toBe("Insert record");
expect(staticScene.nodes.find((node) => node.backendNodeId === 3)?.name).toBe("Localized name");
expect(staticScene.nodes.find((node) => node.backendNodeId === 4)?.name).toBeUndefined();
expect(enrichedScene.nodes.find((node) => node.backendNodeId === 2)?.name).toBe("插入行");
expect(enrichedScene.nodes.find((node) => node.backendNodeId === 3)?.name).toBe(
"Localized name",
);
});
it("isolates equal backend node ids in sibling frames", () => {
const scene = buildSemanticVomScene({
viewport: { width: 1000, height: 800 },
rootFrameId: "main",
documents: [
document(
"main",
[
{
nodeId: "root",
backendDOMNodeId: 1,
role: { type: "role", value: "RootWebArea" },
childIds: ["owner-a", "owner-b"],
},
{
nodeId: "owner-a",
parentId: "root",
backendDOMNodeId: 10,
role: { type: "role", value: "Iframe" },
},
{
nodeId: "owner-b",
parentId: "root",
backendDOMNodeId: 20,
role: { type: "role", value: "Iframe" },
},
],
[dom(1, null, "body"), dom(10, 1, "iframe"), dom(20, 1, "iframe")],
),
document(
"frame-a",
[
{
nodeId: "button-a",
backendDOMNodeId: 5,
role: { type: "role", value: "button" },
childIds: ["text-a"],
},
{
nodeId: "text-a",
parentId: "button-a",
role: { type: "role", value: "StaticText" },
name: { type: "computedString", value: "Accept" },
childIds: ["inline-a"],
},
{
nodeId: "inline-a",
parentId: "text-a",
role: { type: "role", value: "InlineTextBox" },
name: { type: "computedString", value: "Accept" },
},
],
[dom(5, null, "button")],
{ parentFrameId: "main", ownerBackendNodeId: 10 },
),
document(
"frame-b",
[
{
nodeId: "button-b",
backendDOMNodeId: 5,
role: { type: "role", value: "button" },
childIds: ["text-b"],
},
{
nodeId: "text-b",
parentId: "button-b",
role: { type: "role", value: "StaticText" },
name: { type: "computedString", value: "Decline" },
childIds: ["inline-b"],
},
{
nodeId: "inline-b",
parentId: "text-b",
role: { type: "role", value: "InlineTextBox" },
name: { type: "computedString", value: "Decline" },
},
],
[dom(5, null, "button")],
{ parentFrameId: "main", ownerBackendNodeId: 20 },
),
],
});
const frameButtons = scene.nodes.filter((node) => node.backendNodeId === 5);
const owners = new Map(
scene.nodes
.filter((node) => node.backendNodeId === 10 || node.backendNodeId === 20)
.map((node) => [node.backendNodeId, node.id]),
);
expect(frameButtons.map((node) => node.frameId)).toEqual(["frame-a", "frame-b"]);
expect(frameButtons.map((node) => node.name)).toEqual(["Accept", "Decline"]);
expect(frameButtons.map((node) => node.parentId)).toEqual([owners.get(10), owners.get(20)]);
expect(renderVom(scene).refs.map((ref) => [ref.frameId, ref.backendNodeId])).toEqual([
["frame-a", 5],
["frame-b", 5],
]);
});
it("requires render evidence before promoting DOM fallback semantics", () => {
const hiddenOwner = {
...dom(10, 1, "iframe", { hidden: "" }),
rect: null,
localRect: null,
rendered: false,
};
const hiddenButton = {
...dom(2, 1, "button", { "aria-label": "Hidden action" }),
rendered: false,
};
const scene = buildSemanticVomScene({
viewport: { width: 800, height: 600 },
rootFrameId: "main",
documents: [
document(
"main",
[
{
nodeId: "root",
backendDOMNodeId: 1,
role: { type: "role", value: "RootWebArea" },
},
],
[dom(1, null, "body"), hiddenButton, hiddenOwner],
),
document("hidden-frame", [], [dom(20, null, "button", {}, "Frame action")], {
parentFrameId: "main",
ownerBackendNodeId: 10,
}),
],
});
expect(scene.nodes.some((node) => node.backendNodeId === 2)).toBe(false);
expect(scene.nodes.some((node) => node.backendNodeId === 10)).toBe(false);
expect(scene.nodes.some((node) => node.frameId === "hidden-frame")).toBe(false);
expect(renderVom(scene).refs).toEqual([]);
});
it("retains rendered DOM fallback controls when AX semantics are unavailable", () => {
const scene = buildSemanticVomScene({
viewport: { width: 800, height: 600 },
rootFrameId: "main",
documents: [
document(
"main",
[],
[dom(1, null, "body"), { ...dom(2, 1, "button", {}, "Save"), rendered: true }],
),
],
});
expect(renderVom(scene).refs).toEqual([
expect.objectContaining({ backendNodeId: 2, role: "button", name: "Save" }),
]);
});
it("lets explicit DOM structure override ignored AX metadata", () => {
const scene = buildSemanticVomScene({
viewport: { width: 800, height: 600 },
rootFrameId: "main",
documents: [
document(
"main",
[
{
nodeId: "root",
backendDOMNodeId: 1,
role: { type: "role", value: "RootWebArea" },
childIds: ["toolbar"],
},
{
nodeId: "toolbar",
parentId: "root",
backendDOMNodeId: 2,
ignored: true,
role: { type: "role", value: "generic" },
childIds: ["wrapper"],
},
{
nodeId: "wrapper",
parentId: "toolbar",
backendDOMNodeId: 5,
role: { type: "role", value: "generic" },
childIds: ["bold", "italic"],
},
{
nodeId: "bold",
parentId: "wrapper",
backendDOMNodeId: 3,
role: { type: "role", value: "button" },
name: { type: "computedString", value: "Bold" },
},
{
nodeId: "italic",
parentId: "wrapper",
backendDOMNodeId: 4,
role: { type: "role", value: "button" },
name: { type: "computedString", value: "Italic" },
},
],
[
dom(1, null, "body"),
dom(2, 1, "div", { role: "toolbar" }),
dom(5, 2, "div"),
dom(3, 5, "button"),
dom(4, 5, "button"),
],
),
],
});
const toolbar = scene.nodes.find((node) => node.backendNodeId === 2);
const buttons = scene.nodes.filter((node) => [3, 4].includes(node.backendNodeId ?? -1));
expect(toolbar).toEqual(expect.objectContaining({ role: "toolbar" }));
expect(toolbar && isVomReferenceNode(toolbar)).toBe(false);
expect(scene.nodes.some((node) => node.role === "group")).toBe(false);
expect(buttons.map((node) => node.parentId)).toEqual([toolbar?.id, toolbar?.id]);
});
it("infers an unnamed group only from independent interaction branches", () => {
const scene = buildSemanticVomScene({
viewport: { width: 800, height: 600 },
rootFrameId: "main",
documents: [
document(
"main",
[
{
nodeId: "root",
backendDOMNodeId: 1,
role: { type: "role", value: "RootWebArea" },
childIds: ["wrapper"],
},
{
nodeId: "wrapper",
parentId: "root",
backendDOMNodeId: 2,
role: { type: "role", value: "generic" },
childIds: ["first", "second"],
},
{
nodeId: "first",
parentId: "wrapper",
backendDOMNodeId: 3,
role: { type: "role", value: "button" },
name: { type: "computedString", value: "First" },
},
{
nodeId: "second",
parentId: "wrapper",
backendDOMNodeId: 4,
role: { type: "role", value: "button" },
name: { type: "computedString", value: "Second" },
},
],
[dom(1, null, "body"), dom(2, 1, "div"), dom(3, 2, "button"), dom(4, 2, "button")],
),
],
});
const group = scene.nodes.find((node) => node.backendNodeId === 2);
const buttons = scene.nodes.filter((node) => [3, 4].includes(node.backendNodeId ?? -1));
expect(group).toEqual(expect.objectContaining({ role: "group" }));
expect(group?.name).toBeUndefined();
expect(group && isVomReferenceNode(group)).toBe(false);
expect(buttons.map((node) => node.parentId)).toEqual([group?.id, group?.id]);
const rendered = renderVom(scene);
const lines = rendered.text.split("\n");
const groupLine = lines.findIndex((line) => line.trim() === "group");
expect(groupLine).toBeGreaterThanOrEqual(0);
for (const ref of rendered.refs) {
expect(ref.line).toBeGreaterThan(groupLine);
expect(lines[ref.line]).toMatch(/^\s+@e\d+ button/);
}
});
it("never promotes a form value into its semantic name through supplemental evidence", () => {
const secret = "user@example.com";
const input = {
...dom(2, 1, "input", { type: "email" }),
formValue: secret,
formDefaultValue: "",
formState: "filled" as const,
};
const scene = buildSemanticVomScene({
viewport: { width: 800, height: 600 },
rootFrameId: "main",
supplementalNames: new Map([["main\u00002", secret]]),
identifierFallback: false,
documents: [
document(
"main",
[
{
nodeId: "root",
backendDOMNodeId: 1,
role: { type: "role", value: "RootWebArea" },
childIds: ["email"],
},
{
nodeId: "email",
parentId: "root",
backendDOMNodeId: 2,
role: { type: "role", value: "textbox" },
value: { value: secret },
},
],
[dom(1, null, "body"), input],
),
],
});
const field = scene.nodes.find((node) => node.backendNodeId === 2);
expect(field?.name).toBeUndefined();
const rendered = renderVom(scene, { redactValues: true });
expect(rendered.text).toContain("•••");
expect(JSON.stringify(rendered)).not.toContain(secret);
});
it("keeps uncertain single-branch wrappers flat", () => {
const scene = buildSemanticVomScene({
viewport: { width: 800, height: 600 },
rootFrameId: "main",
documents: [
document(
"main",
[
{
nodeId: "root",
backendDOMNodeId: 1,
role: { type: "role", value: "RootWebArea" },
childIds: ["wrapper"],
},
{
nodeId: "wrapper",
parentId: "root",
backendDOMNodeId: 2,
role: { type: "role", value: "generic" },
childIds: ["button"],
},
{
nodeId: "button",
parentId: "wrapper",
backendDOMNodeId: 3,
role: { type: "role", value: "button" },
name: { type: "computedString", value: "Only action" },
},
],
[dom(1, null, "body"), dom(2, 1, "div"), dom(3, 2, "button")],
),
],
});
expect(scene.nodes.some((node) => node.role === "group")).toBe(false);
expect(scene.nodes.find((node) => node.backendNodeId === 3)?.parentId).toBe(
scene.nodes.find((node) => node.backendNodeId === 1)?.id,
);
});
it("does not infer across mixed content boundaries", () => {
const scene = buildSemanticVomScene({
viewport: { width: 800, height: 600 },
rootFrameId: "main",
documents: [
document(
"main",
[
{
nodeId: "root",
backendDOMNodeId: 1,
role: { type: "role", value: "RootWebArea" },
childIds: ["wrapper"],
},
{
nodeId: "wrapper",
parentId: "root",
backendDOMNodeId: 2,
role: { type: "role", value: "generic" },
childIds: ["first", "second", "paragraph"],
},
{
nodeId: "first",
parentId: "wrapper",
backendDOMNodeId: 3,
role: { type: "role", value: "button" },
name: { type: "computedString", value: "First" },
},
{
nodeId: "second",
parentId: "wrapper",
backendDOMNodeId: 4,
role: { type: "role", value: "button" },
name: { type: "computedString", value: "Second" },
},
{
nodeId: "paragraph",
parentId: "wrapper",
backendDOMNodeId: 5,
role: { type: "role", value: "paragraph" },
},
],
[
dom(1, null, "body"),
dom(2, 1, "div"),
dom(3, 2, "button"),
dom(4, 2, "button"),
dom(5, 2, "p", {}, "Description"),
],
),
],
});
expect(scene.nodes.some((node) => node.role === "group")).toBe(false);
});
it("keeps only the innermost reliable inferred group", () => {
const scene = buildSemanticVomScene({
viewport: { width: 800, height: 600 },
rootFrameId: "main",
documents: [
document(
"main",
[
{
nodeId: "root",
backendDOMNodeId: 1,
role: { type: "role", value: "RootWebArea" },
childIds: ["outer"],
},
{
nodeId: "outer",
parentId: "root",
backendDOMNodeId: 2,
role: { type: "role", value: "generic" },
childIds: ["inner", "third"],
},
{
nodeId: "inner",
parentId: "outer",
backendDOMNodeId: 3,
role: { type: "role", value: "generic" },
childIds: ["first", "second"],
},
{
nodeId: "first",
parentId: "inner",
backendDOMNodeId: 4,
role: { type: "role", value: "button" },
name: { type: "computedString", value: "First" },
},
{
nodeId: "second",
parentId: "inner",
backendDOMNodeId: 5,
role: { type: "role", value: "button" },
name: { type: "computedString", value: "Second" },
},
{
nodeId: "third",
parentId: "outer",
backendDOMNodeId: 6,
role: { type: "role", value: "button" },
name: { type: "computedString", value: "Third" },
},
],
[
dom(1, null, "body"),
dom(2, 1, "div"),
dom(3, 2, "div"),
dom(4, 3, "button"),
dom(5, 3, "button"),
dom(6, 2, "button"),
],
),
],
});
const groups = scene.nodes.filter((node) => node.role === "group");
expect(groups.map((node) => node.backendNodeId)).toEqual([3]);
expect(scene.nodes.find((node) => node.backendNodeId === 6)?.parentId).not.toBe(groups[0]?.id);
});
});
+30 -36
View File
@@ -1,16 +1,17 @@
// CDP capture adapter: DOMSnapshot.captureSnapshot + Page.getLayoutMetrics
// → CapturedNode[] + Viewport. This is the ONLY VOM module that touches
// raw CDP. The captureSnapshot reply is columnar (parallel arrays + a
// shared string table); we request exactly three computed styles so the
// `styles` columns are [position, pointer-events, cursor] in that order.
// shared string table); requested computed styles are decoded through
// STYLE_COL so capture and visibility policy share one explicit contract.
import type { Rect, Viewport } from "@browser-skill/vom";
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, waitForHover } from "./hover-perception";
const REQUESTED_STYLES = ["position", "pointer-events", "cursor"] as const;
const REQUESTED_STYLES = ["position", "pointer-events", "cursor", "visibility", "opacity"] as const;
const STYLE_COL = Object.fromEntries(
REQUESTED_STYLES.map((name, index) => [name, index]),
) as Record<(typeof REQUESTED_STYLES)[number], number>;
@@ -38,6 +39,12 @@ export interface CapturedNode {
* `textContent`: the live parser always sets it, hand-built fixtures may not.
*/
cursor?: string;
/**
* 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
* nodes remain authoritative even when they are outside the viewport.
*/
rendered?: boolean;
textContent?: string;
formState?: "empty" | "filled" | "default";
formValue?: string;
@@ -460,36 +467,6 @@ function hoverStateExpression(): string {
})()`;
}
async function wait(ms: number, signal?: AbortSignal): Promise<void> {
throwIfAborted(signal);
await new Promise<void>((resolve, reject) => {
const timer = setTimeout(() => {
signal?.removeEventListener("abort", onAbort);
resolve();
}, ms);
const onAbort = () => {
clearTimeout(timer);
signal?.removeEventListener("abort", onAbort);
reject(captureAbortError());
};
signal?.addEventListener("abort", onAbort, { once: true });
});
}
async function clearHover(cdp: CdpRunner, tabId: number): Promise<void> {
await cdp
.send(tabId, "Input.dispatchMouseEvent", {
type: "mouseMoved",
x: -10,
y: -10,
})
.catch(() =>
cdp.send(tabId, "Input.dispatchMouseEvent", { type: "mouseMoved", x: 0, y: 0 }).catch(() => {
// best effort
}),
);
}
function capturedText(node: CapturedNode): string | undefined {
const value =
node.attrs["aria-label"] ?? node.attrs.title ?? node.attrs.alt ?? node.textContent ?? "";
@@ -669,7 +646,7 @@ async function probeHoverSurfaces(
try {
await clearHover(cdp, tabId);
throwIfAborted(options.signal);
await wait(HOVER_SETTLE_MS, options.signal);
await waitForHover(HOVER_SETTLE_MS, options.signal);
const baselineReply = await cdp.send<RuntimeEvaluateReply>(tabId, "Runtime.evaluate", {
expression: hoverStateExpression(),
returnByValue: true,
@@ -684,7 +661,7 @@ async function probeHoverSurfaces(
y: candidate.y,
});
throwIfAborted(options.signal);
await wait(HOVER_SETTLE_MS, options.signal);
await waitForHover(HOVER_SETTLE_MS, options.signal);
const collected = await cdp.send<RuntimeEvaluateReply>(tabId, "Runtime.evaluate", {
expression: hoverStateExpression(),
returnByValue: true,
@@ -817,6 +794,8 @@ function parseDocumentNodes(
const posCol = STYLE_COL.position;
const peCol = STYLE_COL["pointer-events"];
const cursorCol = STYLE_COL.cursor;
const visibilityCol = STYLE_COL.visibility;
const opacityCol = STYLE_COL.opacity;
const inputValues = sparseIndexMap(dn.inputValue);
const textValues = sparseIndexMap(dn.textValue);
const checkedInputs = new Set(dn.inputChecked?.index ?? []);
@@ -861,6 +840,7 @@ function parseDocumentNodes(
let position = "static";
let pointerEvents = "auto";
let cursor = "auto";
let rendered = false;
const li = layoutByNode.get(n);
if (li !== undefined) {
const b = dl?.bounds?.[li];
@@ -881,6 +861,13 @@ function parseDocumentNodes(
position = str(strings, styleRow[posCol]) || "static";
pointerEvents = str(strings, styleRow[peCol]) || "auto";
cursor = str(strings, styleRow[cursorCol]) || "auto";
const visibility = str(strings, styleRow[visibilityCol]) || "visible";
const opacity = str(strings, styleRow[opacityCol]) || "1";
rendered =
localRect !== null &&
visibility !== "hidden" &&
visibility !== "collapse" &&
(Number.parseFloat(opacity) || 0) > 0;
}
// Skip non-element nodes (#text, #cdata-section, etc.) — they carry no
@@ -916,6 +903,7 @@ function parseDocumentNodes(
position,
pointerEvents,
cursor,
rendered,
textContent,
...(formValue !== undefined ? { formValue } : {}),
...(tag === "input" || tag === "textarea"
@@ -1020,7 +1008,13 @@ export async function captureViewModel(
options: CaptureViewModelOptions = {},
): Promise<CapturedViewModel> {
throwIfAborted(options.signal);
const metrics = await cdp.send<LayoutMetricsReply>(tabId, "Page.getLayoutMetrics", {});
let metrics: LayoutMetricsReply = {};
try {
metrics = await cdp.send<LayoutMetricsReply>(tabId, "Page.getLayoutMetrics", {});
} catch (error) {
if (isAbortError(error)) throw error;
console.debug("[bsk capture] layout metrics unavailable", error);
}
throwIfAborted(options.signal);
const dpr = devicePixelRatio(metrics);
const vpSrc = metrics.cssLayoutViewport ?? metrics.layoutViewport ?? {};
+24 -22
View File
@@ -33,25 +33,19 @@ async function discoverFrameGraph(cdp: CdpRunner, tabId: number): Promise<CdpFra
function transformFrameNodes(
nodes: CapturedNode[],
projection: GeometryProjection,
projection: GeometryProjection | null,
): CapturedNode[] {
return nodes.map((node) => ({
...node,
rect: (() => {
const bounds = projectRectToViewport(node.rect, projection);
return bounds ? { x: bounds.x, y: bounds.y, w: bounds.width, h: bounds.height } : null;
})(),
rect: projection
? (() => {
const bounds = projectRectToViewport(node.rect, projection);
return bounds ? { x: bounds.x, y: bounds.y, w: bounds.width, h: bounds.height } : null;
})()
: null,
}));
}
function frameOwnerNode(captured: CapturedViewModel, frame: CdpFrame): CapturedNode | undefined {
if (frame.ownerBackendNodeId === undefined) return undefined;
const parentNodes = frame.parentFrameId
? captured.frameNodes?.get(frame.parentFrameId)
: captured.nodes;
return parentNodes?.find((node) => node.backendNodeId === frame.ownerBackendNodeId);
}
async function captureMissingFrameDocuments(
cdp: CdpRunner,
tabId: number,
@@ -65,14 +59,20 @@ async function captureMissingFrameDocuments(
for (const frame of graph.frames) {
if (!frame.target.sessionId || captured.frameNodes.has(frame.frameId)) continue;
const owner = frameOwnerNode(captured, frame);
if (!owner?.rect || frame.ownerBackendNodeId === undefined) continue;
try {
const child = await captureViewModel(cdpRunnerForTarget(cdp, frame.target), tabId, {
signal,
});
const projection = await resolveFrameProjection(cdp, graph, frame.frameId);
if (!projection) continue;
let projection: GeometryProjection | null = null;
try {
projection = await resolveFrameProjection(cdp, graph, frame.frameId);
} catch (err) {
if (isAbortError(err)) throw err;
console.debug("[bsk observation] child frame geometry projection failed", {
frameId: frame.frameId,
err,
});
}
const rootFrameId = child.rootFrameId ?? frame.frameId;
const childFrames = child.frameNodes ?? new Map<string, CapturedNode[]>();
if (!childFrames.has(rootFrameId)) childFrames.set(rootFrameId, child.nodes);
@@ -85,11 +85,13 @@ async function captureMissingFrameDocuments(
for (const [childFrameId, nodes] of childFrames) {
captured.frameNodes.set(childFrameId, transformFrameNodes(nodes, projection));
}
captured.iframeNodes.set(
frame.ownerBackendNodeId,
captured.frameNodes.get(frame.frameId) ?? [],
);
captured.frameOwnerBackendNodeIds.set(frame.frameId, frame.ownerBackendNodeId);
if (frame.ownerBackendNodeId !== undefined) {
captured.iframeNodes.set(
frame.ownerBackendNodeId,
captured.frameNodes.get(frame.frameId) ?? [],
);
captured.frameOwnerBackendNodeIds.set(frame.frameId, frame.ownerBackendNodeId);
}
if (frame.parentFrameId) captured.frameParentIds.set(frame.frameId, frame.parentFrameId);
for (const [childFrameId, ownerBackendNodeId] of child.frameOwnerBackendNodeIds ?? []) {
captured.frameOwnerBackendNodeIds.set(childFrameId, ownerBackendNodeId);
@@ -0,0 +1,37 @@
import type { CdpRunner } from "../shared";
function abortError(): Error {
const error = new Error("hover perception aborted");
error.name = "AbortError";
return error;
}
export async function waitForHover(ms: number, signal?: AbortSignal): Promise<void> {
if (signal?.aborted) throw abortError();
await new Promise<void>((resolve, reject) => {
const timer = setTimeout(() => {
signal?.removeEventListener("abort", onAbort);
resolve();
}, ms);
const onAbort = () => {
clearTimeout(timer);
signal?.removeEventListener("abort", onAbort);
reject(abortError());
};
signal?.addEventListener("abort", onAbort, { once: true });
});
}
export async function clearHover(cdp: CdpRunner, tabId: number): Promise<void> {
await cdp
.send(tabId, "Input.dispatchMouseEvent", {
type: "mouseMoved",
x: -10,
y: -10,
})
.catch(() =>
cdp.send(tabId, "Input.dispatchMouseEvent", { type: "mouseMoved", x: 0, y: 0 }).catch(() => {
// Best effort cleanup.
}),
);
}
@@ -0,0 +1,296 @@
import { isVomReferenceNode } from "@browser-skill/vom";
import { type CdpRunner, sendToCdpTarget } from "../shared";
import type { FrameDocument, FrameOwnedAxNode } from "./frame-document";
import { clearHover, waitForHover } from "./hover-perception";
import { stableIdentifierName } from "./semantic-graph/name-evidence";
import { frameBackendKey, type ResolvedSemanticGraph } from "./semantic-graph/types";
const MAX_TOOLTIP_CANDIDATES = 8;
const MAX_TOOLTIP_PROBE_MS = 2_400;
const TOOLTIP_SETTLE_MS = 250;
const CACHE_TTL_MS = 5 * 60_000;
const MAX_CACHE_ENTRIES = 500;
const FORM_CONTROL_TAGS = new Set(["input", "select", "textarea"]);
interface TooltipCacheEntry {
name: string;
expiresAt: number;
}
interface TooltipCandidate<T extends FrameOwnedAxNode> {
document: FrameDocument<T>;
backendNodeId: number;
x: number;
y: number;
stableIdentifier: boolean;
}
export interface TooltipNameProbeOptions {
signal?: AbortSignal;
hoverProbeBypassOverlay?: (tabId: number, enabled: boolean) => Promise<void>;
}
const tooltipCache = new Map<string, TooltipCacheEntry>();
function cacheKey<T extends FrameOwnedAxNode>(
document: FrameDocument<T>,
backendNodeId: number,
): string {
return `${document.target.tabId}\u0000${document.frameId}\u0000${document.url ?? ""}\u0000${backendNodeId}`;
}
function cachedName<T extends FrameOwnedAxNode>(
document: FrameDocument<T>,
backendNodeId: number,
): string | undefined {
const key = cacheKey(document, backendNodeId);
const entry = tooltipCache.get(key);
if (!entry) return undefined;
if (entry.expiresAt <= Date.now()) {
tooltipCache.delete(key);
return undefined;
}
return entry.name;
}
function storeName<T extends FrameOwnedAxNode>(
document: FrameDocument<T>,
backendNodeId: number,
name: string,
): void {
if (tooltipCache.size >= MAX_CACHE_ENTRIES) {
const oldest = tooltipCache.keys().next().value;
if (oldest !== undefined) tooltipCache.delete(oldest);
}
tooltipCache.set(cacheKey(document, backendNodeId), {
name,
expiresAt: Date.now() + CACHE_TTL_MS,
});
}
function tooltipCollectorFunction(): string {
return `function() {
const ownRoot = this && typeof this.getRootNode === "function" ? this.getRootNode() : null;
const roots = ownRoot && ownRoot !== document ? [ownRoot, document] : [document];
const items = [];
const seen = new Set();
for (const root of roots) {
for (const el of Array.from(root.querySelectorAll("[role='tooltip']"))) {
const style = getComputedStyle(el);
const rect = el.getBoundingClientRect();
if (rect.width <= 0 || rect.height <= 0) continue;
if (style.display === "none" || style.visibility === "hidden" || style.opacity === "0") continue;
const text = String(el.getAttribute("aria-label") || el.getAttribute("title") || el.textContent || "")
.replace(/\\s+/g, " ")
.trim();
if (!text || text.length > 160 || seen.has(text)) continue;
seen.add(text);
items.push(text);
}
}
return items.slice(0, 20);
}`;
}
function cleanTooltipDiff(before: string[], after: string[]): string | undefined {
const existing = new Set(before.map((value) => value.toLowerCase()));
const added = [...new Set(after.map((value) => value.replace(/\s+/g, " ").trim()))].filter(
(value) => value && !existing.has(value.toLowerCase()),
);
return added.length === 1 ? added[0] : undefined;
}
function candidates<T extends FrameOwnedAxNode>(
documents: FrameDocument<T>[],
graph: ResolvedSemanticGraph,
): TooltipCandidate<T>[] {
const documentByFrame = new Map(documents.map((document) => [document.frameId, document]));
const byFrame = new Map<string, TooltipCandidate<T>[]>();
for (const node of graph.nodes.values()) {
if (
node.vom.name ||
FORM_CONTROL_TAGS.has(node.dom?.tag.toLowerCase() ?? "") ||
!node.referenceable ||
node.backendNodeId === undefined ||
!isVomReferenceNode({
...node.vom,
id: 0,
parentId: null,
backendNodeId: node.backendNodeId,
frameId: node.frameId,
referenceable: true,
})
) {
continue;
}
const document = documentByFrame.get(node.frameId);
const rect = node.dom?.rect;
if (!document || !rect || rect.w <= 0 || rect.h <= 0) continue;
if (
rect.x + rect.w <= 0 ||
rect.y + rect.h <= 0 ||
rect.x >= graph.viewport.width ||
rect.y >= graph.viewport.height
) {
continue;
}
if (
node.vom.disabled ||
node.vom.inert ||
node.vom.pointerEvents === "none" ||
node.vom.sensitive
) {
continue;
}
const frameCandidates = byFrame.get(node.frameId) ?? [];
frameCandidates.push({
document,
backendNodeId: node.backendNodeId,
x: (Math.max(0, rect.x) + Math.min(graph.viewport.width, rect.x + rect.w)) / 2,
y: (Math.max(0, rect.y) + Math.min(graph.viewport.height, rect.y + rect.h)) / 2,
stableIdentifier: stableIdentifierName(node.dom?.attrs ?? {}) !== undefined,
});
byFrame.set(node.frameId, frameCandidates);
}
const out: TooltipCandidate<T>[] = [];
for (const stableIdentifier of [true, false]) {
const queues = documents.map((document) =>
(byFrame.get(document.frameId) ?? []).filter(
(candidate) => candidate.stableIdentifier === stableIdentifier,
),
);
for (let index = 0; queues.some((queue) => index < queue.length); index += 1) {
for (const queue of queues) {
const candidate = queue[index];
if (candidate) out.push(candidate);
}
}
}
return out;
}
async function executionContextId<T extends FrameOwnedAxNode>(
cdp: CdpRunner,
document: FrameDocument<T>,
): Promise<number | undefined> {
const result = await sendToCdpTarget<{ executionContextId?: number }>(
cdp,
document.target,
"Page.createIsolatedWorld",
{
frameId: document.frameId,
worldName: "bsk-vom-tooltip",
grantUniveralAccess: false,
},
);
return result.executionContextId;
}
async function resolvedObjectId<T extends FrameOwnedAxNode>(
cdp: CdpRunner,
document: FrameDocument<T>,
contextId: number,
backendNodeId: number,
): Promise<string | undefined> {
const result = await sendToCdpTarget<{
object?: { objectId?: string };
}>(cdp, document.target, "DOM.resolveNode", {
backendNodeId,
executionContextId: contextId,
});
return result.object?.objectId;
}
async function tooltipTexts<T extends FrameOwnedAxNode>(
cdp: CdpRunner,
document: FrameDocument<T>,
objectId: string,
): Promise<string[]> {
const result = await sendToCdpTarget<{
result?: { value?: string[] };
}>(cdp, document.target, "Runtime.callFunctionOn", {
functionDeclaration: tooltipCollectorFunction(),
objectId,
returnByValue: true,
});
return result.result?.value ?? [];
}
async function releaseObject<T extends FrameOwnedAxNode>(
cdp: CdpRunner,
document: FrameDocument<T>,
objectId: string,
): Promise<void> {
await sendToCdpTarget(cdp, document.target, "Runtime.releaseObject", { objectId }).catch(
() => undefined,
);
}
export async function probeTooltipNames<T extends FrameOwnedAxNode>(
cdp: CdpRunner,
tabId: number,
documents: FrameDocument<T>[],
graph: ResolvedSemanticGraph,
options: TooltipNameProbeOptions = {},
): Promise<Map<string, string>> {
const names = new Map<string, string>();
const pending: TooltipCandidate<T>[] = [];
for (const candidate of candidates(documents, graph)) {
const cached = cachedName(candidate.document, candidate.backendNodeId);
if (cached) {
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 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 (Date.now() - started >= MAX_TOOLTIP_PROBE_MS) break;
let objectId: string | undefined;
try {
let contextId = contexts.get(candidate.document.frameId);
if (contextId === undefined) {
contextId = await executionContextId(cdp, candidate.document);
if (contextId === undefined) continue;
contexts.set(candidate.document.frameId, contextId);
}
objectId = await resolvedObjectId(
cdp,
candidate.document,
contextId,
candidate.backendNodeId,
);
if (!objectId) continue;
await clearHover(cdp, tabId);
const before = await tooltipTexts(cdp, candidate.document, objectId);
await cdp.send(tabId, "Input.dispatchMouseEvent", {
type: "mouseMoved",
x: candidate.x,
y: candidate.y,
});
await waitForHover(TOOLTIP_SETTLE_MS, options.signal);
const after = await tooltipTexts(cdp, candidate.document, objectId);
const name = cleanTooltipDiff(before, after);
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 {
if (objectId) await releaseObject(cdp, candidate.document, objectId);
await clearHover(cdp, tabId);
}
}
} finally {
await clearHover(cdp, tabId);
await options.hoverProbeBypassOverlay?.(tabId, false).catch(() => undefined);
}
return names;
}
@@ -1,5 +1,6 @@
import type { Rect, RenderedRef, VomResult } from "@browser-skill/vom";
import type { CdpTarget } from "@/browser-driver/frame-graph";
import { readRecordingDocumentIdentity } from "@/shared/recording-document-identity";
import type { CapturedSurfaceProbe } from "./capture";
import type { CapturedFrameDocument, FrameAxNode } from "./frame-capture";
@@ -9,6 +10,8 @@ export interface CaptureVomFrame {
target: CdpTarget;
parentFrameId?: string;
ownerBackendNodeId?: number;
/** BrowserSkill-generated identity for the recording agent in this Document. */
recordingDocumentId?: string;
}
/** Allowlisted geometry used to match a recorded action to a rendered ref. */
@@ -37,14 +40,22 @@ export interface CaptureVomObservationResult {
}
function projectFrames(documents: CapturedFrameDocument<FrameAxNode>[]): CaptureVomFrame[] {
return documents.map((document) => ({
frameId: document.frameId,
target: document.target,
...(document.parentFrameId ? { parentFrameId: document.parentFrameId } : {}),
...(document.ownerBackendNodeId !== undefined
? { ownerBackendNodeId: document.ownerBackendNodeId }
: {}),
}));
return documents.map((document) => {
let recordingDocumentId: string | undefined;
for (const node of document.domNodes) {
recordingDocumentId = readRecordingDocumentIdentity(node.attrs);
if (recordingDocumentId) break;
}
return {
frameId: document.frameId,
target: document.target,
...(document.parentFrameId ? { parentFrameId: document.parentFrameId } : {}),
...(document.ownerBackendNodeId !== undefined
? { ownerBackendNodeId: document.ownerBackendNodeId }
: {}),
...(recordingDocumentId ? { recordingDocumentId } : {}),
};
});
}
function projectMatchNodes(documents: CapturedFrameDocument<FrameAxNode>[]): CaptureVomMatchNode[] {
@@ -0,0 +1,154 @@
import type { Viewport } from "@browser-skill/vom";
import type { FrameDocument } from "../frame-document";
import {
frameAxKey,
frameBackendKey,
type SemanticAxNode,
type SemanticFrame,
type SemanticGraph,
type SemanticGraphNode,
type SemanticNodeId,
} from "./types";
export interface BuildSemanticGraphInput {
documents: FrameDocument<SemanticAxNode>[];
viewport: Viewport;
rootFrameId?: string;
excludedBackendNodeIds?: ReadonlySet<number>;
}
function domNodeId(frameId: string, backendNodeId: number): SemanticNodeId {
return JSON.stringify([frameId, "dom", backendNodeId]);
}
function axNodeId(frameId: string, nodeId: string): SemanticNodeId {
return JSON.stringify([frameId, "ax", nodeId]);
}
export function buildSemanticGraph(input: BuildSemanticGraphInput): SemanticGraph {
const rootFrameId =
input.rootFrameId ??
input.documents.find((document) => !document.parentFrameId)?.frameId ??
input.documents[0]?.frameId ??
"root";
const frames = new Map<string, SemanticFrame>();
const nodes = new Map<SemanticNodeId, SemanticGraphNode>();
const nodeIdsByFrame = new Map<string, SemanticNodeId[]>();
const nodeByFrameBackend = new Map<string, SemanticNodeId>();
const nodeByFrameAx = new Map<string, SemanticNodeId>();
for (const [frameOrder, document] of input.documents.entries()) {
frames.set(document.frameId, {
frameId: document.frameId,
...(document.parentFrameId ? { parentFrameId: document.parentFrameId } : {}),
...(document.ownerBackendNodeId !== undefined
? { ownerBackendNodeId: document.ownerBackendNodeId }
: {}),
contextScopeId: document.contextScopeId,
target: document.target,
...(document.url ? { url: document.url } : {}),
order: frameOrder,
});
const frameNodeIds: SemanticNodeId[] = [];
nodeIdsByFrame.set(document.frameId, frameNodeIds);
for (const [index, dom] of document.domNodes.entries()) {
const id = domNodeId(document.frameId, dom.backendNodeId);
const node: SemanticGraphNode = {
id,
frameId: document.frameId,
backendNodeId: dom.backendNodeId,
dom,
sourceOrder: index,
excluded:
document.frameId === rootFrameId &&
input.excludedBackendNodeIds?.has(dom.backendNodeId) === true,
};
nodes.set(id, node);
frameNodeIds.push(id);
nodeByFrameBackend.set(frameBackendKey(document.frameId, dom.backendNodeId), id);
}
for (const [index, ax] of document.axNodes.entries()) {
const joinedId =
typeof ax.backendDOMNodeId === "number"
? nodeByFrameBackend.get(frameBackendKey(document.frameId, ax.backendDOMNodeId))
: undefined;
const id = joinedId ?? axNodeId(document.frameId, ax.nodeId);
const existing = nodes.get(id);
if (existing) {
existing.ax = ax;
existing.axNodeId = ax.nodeId;
} else {
nodes.set(id, {
id,
frameId: document.frameId,
...(typeof ax.backendDOMNodeId === "number"
? { backendNodeId: ax.backendDOMNodeId }
: {}),
axNodeId: ax.nodeId,
ax,
sourceOrder: document.domNodes.length + index,
excluded:
document.frameId === rootFrameId &&
typeof ax.backendDOMNodeId === "number" &&
input.excludedBackendNodeIds?.has(ax.backendDOMNodeId) === true,
});
frameNodeIds.push(id);
}
nodeByFrameAx.set(frameAxKey(document.frameId, ax.nodeId), id);
}
}
for (const document of input.documents) {
for (const dom of document.domNodes) {
const id = nodeByFrameBackend.get(frameBackendKey(document.frameId, dom.backendNodeId));
const node = id ? nodes.get(id) : undefined;
if (!node || dom.parentBackendNodeId === null) continue;
node.domParentId = nodeByFrameBackend.get(
frameBackendKey(document.frameId, dom.parentBackendNodeId),
);
}
for (const ax of document.axNodes) {
const id = nodeByFrameAx.get(frameAxKey(document.frameId, ax.nodeId));
const node = id ? nodes.get(id) : undefined;
if (!node || !ax.parentId) continue;
node.axParentId = nodeByFrameAx.get(frameAxKey(document.frameId, ax.parentId));
}
}
const axChildren = new Map<SemanticNodeId, SemanticNodeId[]>();
const excludedQueue: SemanticNodeId[] = [];
for (const node of nodes.values()) {
if (node.axParentId) {
const children = axChildren.get(node.axParentId) ?? [];
children.push(node.id);
axChildren.set(node.axParentId, children);
}
if (node.excluded) excludedQueue.push(node.id);
}
for (let index = 0; index < excludedQueue.length; index += 1) {
for (const childId of axChildren.get(excludedQueue[index]) ?? []) {
const child = nodes.get(childId);
if (!child || child.excluded) continue;
child.excluded = true;
excludedQueue.push(childId);
}
}
for (const frame of frames.values()) {
if (!frame.parentFrameId || frame.ownerBackendNodeId === undefined) continue;
frame.ownerNodeId = nodeByFrameBackend.get(
frameBackendKey(frame.parentFrameId, frame.ownerBackendNodeId),
);
}
return {
viewport: input.viewport,
rootFrameId,
frames,
nodes,
nodeIdsByFrame,
nodeByFrameBackend,
};
}
@@ -0,0 +1,59 @@
import type { Viewport, VomScene } from "@browser-skill/vom";
import type { FrameDocument } from "../frame-document";
import { buildSemanticGraph as buildGraph } from "./build";
import { projectSemanticGraph as projectGraph } from "./project";
import { resolveSemanticGraph as resolveGraph } from "./resolve";
import { normalizeSemanticStructure as normalizeStructure } from "./structure";
import type {
ResolvedSemanticGraph,
SemanticAxNode,
SemanticGraph,
StructuredSemanticGraph,
} from "./types";
export type { SemanticAxNode } from "./types";
export interface SemanticVomInput {
documents: FrameDocument<SemanticAxNode>[];
viewport: Viewport;
rootFrameId?: string;
excludedBackendNodeIds?: ReadonlySet<number>;
}
export function buildSemanticGraph(input: SemanticVomInput): SemanticGraph {
return buildGraph(input);
}
export function resolveSemanticGraph(
graph: SemanticGraph,
options: {
supplementalNames?: ReadonlyMap<string, string>;
identifierFallback?: boolean;
} = {},
): ResolvedSemanticGraph {
return resolveGraph(graph, options);
}
export function normalizeSemanticStructure(graph: ResolvedSemanticGraph): StructuredSemanticGraph {
return normalizeStructure(graph);
}
export function projectSemanticGraph(graph: StructuredSemanticGraph): VomScene {
return projectGraph(graph);
}
export function buildSemanticVomScene(
input: SemanticVomInput & {
supplementalNames?: ReadonlyMap<string, string>;
identifierFallback?: boolean;
},
): VomScene {
return projectSemanticGraph(
normalizeSemanticStructure(
resolveSemanticGraph(buildSemanticGraph(input), {
supplementalNames: input.supplementalNames,
identifierFallback: input.identifierFallback,
}),
),
);
}
@@ -0,0 +1,34 @@
const IDENTIFIER_ATTRS = ["data-testid", "data-test-id", "id"] as const;
const GENERIC_TOKENS = new Set(["btn", "button", "control", "icon", "item", "test", "tool"]);
const GENERATED_PREFIXES = new Set(["ember", "headlessui", "mui", "radix", "react", "rc"]);
function identifierTokens(value: string): string[] {
return value
.replace(/([a-z0-9])([A-Z])/g, "$1 $2")
.split(/[^a-zA-Z0-9]+/)
.map((token) => token.toLowerCase())
.filter(Boolean);
}
function readableIdentifier(value: string | undefined): string | undefined {
const raw = value?.trim();
if (!raw || raw.length < 3 || raw.length > 80) return undefined;
if (/^[a-f0-9]{8,}$/i.test(raw)) return undefined;
const tokens = identifierTokens(raw);
if (GENERATED_PREFIXES.has(tokens[0])) return undefined;
while (tokens.length > 1 && GENERIC_TOKENS.has(tokens[0])) tokens.shift();
while (tokens.length > 1 && GENERIC_TOKENS.has(tokens[tokens.length - 1])) tokens.pop();
if (tokens.length === 0 || !tokens.some((token) => /[a-z]{3,}/.test(token))) return undefined;
return tokens
.map((token, index) => (index === 0 ? `${token[0].toUpperCase()}${token.slice(1)}` : token))
.join(" ");
}
export function stableIdentifierName(attrs: Readonly<Record<string, string>>): string | undefined {
for (const name of IDENTIFIER_ATTRS) {
const label = readableIdentifier(attrs[name]);
if (label) return label;
}
return undefined;
}
@@ -0,0 +1,85 @@
import type { VomNode, VomScene } from "@browser-skill/vom";
import type { SemanticNodeId, StructuredSemanticGraph, StructuredSemanticNode } from "./types";
function retained(node: StructuredSemanticNode): boolean {
return node.structure.kind === "keep" || node.structure.kind === "inferred-group";
}
function connectedFrameIds(graph: StructuredSemanticGraph): Set<string> {
const connected = new Set<string>([graph.rootFrameId]);
const childrenByParent = new Map<string, string[]>();
for (const frame of graph.frames.values()) {
if (!frame.parentFrameId) continue;
const children = childrenByParent.get(frame.parentFrameId) ?? [];
children.push(frame.frameId);
childrenByParent.set(frame.parentFrameId, children);
}
const queue = [graph.rootFrameId];
for (let index = 0; index < queue.length; index += 1) {
for (const frameId of childrenByParent.get(queue[index]) ?? []) {
if (connected.has(frameId)) continue;
const frame = graph.frames.get(frameId);
if (!frame?.ownerNodeId) continue;
const owner = graph.nodes.get(frame.ownerNodeId);
if (!owner || !retained(owner)) continue;
connected.add(frame.frameId);
queue.push(frame.frameId);
}
}
return connected;
}
function domAncestors(
graph: StructuredSemanticGraph,
node: StructuredSemanticNode,
numericId: ReadonlyMap<SemanticNodeId, number>,
): number[] {
const ancestors: number[] = [];
const seen = new Set<SemanticNodeId>();
let current = node.domParentId;
while (current && !seen.has(current)) {
seen.add(current);
const id = numericId.get(current);
if (id !== undefined) ancestors.push(id);
current = graph.nodes.get(current)?.domParentId;
}
return ancestors;
}
export function projectSemanticGraph(graph: StructuredSemanticGraph): VomScene {
const connectedFrames = connectedFrameIds(graph);
const orderedNodes = [...graph.frames.values()]
.filter((frame) => connectedFrames.has(frame.frameId))
.sort((a, b) => a.order - b.order)
.flatMap((frame) =>
(graph.nodeIdsByFrame.get(frame.frameId) ?? [])
.map((id) => graph.nodes.get(id))
.filter((node): node is StructuredSemanticNode => node !== undefined && retained(node))
.sort((a, b) => a.sourceOrder - b.sourceOrder),
);
const numericId = new Map<SemanticNodeId, number>();
orderedNodes.forEach((node, index) => numericId.set(node.id, index + 1));
const nodes: VomNode[] = orderedNodes.map((node) => ({
...node.vom,
id: numericId.get(node.id) as number,
parentId: node.semanticParentId ? (numericId.get(node.semanticParentId) ?? null) : null,
...(node.backendNodeId !== undefined ? { backendNodeId: node.backendNodeId } : {}),
frameId: node.frameId,
contextScopeId: graph.frames.get(node.frameId)?.contextScopeId ?? node.frameId,
referenceable: node.referenceable,
...(node.dom
? {
domParentId:
node.domParentId === undefined ? null : (numericId.get(node.domParentId) ?? null),
domAncestorIds: domAncestors(graph, node, numericId),
}
: {}),
}));
return {
viewport: graph.viewport,
nodes,
rootFrameId: graph.rootFrameId,
};
}
@@ -0,0 +1,521 @@
import type { VomNode } from "@browser-skill/vom";
import { stableIdentifierName } from "./name-evidence";
import type {
ResolvedSemanticGraph,
ResolvedSemanticNode,
SemanticAxNode,
SemanticGraph,
SemanticGraphNode,
SemanticNodeId,
} from "./types";
import { frameBackendKey } from "./types";
const FORM_TAGS = new Set(["input", "textarea", "select"]);
const NATIVE_CONTROL_TAGS = new Set(["button", "input", "select", "textarea"]);
const TEXT_AX_ROLES = new Set(["statictext", "inlinetextbox", "labeltext", "text"]);
const INTERACTIVE_ROLES = new Set([
"button",
"checkbox",
"combobox",
"link",
"listbox",
"menuitem",
"menuitemcheckbox",
"menuitemradio",
"option",
"radio",
"searchbox",
"slider",
"spinbutton",
"switch",
"tab",
"textbox",
"treeitem",
]);
const SENSITIVE_INPUT_TYPES = new Set([
"password",
"credit-card",
"one-time-code",
"current-password",
"new-password",
]);
function clean(value: string | undefined): string | undefined {
const normalized = value?.replace(/\s+/g, " ").trim();
return normalized || undefined;
}
function axValue(field?: { value?: string | number | boolean }): string | undefined {
return field?.value === undefined ? undefined : clean(String(field.value));
}
function axProperty(node: SemanticAxNode | undefined, name: string): string | undefined {
return axValue(node?.properties?.find((property) => property.name === name)?.value);
}
function normalizeRole(value: string | undefined): string | undefined {
const role = clean(value)?.toLowerCase();
return role || undefined;
}
function nativeRole(tag: string, attrs: Record<string, string>): string | undefined {
if (tag === "button") return "button";
if (tag === "textarea") return "textbox";
if (tag === "select") return attrs.multiple === undefined ? "combobox" : "listbox";
if (tag === "a" && attrs.href) return "link";
if (tag === "iframe") return "Iframe";
if (tag === "dialog") return "dialog";
if (tag === "img") return "img";
if (tag === "table") return "table";
if (tag === "tr") return "row";
if (tag === "th") return "columnheader";
if (tag === "td") return "cell";
if (tag === "ul" || tag === "ol") return "list";
if (tag === "li") return "listitem";
if (/^h[1-6]$/.test(tag)) return "heading";
if (tag === "nav") return "navigation";
if (tag === "main") return "main";
if (tag === "form") return "form";
if (tag === "input") {
const type = (attrs.type ?? "text").toLowerCase();
if (type === "hidden") return undefined;
if (type === "checkbox") return "checkbox";
if (type === "radio") return "radio";
if (type === "range") return "slider";
if (type === "number") return "spinbutton";
if (type === "search") return "searchbox";
if (["button", "submit", "reset", "image"].includes(type)) return "button";
return "textbox";
}
return undefined;
}
function domRole(tag: string, attrs: Record<string, string>): string | undefined {
return normalizeRole(attrs.role) ?? nativeRole(tag, attrs);
}
function resolvedRole(node: SemanticGraphNode): {
value?: string;
source: ResolvedSemanticNode["roleSource"];
} {
const rawAxRole = clean(axValue(node.ax?.role));
const axRole = normalizeRole(rawAxRole);
const attrs = node.dom?.attrs ?? {};
const explicitDomRole = normalizeRole(attrs.role);
const nativeDomRole = nativeRole(node.dom?.tag.toLowerCase() ?? "", attrs);
const usableAxRole = node.ax?.ignored !== true;
if (usableAxRole && axRole && !["generic", "none", "presentation"].includes(axRole)) {
return { value: rawAxRole, source: "ax" };
}
if (explicitDomRole) return { value: explicitDomRole, source: "dom-explicit" };
if (nativeDomRole) return { value: nativeDomRole, source: "dom-native" };
if (rawAxRole) {
return { value: rawAxRole, source: usableAxRole ? "ax" : "ax-ignored" };
}
return { source: "none" };
}
function sensitive(node: SemanticGraphNode): boolean {
const attrs = node.dom?.attrs ?? {};
const type = (attrs.type ?? "").toLowerCase();
const autocomplete = (attrs.autocomplete ?? "").toLowerCase();
return (
type === "password" ||
autocomplete.startsWith("cc-") ||
SENSITIVE_INPUT_TYPES.has(autocomplete) ||
SENSITIVE_INPUT_TYPES.has(axProperty(node.ax, "inputType") ?? "")
);
}
function inputState(
node: SemanticGraphNode,
role: string | undefined,
value: string | undefined,
isSensitive: boolean,
): VomNode["inputState"] {
if (node.dom?.formState) return node.dom.formState;
const tag = node.dom?.tag.toLowerCase() ?? "";
if (!FORM_TAGS.has(tag) && !["textbox", "searchbox"].includes(role ?? "")) return undefined;
if (isSensitive) return value ? "filled" : "empty";
if (node.dom?.formValue !== undefined) {
if (node.dom.formValue === "") return "empty";
return node.dom.formValue === (node.dom.formDefaultValue ?? "") ? "default" : "filled";
}
return value ? "filled" : "empty";
}
function externalHrefHost(
href: string | undefined,
pageUrl: string | undefined,
): string | undefined {
if (!href || !pageUrl) return undefined;
try {
const page = new URL(pageUrl);
const target = new URL(href, page);
return target.origin === page.origin ? undefined : target.hostname;
} catch {
return undefined;
}
}
interface ResolveIndexes {
domChildren: Map<SemanticNodeId, SemanticNodeId[]>;
axChildren: Map<SemanticNodeId, SemanticNodeId[]>;
domIdIndex: Map<string, SemanticNodeId>;
labelForIndex: Map<string, SemanticNodeId[]>;
domText: (nodeId: SemanticNodeId) => string | undefined;
axText: (nodeId: SemanticNodeId) => string | undefined;
hasNativeDescendant: (nodeId: SemanticNodeId) => boolean;
}
function buildIndexes(graph: SemanticGraph): ResolveIndexes {
const domChildren = new Map<SemanticNodeId, SemanticNodeId[]>();
const axChildren = new Map<SemanticNodeId, SemanticNodeId[]>();
const domIdIndex = new Map<string, SemanticNodeId>();
const labelForIndex = new Map<string, SemanticNodeId[]>();
for (const node of graph.nodes.values()) {
if (node.domParentId) {
const children = domChildren.get(node.domParentId) ?? [];
children.push(node.id);
domChildren.set(node.domParentId, children);
}
if (node.axParentId) {
const children = axChildren.get(node.axParentId) ?? [];
children.push(node.id);
axChildren.set(node.axParentId, children);
}
const id = clean(node.dom?.attrs.id);
if (id) domIdIndex.set(`${node.frameId}\u0000${id}`, node.id);
const targetId = clean(node.dom?.attrs.for);
if (node.dom?.tag.toLowerCase() === "label" && targetId) {
const key = `${node.frameId}\u0000${targetId}`;
const labels = labelForIndex.get(key) ?? [];
labels.push(node.id);
labelForIndex.set(key, labels);
}
}
for (const [parentId, children] of axChildren) {
const order = graph.nodes.get(parentId)?.ax?.childIds ?? [];
const orderByAxId = new Map(order.map((nodeId, index) => [nodeId, index]));
children.sort((a, b) => {
const aIndex = orderByAxId.get(graph.nodes.get(a)?.axNodeId ?? "") ?? -1;
const bIndex = orderByAxId.get(graph.nodes.get(b)?.axNodeId ?? "") ?? -1;
return (
(aIndex < 0 ? Number.MAX_SAFE_INTEGER : aIndex) -
(bIndex < 0 ? Number.MAX_SAFE_INTEGER : bIndex)
);
});
}
const textMemo = new Map<SemanticNodeId, string | undefined>();
const resolvingText = new Set<SemanticNodeId>();
const domText = (nodeId: SemanticNodeId): string | undefined => {
if (textMemo.has(nodeId)) return textMemo.get(nodeId);
if (resolvingText.has(nodeId)) return undefined;
resolvingText.add(nodeId);
const node = graph.nodes.get(nodeId);
const parts: string[] = [];
const own = clean(node?.dom?.textContent);
if (own) parts.push(own);
for (const childId of domChildren.get(nodeId) ?? []) {
const child = graph.nodes.get(childId);
const childRole = child
? domRole(child.dom?.tag.toLowerCase() ?? "", child.dom?.attrs ?? {})
: undefined;
if (childRole && INTERACTIVE_ROLES.has(childRole.toLowerCase())) continue;
const childText = domText(childId);
if (childText) parts.push(childText);
if (parts.join(" ").length >= 240) break;
}
resolvingText.delete(nodeId);
const result = clean(parts.join(" ").slice(0, 240));
textMemo.set(nodeId, result);
return result;
};
const axTextMemo = new Map<SemanticNodeId, string | undefined>();
const resolvingAxText = new Set<SemanticNodeId>();
const axText = (nodeId: SemanticNodeId): string | undefined => {
if (axTextMemo.has(nodeId)) return axTextMemo.get(nodeId);
if (resolvingAxText.has(nodeId)) return undefined;
resolvingAxText.add(nodeId);
const complete = (value: string | undefined): string | undefined => {
resolvingAxText.delete(nodeId);
axTextMemo.set(nodeId, value);
return value;
};
const node = graph.nodes.get(nodeId);
const role = normalizeRole(axValue(node?.ax?.role)) ?? "";
if (TEXT_AX_ROLES.has(role)) {
const own = clean(axValue(node?.ax?.name) ?? axValue(node?.ax?.value));
if (own) return complete(own);
}
const parts: string[] = [];
for (const childId of axChildren.get(nodeId) ?? []) {
const child = graph.nodes.get(childId);
const childRole = normalizeRole(axValue(child?.ax?.role)) ?? "";
if (INTERACTIVE_ROLES.has(childRole)) continue;
const childText = axText(childId);
if (childText) parts.push(childText);
if (parts.join(" ").length >= 240) break;
}
return complete(clean(parts.join(" ").slice(0, 240)));
};
const nativeMemo = new Map<SemanticNodeId, boolean>();
const hasNativeDescendant = (nodeId: SemanticNodeId): boolean => {
const cached = nativeMemo.get(nodeId);
if (cached !== undefined) return cached;
nativeMemo.set(nodeId, false);
for (const childId of domChildren.get(nodeId) ?? []) {
const child = graph.nodes.get(childId);
const childRole =
(child?.ax?.ignored === true ? undefined : normalizeRole(axValue(child?.ax?.role))) ??
(child ? domRole(child.dom?.tag.toLowerCase() ?? "", child.dom?.attrs ?? {}) : undefined);
if (
(child?.dom && NATIVE_CONTROL_TAGS.has(child.dom.tag.toLowerCase())) ||
INTERACTIVE_ROLES.has(childRole ?? "") ||
hasNativeDescendant(childId)
) {
nativeMemo.set(nodeId, true);
return true;
}
}
return false;
};
return {
domChildren,
axChildren,
domIdIndex,
labelForIndex,
domText,
axText,
hasNativeDescendant,
};
}
function ariaLabelledText(node: SemanticGraphNode, indexes: ResolveIndexes): string | undefined {
const ids = (node.dom?.attrs["aria-labelledby"] ?? "").split(/\s+/).filter(Boolean);
const parts = ids.flatMap((id) => {
const target = indexes.domIdIndex.get(`${node.frameId}\u0000${id}`);
return target ? (indexes.domText(target) ?? []) : [];
});
return clean(parts.join(" "));
}
function nativeLabelText(
node: SemanticGraphNode,
graph: SemanticGraph,
indexes: ResolveIndexes,
): string | undefined {
const id = clean(node.dom?.attrs.id);
if (id) {
const labels = indexes.labelForIndex.get(`${node.frameId}\u0000${id}`) ?? [];
const explicit = clean(
labels
.map((labelId) => indexes.domText(labelId))
.filter(Boolean)
.join(" "),
);
if (explicit) return explicit;
}
let parentId = node.domParentId;
let guard = 0;
while (parentId && guard <= graph.nodes.size) {
const parent = graph.nodes.get(parentId);
if (!parent) break;
if (parent.dom?.tag.toLowerCase() === "label") return indexes.domText(parentId);
parentId = parent.domParentId;
guard += 1;
}
return undefined;
}
function precedingAxText(
node: SemanticGraphNode,
graph: SemanticGraph,
indexes: ResolveIndexes,
): string | undefined {
if (!node.axParentId) return undefined;
const siblings = indexes.axChildren.get(node.axParentId) ?? [];
const index = siblings.indexOf(node.id);
for (let offset = 1; offset <= 4 && index - offset >= 0; offset += 1) {
const sibling = graph.nodes.get(siblings[index - offset]);
const role = normalizeRole(axValue(sibling?.ax?.role)) ?? "";
if (INTERACTIVE_ROLES.has(role)) break;
if (!TEXT_AX_ROLES.has(role) && role !== "generic") continue;
const text = clean(axValue(sibling?.ax?.name) ?? axValue(sibling?.ax?.value))?.replace(
/[::]\s*$/,
"",
);
if (text && text.length <= 80) return text;
}
return undefined;
}
function precedingDomText(
node: SemanticGraphNode,
graph: SemanticGraph,
indexes: ResolveIndexes,
): string | undefined {
if (!node.domParentId) return undefined;
const siblings = indexes.domChildren.get(node.domParentId) ?? [];
const index = siblings.indexOf(node.id);
for (let offset = 1; offset <= 4 && index - offset >= 0; offset += 1) {
const sibling = graph.nodes.get(siblings[index - offset]);
const siblingRole = sibling
? domRole(sibling.dom?.tag.toLowerCase() ?? "", sibling.dom?.attrs ?? {})
: undefined;
if (siblingRole && INTERACTIVE_ROLES.has(siblingRole.toLowerCase())) break;
const text = indexes.domText(siblings[index - offset])?.replace(/[::]\s*$/, "");
if (text && text.length <= 80) return text;
}
return undefined;
}
function resolvedName(
node: SemanticGraphNode,
role: string | undefined,
graph: SemanticGraph,
indexes: ResolveIndexes,
supplementalNames: ReadonlyMap<string, string>,
identifierFallback: boolean,
): string | undefined {
const attrs = node.dom?.attrs ?? {};
const formControl = FORM_TAGS.has(node.dom?.tag.toLowerCase() ?? "");
const descendantText = indexes.domText(node.id);
const interactive =
INTERACTIVE_ROLES.has(role?.toLowerCase() ?? "") ||
NATIVE_CONTROL_TAGS.has(node.dom?.tag.toLowerCase() ?? "");
const candidates = [
node.ax?.ignored === true ? undefined : axValue(node.ax?.name),
ariaLabelledText(node, indexes),
clean(attrs["aria-label"]),
nativeLabelText(node, graph, indexes),
clean(attrs.title),
clean(attrs.alt),
node.ax?.ignored === true ? undefined : indexes.axText(node.id),
descendantText,
formControl ? precedingAxText(node, graph, indexes) : undefined,
formControl ? precedingDomText(node, graph, indexes) : undefined,
formControl ? clean(node.dom?.formPlaceholder ?? attrs.placeholder) : undefined,
formControl || !interactive || node.backendNodeId === undefined
? undefined
: clean(supplementalNames.get(frameBackendKey(node.frameId, node.backendNodeId))),
identifierFallback && interactive ? stableIdentifierName(attrs) : undefined,
];
let name = candidates.find((candidate) => candidate !== undefined);
const hasPopup = axProperty(node.ax, "hasPopup");
if (name && hasPopup && hasPopup !== "false") {
name =
axProperty(node.ax, "expanded") === "true" ? `${name} [expanded]` : `${name} [has-submenu]`;
}
if (!name && role?.toLowerCase() === "iframe") return clean(attrs.id);
return name;
}
function insideNative(node: SemanticGraphNode, graph: SemanticGraph): boolean {
let parentId = node.domParentId;
let guard = 0;
while (parentId && guard <= graph.nodes.size) {
const parent = graph.nodes.get(parentId);
if (!parent) break;
if (parent.dom && NATIVE_CONTROL_TAGS.has(parent.dom.tag.toLowerCase())) return true;
parentId = parent.domParentId;
guard += 1;
}
return false;
}
function nearbyText(
node: SemanticGraphNode,
graph: SemanticGraph,
indexes: ResolveIndexes,
): string | undefined {
if (!node.domParentId) return undefined;
const siblings = indexes.domChildren.get(node.domParentId) ?? [];
const index = siblings.indexOf(node.id);
if (index < 0) return undefined;
const texts = siblings
.slice(Math.max(0, index - 3), index + 4)
.filter((id) => id !== node.id)
.flatMap((id) => indexes.domText(id) ?? [])
.slice(0, 3);
return clean(texts.join(" "));
}
export function resolveSemanticGraph(
graph: SemanticGraph,
options: {
supplementalNames?: ReadonlyMap<string, string>;
identifierFallback?: boolean;
} = {},
): ResolvedSemanticGraph {
const indexes = buildIndexes(graph);
const supplementalNames = options.supplementalNames ?? new Map<string, string>();
const identifierFallback = options.identifierFallback ?? true;
const nodes = new Map<SemanticNodeId, ResolvedSemanticNode>();
for (const node of graph.nodes.values()) {
const roleResolution = resolvedRole(node);
const role = roleResolution.value;
const name = resolvedName(node, role, graph, indexes, supplementalNames, identifierFallback);
const isSensitive = sensitive(node);
const sourceValue = node.dom?.formValue ?? axValue(node.ax?.value);
const value = isSensitive ? undefined : sourceValue;
const attrs = { ...(node.dom?.attrs ?? {}) };
if (isSensitive) delete attrs.value;
const frame = graph.frames.get(node.frameId);
const tag = node.dom?.tag.toLowerCase() ?? "";
const vom: ResolvedSemanticNode["vom"] = {
tag,
rect: node.dom?.rect ?? null,
paintOrder: node.dom?.paintOrder ?? 0,
position: node.dom?.position || "static",
pointerEvents: node.dom?.pointerEvents || "auto",
...(node.dom?.cursor ? { cursor: node.dom.cursor } : {}),
...(role ? { role } : {}),
...(name ? { name } : {}),
...(value !== undefined ? { value } : {}),
...(clean(node.dom?.formPlaceholder ?? attrs.placeholder)
? { placeholder: clean(node.dom?.formPlaceholder ?? attrs.placeholder) }
: {}),
attrs,
...(clean(node.dom?.textContent) ? { text: clean(node.dom?.textContent) } : {}),
...(nearbyText(node, graph, indexes) ? { nearbyText: nearbyText(node, graph, indexes) } : {}),
inputState: inputState(node, role, sourceValue, isSensitive),
sensitive: isSensitive,
modal:
["dialog", "alertdialog"].includes(role?.toLowerCase() ?? "") ||
tag === "dialog" ||
(attrs["aria-modal"] ?? "").toLowerCase() === "true",
disabled:
Object.prototype.hasOwnProperty.call(attrs, "disabled") ||
(attrs["aria-disabled"] ?? "").toLowerCase() === "true",
inert: Object.prototype.hasOwnProperty.call(attrs, "inert"),
hasNativeDescendant: indexes.hasNativeDescendant(node.id),
insideNative: insideNative(node, graph),
...(role?.toLowerCase() === "link" ? { href: externalHrefHost(attrs.href, frame?.url) } : {}),
};
const selected = node.ax?.ignored === true ? undefined : axProperty(node.ax, "selected");
const expanded = node.ax?.ignored === true ? undefined : axProperty(node.ax, "expanded");
const controls = node.ax?.ignored === true ? undefined : axProperty(node.ax, "controls");
if (selected === "true" && !vom.attrs?.["aria-selected"]) {
vom.attrs = { ...(vom.attrs ?? {}), "aria-selected": "true" };
}
if (expanded === "true" && !vom.attrs?.["aria-expanded"]) {
vom.attrs = { ...(vom.attrs ?? {}), "aria-expanded": "true" };
}
if (controls && !vom.attrs?.["aria-controls"]) {
vom.attrs = { ...(vom.attrs ?? {}), "aria-controls": controls };
}
nodes.set(node.id, {
...node,
vom,
referenceable: node.backendNodeId !== undefined,
roleSource: roleResolution.source,
});
}
return { ...graph, nodes };
}
@@ -0,0 +1,456 @@
import {
applyVomInteractionRecovery,
isVomReferenceNode,
isVomStructuralRole,
type VomNode,
} from "@browser-skill/vom";
import type {
ResolvedSemanticGraph,
ResolvedSemanticNode,
SemanticNodeId,
StructureDecision,
StructuredSemanticGraph,
StructuredSemanticNode,
} from "./types";
const TRANSPARENT_ROLES = new Set(["", "generic", "none", "presentation", "inlinetextbox"]);
const ROOT_ROLES = new Set(["rootwebarea", "webarea"]);
const CONTENT_BOUNDARY_ROLES = new Set([
"article",
"dialog",
"document",
"feed",
"form",
"grid",
"heading",
"iframe",
"list",
"main",
"navigation",
"paragraph",
"region",
"section",
"table",
"tree",
]);
function hasMeaningfulRelations(node: ResolvedSemanticNode): boolean {
const attrs = node.vom.attrs ?? {};
return ["aria-controls", "aria-labelledby", "aria-owns", "aria-describedby"].some((name) =>
attrs[name]?.trim(),
);
}
function explicitStructural(node: ResolvedSemanticNode): boolean {
return (
isVomStructuralRole(node.vom.role) &&
node.roleSource !== "none" &&
node.roleSource !== "ax-ignored"
);
}
function hasUsableAxSemantics(node: ResolvedSemanticNode): boolean {
return node.ax !== undefined && node.ax.ignored !== true;
}
function isRenderedDomFallback(node: ResolvedSemanticNode): boolean {
if (!node.dom) return false;
const attrs = node.dom.attrs;
if (
Object.prototype.hasOwnProperty.call(attrs, "hidden") ||
Object.prototype.hasOwnProperty.call(attrs, "inert") ||
(attrs["aria-hidden"] ?? "").toLowerCase() === "true"
) {
return false;
}
return node.dom.rendered ?? (node.dom.localRect != null || node.dom.rect != null);
}
function initialParentId(
graph: ResolvedSemanticGraph,
node: ResolvedSemanticNode,
): SemanticNodeId | undefined {
let parentId = node.axParentId ?? node.domParentId;
if (node.axParentId && node.domParentId && node.axParentId !== node.domParentId) {
const axParent = graph.nodes.get(node.axParentId);
const axRole = axParent?.vom.role?.toLowerCase() ?? "";
const meaningfulAxParent =
axParent !== undefined &&
!ROOT_ROLES.has(axRole) &&
(explicitStructural(axParent) ||
!TRANSPARENT_ROLES.has(axRole) ||
Boolean(axParent.vom.name || axParent.vom.text || hasMeaningfulRelations(axParent)));
parentId = meaningfulAxParent ? node.axParentId : node.domParentId;
}
if (parentId) return parentId;
if (node.frameId === graph.rootFrameId) return undefined;
return graph.frames.get(node.frameId)?.ownerNodeId;
}
function orderedNodes(graph: ResolvedSemanticGraph): ResolvedSemanticNode[] {
return [...graph.frames.values()]
.sort((a, b) => a.order - b.order)
.flatMap((frame) =>
(graph.nodeIdsByFrame.get(frame.frameId) ?? [])
.map((id) => graph.nodes.get(id))
.filter((node): node is ResolvedSemanticNode => node !== undefined)
.sort((a, b) => a.sourceOrder - b.sourceOrder),
);
}
function recoveredNodes(
graph: ResolvedSemanticGraph,
nodes: ResolvedSemanticNode[],
): Map<SemanticNodeId, VomNode> {
const numericId = new Map(nodes.map((node, index) => [node.id, index + 1]));
const semanticId = new Map([...numericId].map(([id, numeric]) => [numeric, id]));
const provisional = nodes.map((node) => ({
...node.vom,
id: numericId.get(node.id) as number,
parentId: (() => {
const parent = initialParentId(graph, node);
return parent ? (numericId.get(parent) ?? null) : null;
})(),
...(node.backendNodeId !== undefined ? { backendNodeId: node.backendNodeId } : {}),
frameId: node.frameId,
contextScopeId: graph.frames.get(node.frameId)?.contextScopeId ?? node.frameId,
referenceable: node.referenceable,
}));
return new Map(
applyVomInteractionRecovery(provisional).flatMap((node) => {
const id = semanticId.get(node.id);
return id ? [[id, node] as const] : [];
}),
);
}
function redundantSourceNodes(
graph: ResolvedSemanticGraph,
nodes: ResolvedSemanticNode[],
recovered: ReadonlyMap<SemanticNodeId, VomNode>,
): Set<SemanticNodeId> {
const byKey = new Map<string, ResolvedSemanticNode[]>();
for (const node of nodes) {
const output = recovered.get(node.id) as VomNode;
const name = output.name?.trim().toLowerCase() ?? "";
if (!name || !isVomReferenceNode({ ...output, referenceable: true })) continue;
const key = `${node.frameId}\u0000${initialParentId(graph, node) ?? ""}\u0000${output.role?.toLowerCase() ?? ""}\u0000${name}\u0000${output.sensitive === true}`;
const matches = byKey.get(key) ?? [];
matches.push(node);
byKey.set(key, matches);
}
const redundant = new Set<SemanticNodeId>();
for (const matches of byKey.values()) {
const axOnly = matches.filter((node) => node.ax && !node.dom);
const domOnly = matches.filter((node) => node.dom && !node.ax);
if (matches.length === 2 && axOnly.length === 1 && domOnly.length === 1) {
redundant.add(domOnly[0].id);
}
}
return redundant;
}
function baseDecision(
graph: ResolvedSemanticGraph,
node: ResolvedSemanticNode,
output: VomNode,
frameOwnerIds: ReadonlySet<SemanticNodeId>,
redundant: ReadonlySet<SemanticNodeId>,
): StructureDecision {
if (node.excluded) return { kind: "excluded", reason: "excluded" };
if (redundant.has(node.id)) return { kind: "excluded", reason: "redundant-source" };
if (!hasUsableAxSemantics(node) && !isRenderedDomFallback(node)) {
return { kind: "excluded", reason: "non-rendered-dom-fallback" };
}
if (frameOwnerIds.has(node.id)) return { kind: "keep", reason: "frame-owner" };
if (node.frameId !== graph.rootFrameId && ROOT_ROLES.has(output.role?.toLowerCase() ?? "")) {
return { kind: "transparent", reason: "child-document-root" };
}
if (node.ax?.ignored === true && !node.dom) {
return { kind: "transparent", reason: "ax-ignored" };
}
const role = output.role?.toLowerCase() ?? "";
if (!TRANSPARENT_ROLES.has(role)) return { kind: "keep", reason: "semantic" };
if (
output.name ||
output.text ||
output.value ||
output.placeholder ||
hasMeaningfulRelations(node) ||
(output.rect && ["fixed", "sticky"].includes(output.position))
) {
return { kind: "keep", reason: "semantic" };
}
return { kind: "transparent", reason: "wrapper" };
}
function buildChildren(
nodes: ResolvedSemanticNode[],
relation: "axParentId" | "domParentId",
): Map<SemanticNodeId, SemanticNodeId[]> {
const children = new Map<SemanticNodeId, SemanticNodeId[]>();
for (const node of nodes) {
const parentId = node[relation];
if (!parentId) continue;
const siblings = children.get(parentId) ?? [];
siblings.push(node.id);
children.set(parentId, siblings);
}
return children;
}
function explicitDomAncestors(
graph: ResolvedSemanticGraph,
nodes: ResolvedSemanticNode[],
domChildren: ReadonlyMap<SemanticNodeId, SemanticNodeId[]>,
): Set<SemanticNodeId> {
const inside = new Set<SemanticNodeId>();
const roots = nodes.filter((node) => !node.domParentId);
const queue = roots.map((node) => node.id);
for (let index = 0; index < queue.length; index += 1) {
const parentId = queue[index];
const parent = graph.nodes.get(parentId);
const parentIsBoundary =
parent !== undefined &&
explicitStructural(parent) &&
!ROOT_ROLES.has(parent.vom.role?.toLowerCase() ?? "");
for (const childId of domChildren.get(parentId) ?? []) {
if (inside.has(parentId) || parentIsBoundary) inside.add(childId);
queue.push(childId);
}
}
return inside;
}
function reliableGroupCandidates(
graph: ResolvedSemanticGraph,
nodes: ResolvedSemanticNode[],
recovered: ReadonlyMap<SemanticNodeId, VomNode>,
decisions: ReadonlyMap<SemanticNodeId, StructureDecision>,
domChildren: ReadonlyMap<SemanticNodeId, SemanticNodeId[]>,
): Set<SemanticNodeId> {
const insideExplicitStructure = explicitDomAncestors(graph, nodes, domChildren);
const remainingChildren = new Map(
nodes.map((node) => [node.id, domChildren.get(node.id)?.length ?? 0]),
);
const summaries = new Map<SemanticNodeId, { blocked: boolean; interactions: number }>();
const queue = nodes
.filter((node) => (remainingChildren.get(node.id) ?? 0) === 0)
.map((node) => node.id);
for (let index = 0; index < queue.length; index += 1) {
const nodeId = queue[index];
const node = graph.nodes.get(nodeId);
const output = recovered.get(nodeId);
if (!node || !output) continue;
const role = output.role?.toLowerCase() ?? "";
const excluded = decisions.get(nodeId)?.kind === "excluded";
const blocked = excluded || explicitStructural(node) || CONTENT_BOUNDARY_ROLES.has(role);
const summary = excluded
? { blocked: true, interactions: 0 }
: isVomReferenceNode(output)
? { blocked: false, interactions: 1 }
: blocked
? { blocked: true, interactions: 0 }
: (domChildren.get(nodeId) ?? []).reduce(
(combined, childId) => {
const child = summaries.get(childId);
return {
blocked: combined.blocked || !child || child.blocked,
interactions: Math.min(2, combined.interactions + (child?.interactions ?? 0)),
};
},
{ blocked: false, interactions: 0 },
);
summaries.set(nodeId, summary);
const parentId = node.domParentId;
if (!parentId) continue;
const remaining = (remainingChildren.get(parentId) ?? 1) - 1;
remainingChildren.set(parentId, remaining);
if (remaining === 0) queue.push(parentId);
}
const candidates = new Set<SemanticNodeId>();
for (const node of nodes) {
const output = recovered.get(node.id) as VomNode;
const role = output.role?.toLowerCase() ?? "";
if (
!node.dom ||
node.dom.tag.toLowerCase() !== "div" ||
!TRANSPARENT_ROLES.has(role) ||
insideExplicitStructure.has(node.id) ||
decisions.get(node.id)?.kind === "excluded" ||
node.dom.attrs.hidden !== undefined ||
(node.dom.attrs["aria-hidden"] ?? "").toLowerCase() === "true" ||
node.dom.attrs.inert !== undefined
) {
continue;
}
const branches = (domChildren.get(node.id) ?? []).map((id) => summaries.get(id));
const blocked = branches.some((summary) => !summary || summary.blocked);
const contributingBranches = branches.filter(
(summary) => (summary?.interactions ?? 0) > 0,
).length;
const interactions = branches.reduce(
(count, summary) => Math.min(2, count + (summary?.interactions ?? 0)),
0,
);
if (!blocked && contributingBranches >= 2 && interactions >= 2) {
candidates.add(node.id);
}
}
return candidates;
}
function leafCandidates(
nodes: ResolvedSemanticNode[],
candidates: ReadonlySet<SemanticNodeId>,
domChildren: ReadonlyMap<SemanticNodeId, SemanticNodeId[]>,
): Set<SemanticNodeId> {
const parentById = new Map(
nodes.flatMap((node) => (node.domParentId ? ([[node.id, node.domParentId]] as const) : [])),
);
const childCount = new Map(nodes.map((node) => [node.id, domChildren.get(node.id)?.length ?? 0]));
const containsCandidate = new Set<SemanticNodeId>();
const queue = nodes.filter((node) => (childCount.get(node.id) ?? 0) === 0).map((node) => node.id);
const selected = new Set<SemanticNodeId>();
for (let index = 0; index < queue.length; index += 1) {
const nodeId = queue[index];
if (candidates.has(nodeId) && !containsCandidate.has(nodeId)) selected.add(nodeId);
const parentId = parentById.get(nodeId);
if (!parentId) continue;
if (candidates.has(nodeId) || containsCandidate.has(nodeId)) containsCandidate.add(parentId);
const remaining = (childCount.get(parentId) ?? 1) - 1;
childCount.set(parentId, remaining);
if (remaining === 0) queue.push(parentId);
}
return selected;
}
function nearestKept(
graph: ResolvedSemanticGraph,
startId: SemanticNodeId | undefined,
relation: "axParentId" | "domParentId",
kept: ReadonlySet<SemanticNodeId>,
memo: Map<SemanticNodeId, SemanticNodeId | undefined>,
): SemanticNodeId | undefined {
let currentId = startId;
const path: SemanticNodeId[] = [];
const seen = new Set<SemanticNodeId>();
let result: SemanticNodeId | undefined;
while (currentId && !seen.has(currentId)) {
if (kept.has(currentId)) {
result = currentId;
break;
}
if (memo.has(currentId)) {
result = memo.get(currentId);
break;
}
seen.add(currentId);
path.push(currentId);
currentId = graph.nodes.get(currentId)?.[relation];
}
for (const id of path) memo.set(id, result);
return result;
}
function finalParentId(
graph: ResolvedSemanticGraph,
node: ResolvedSemanticNode,
kept: ReadonlySet<SemanticNodeId>,
decisions: ReadonlyMap<SemanticNodeId, StructureDecision>,
axParentMemo: Map<SemanticNodeId, SemanticNodeId | undefined>,
domParentMemo: Map<SemanticNodeId, SemanticNodeId | undefined>,
): SemanticNodeId | undefined {
const axParent = nearestKept(graph, node.axParentId, "axParentId", kept, axParentMemo);
const domParent = nearestKept(graph, node.domParentId, "domParentId", kept, domParentMemo);
const axNode = axParent ? graph.nodes.get(axParent) : undefined;
const domNode = domParent ? graph.nodes.get(domParent) : undefined;
const axIsRoot = ROOT_ROLES.has(axNode?.vom.role?.toLowerCase() ?? "");
const axIsExplicit = axNode ? explicitStructural(axNode) && !axIsRoot : false;
const domIsStructural =
domNode !== undefined &&
(explicitStructural(domNode) || decisions.get(domNode.id)?.kind === "inferred-group");
if (axIsExplicit) return axParent;
if (domIsStructural) return domParent;
if (axParent && !axIsRoot) return axParent;
if (domParent) return domParent;
if (axParent) return axParent;
if (node.frameId === graph.rootFrameId) return undefined;
return graph.frames.get(node.frameId)?.ownerNodeId;
}
function recoveredVom(node: VomNode, inferredGroup: boolean): ResolvedSemanticNode["vom"] {
const {
id: _id,
parentId: _parentId,
backendNodeId: _backendNodeId,
frameId: _frameId,
contextScopeId: _contextScopeId,
referenceable: _referenceable,
...vom
} = node;
if (!inferredGroup) return vom;
const { name: _name, text: _text, nearbyText: _nearbyText, ...groupVom } = vom;
return { ...groupVom, role: "group" };
}
export function normalizeSemanticStructure(graph: ResolvedSemanticGraph): StructuredSemanticGraph {
const nodes = orderedNodes(graph);
const recovered = recoveredNodes(graph, nodes);
const redundant = redundantSourceNodes(graph, nodes, recovered);
const frameOwnerIds = new Set(
[...graph.frames.values()].flatMap((frame) => frame.ownerNodeId ?? []),
);
const decisions = new Map<SemanticNodeId, StructureDecision>();
for (const node of nodes) {
decisions.set(
node.id,
baseDecision(graph, node, recovered.get(node.id) as VomNode, frameOwnerIds, redundant),
);
}
const domChildren = buildChildren(nodes, "domParentId");
const candidates = reliableGroupCandidates(graph, nodes, recovered, decisions, domChildren);
for (const nodeId of leafCandidates(nodes, candidates, domChildren)) {
decisions.set(nodeId, {
kind: "inferred-group",
reason: "independent-interaction-branches",
});
}
const kept = new Set(
nodes
.filter((node) => {
const kind = decisions.get(node.id)?.kind;
return kind === "keep" || kind === "inferred-group";
})
.map((node) => node.id),
);
const axParentMemo = new Map<SemanticNodeId, SemanticNodeId | undefined>();
const domParentMemo = new Map<SemanticNodeId, SemanticNodeId | undefined>();
const structuredNodes = new Map<SemanticNodeId, StructuredSemanticNode>();
for (const node of nodes) {
const structure = decisions.get(node.id) as StructureDecision;
const inferredGroup = structure.kind === "inferred-group";
structuredNodes.set(node.id, {
...node,
vom: recoveredVom(recovered.get(node.id) as VomNode, inferredGroup),
structure,
...(kept.has(node.id)
? {
semanticParentId: finalParentId(
graph,
node,
kept,
decisions,
axParentMemo,
domParentMemo,
),
}
: {}),
});
}
return { ...graph, nodes: structuredNodes };
}
@@ -0,0 +1,88 @@
import type { Viewport, VomNode } from "@browser-skill/vom";
import type { CdpTarget } from "@/browser-driver/frame-graph";
import type { CapturedNode } from "../capture";
export interface SemanticAxNode {
nodeId: string;
frameId?: string;
parentId?: string;
backendDOMNodeId?: number;
ignored?: boolean;
role?: { type: string; value?: string };
name?: { type: string; value?: string };
description?: { value?: string };
value?: { value?: string | number | boolean };
properties?: Array<{ name?: string; value?: { value?: string | number | boolean } }>;
childIds?: string[];
}
export type SemanticNodeId = string;
export interface SemanticFrame {
frameId: string;
parentFrameId?: string;
ownerBackendNodeId?: number;
ownerNodeId?: SemanticNodeId;
contextScopeId: string;
target: CdpTarget;
url?: string;
order: number;
}
export interface SemanticGraphNode {
id: SemanticNodeId;
frameId: string;
backendNodeId?: number;
axNodeId?: string;
ax?: SemanticAxNode;
dom?: CapturedNode;
axParentId?: SemanticNodeId;
domParentId?: SemanticNodeId;
sourceOrder: number;
excluded: boolean;
}
export interface SemanticGraph {
viewport: Viewport;
rootFrameId: string;
frames: Map<string, SemanticFrame>;
nodes: Map<SemanticNodeId, SemanticGraphNode>;
nodeIdsByFrame: Map<string, SemanticNodeId[]>;
nodeByFrameBackend: Map<string, SemanticNodeId>;
}
export interface ResolvedSemanticNode extends SemanticGraphNode {
vom: Omit<VomNode, "id" | "parentId" | "backendNodeId" | "frameId" | "contextScopeId">;
referenceable: boolean;
roleSource: "ax" | "ax-ignored" | "dom-explicit" | "dom-native" | "none";
}
export interface ResolvedSemanticGraph extends Omit<SemanticGraph, "nodes"> {
nodes: Map<SemanticNodeId, ResolvedSemanticNode>;
}
export type StructureDecision =
| { kind: "keep"; reason: "semantic" | "frame-owner" }
| { kind: "inferred-group"; reason: "independent-interaction-branches" }
| { kind: "transparent"; reason: "wrapper" | "child-document-root" | "ax-ignored" }
| {
kind: "excluded";
reason: "excluded" | "redundant-source" | "non-rendered-dom-fallback";
};
export interface StructuredSemanticNode extends ResolvedSemanticNode {
structure: StructureDecision;
semanticParentId?: SemanticNodeId;
}
export interface StructuredSemanticGraph extends Omit<ResolvedSemanticGraph, "nodes"> {
nodes: Map<SemanticNodeId, StructuredSemanticNode>;
}
export function frameBackendKey(frameId: string, backendNodeId: number): string {
return `${frameId}\u0000${backendNodeId}`;
}
export function frameAxKey(frameId: string, axNodeId: string): string {
return `${frameId}\u0000${axNodeId}`;
}
@@ -65,6 +65,11 @@ function deferredFakeTransport(): { transport: Transport; emit: (frame: Protocol
}
describe("performHandshake", () => {
it("advertises the protocol compatibility boundary", () => {
expect(PROTOCOL_VERSION).toBe("1.1");
expect(MIN_COMPATIBLE_PROTOCOL).toBe("1.0");
});
it("sends system.handshake with identity and both compat fields", async () => {
let sentFrame: ProtocolFrame | null = null;
const transport = fakeTransport((req) => {
@@ -74,7 +79,7 @@ describe("performHandshake", () => {
result: {
server: "browser-skill-daemon",
version: "0.1.0",
protocol_version: "1.0",
protocol_version: "1.1",
min_compatible_peer: "0.0.0",
min_compatible_protocol: "1.0",
},
@@ -129,8 +134,8 @@ describe("performHandshake", () => {
const response = {
server: "browser-skill-daemon",
version: "0.1.0",
protocol_version: "1.0",
min_compatible_protocol: "1.0",
protocol_version: "1.1",
min_compatible_protocol: "1.1",
} satisfies HandshakeResult;
const transport = fakeTransport((req) => ({
id: (req as { id: string }).id,
@@ -145,7 +150,7 @@ describe("performHandshake", () => {
});
expect(outcome.result.min_compatible_peer).toBeUndefined();
expect(outcome.result.min_compatible_protocol).toBe("1.0");
expect(outcome.result.min_compatible_protocol).toBe("1.1");
});
it("rejects when the daemon responds with an error", async () => {
@@ -178,9 +183,9 @@ describe("performHandshake", () => {
result: {
server: "browser-skill-daemon",
version: "0.1.0",
protocol_version: "1.0",
protocol_version: "1.1",
min_compatible_peer: "0.0.0",
min_compatible_protocol: "1.0",
min_compatible_protocol: "1.1",
},
});
emit({
@@ -188,14 +193,14 @@ describe("performHandshake", () => {
result: {
server: "browser-skill-daemon",
version: "0.1.0",
protocol_version: "1.0",
protocol_version: "1.1",
min_compatible_peer: "0.0.0",
min_compatible_protocol: "1.0",
min_compatible_protocol: "1.1",
},
});
await expect(pending).resolves.toMatchObject({
result: { server: "browser-skill-daemon", protocol_version: "1.0" },
result: { server: "browser-skill-daemon", protocol_version: "1.1" },
});
});
+1 -1
View File
@@ -7,7 +7,7 @@ import type {
ResponseFrame,
} from "./types";
export const PROTOCOL_VERSION = "1.0";
export const PROTOCOL_VERSION = "1.1";
/**
* Extension semver, injected at build time from `package.json` via
* Vite's `define` (see `wxt.config.ts` and `vitest.config.ts`).
+7 -42
View File
@@ -744,48 +744,6 @@ export interface StepCommonV2 {
effect?: StepEffectV2;
}
/** Capture/buffer draft before v2 reduction. */
export type DraftTraceStep =
| {
op: "click";
target: TargetDescriptorV2;
navigated_to?: string;
page_url?: string;
}
| {
op: "hover";
target: TargetDescriptorV2;
page_url?: string;
}
| {
op: "fill";
target: TargetDescriptorV2;
value: string;
redacted?: boolean;
page_url?: string;
}
| {
op: "press";
key: string;
target?: TargetDescriptorV2;
modifiers?: KeyModifier[];
navigated_to?: string;
page_url?: string;
}
| {
op: "select";
target: TargetDescriptorV2;
values: string[];
labels?: string[];
navigated_to?: string;
page_url?: string;
}
| {
op: "navigate";
url: string;
page_url?: string;
};
/** Exported record-only step (trace v2). */
export type StepV2 =
| ({ op: "navigate" } & StepCommonV2 & { to: string })
@@ -822,6 +780,7 @@ export interface SelectedOptionV3 {
export type StepV3 =
| ({ op: "navigate" } & StepCommonV3 & { to: string; cause: NavigationCause })
| ({ op: "switch_tab" } & StepCommonV3)
| ({ op: "click" } & StepCommonV3 & { target: TargetDescriptorV3 })
| ({ op: "hover" } & StepCommonV3 & { target: TargetDescriptorV3 })
| ({ op: "fill" } & StepCommonV3 & {
@@ -861,6 +820,12 @@ export interface RecordStartParams {
tab_id?: number;
url?: string;
purpose?: string;
max_page_tokens?: number;
redact_values?: boolean;
/** Omitted means v2; `3` requests a state-linked v3 trace. */
trace_version?: number;
/** Client can decode the v3 `switch_tab` step variant. */
supports_tab_switch_steps?: boolean;
}
export interface RecordStartResult {
+1
View File
@@ -44,6 +44,7 @@
"!packages/dsh-plugin-browserskill/src/client/bsk-ui.nomodule.css",
"!packages/dsh-plugin-browserskill/.bsk-ui.raw.css",
"!packages/dsh-plugin-browserskill/.bsk-ui.input.css",
"!evals/browser/results/**",
"!**/target/**",
"!**/.wxt/**",
"!**/.output/**",
+14 -6
View File
@@ -265,19 +265,27 @@ retry — complete the task autonomously or stop gracefully.
### Recording — `bsk record`
Capture the user's own actions in the Agent Window to a `trace.json`, for later LLM-driven automation:
Capture the user's own actions in the Agent Window to a **trace bundle**, for later LLM-driven automation. New CLI builds request **trace v3** (page observations + action chain); older extensions may still return **trace v2** (actions only), which the CLI exports as a single `trace.json` without `states/`.
```bash
bsk record start --browser <instance-id-or-label> [--url https://…] [--purpose "publish a wiki doc"] [--output trace.json]
bsk record start --browser <instance-id-or-label> \
[--url https://…] [--purpose "publish a wiki doc"] \
[--max-page-tokens 3000] [--redact-values] \
[--output trace]
# `--url` is optional; default https://example.com/ when omitted (must be http(s)).
# Blocks until the user clicks Finish in the recording panel, then writes ./trace.json and closes the window.
# Blocks until the user clicks Finish in the recording panel, then writes:
# trace/trace.json — action chain (+ state index when v3)
# trace/states/ — v3 only: one `sN.txt` observe snapshot per settled page state
bsk record stop [--output trace.json] # terminal fallback if the browser panel is unavailable
bsk record stop [--output trace] # terminal fallback if the browser panel is unavailable
```
- The trace is a **record-only action log** (a `pages[]` dictionary + `navigate`/`click`/`fill`/`select`/`press` steps with `target` descriptors). It records *what the user did*; deciding which inputs are variable is left to the executing agent.
- **v3 bundle (preferred):** `--output` is a directory (default `./trace`) containing `trace.json` and `states/`. `trace.json` lists `states[]` (page observation ids) and `steps[]` (each step binds `state` = observe snapshot *before* the action and `result.state` = snapshot *after* settle). Page bodies live in `states/sN.txt` using the same VOM format as `bsk observe`.
- **v2 fallback:** when the connected extension is older, export may contain only legacy `trace.json` with `pages[]` action context and no `states/` observation files. Update the extension for full v3 bundles.
- Each `states[]` entry is one **settled page observation** (captured after recording start, navigation landing, or action settle — not on a timer).
- `target.ref` values like `@e12` exist **only inside** the bundle for disambiguation; do **not** copy `@eN` refs into SKILL.md or agent runbooks — use visible names from the observation text instead.
- `--purpose` is optional context metadata; it does **not** change what gets captured.
- There is **no** `bsk replay` — to redo a flow, read the trace and reuse the existing `session` / `snapshot` / `@eN` / `click` / `fill` tools. Follow **Stop when the goal is met**.
- `--redact-values` masks all form values in page files as `[filled]` / `[empty]`.
- Do **not** record on banking/SSO/password-manager pages; passwords are redacted but traces may still contain sensitive text.
## Error handling
+1
View File
@@ -22,6 +22,7 @@ pub mod navigate;
pub mod network;
pub mod observe;
pub mod record;
pub mod record_recovery;
pub mod record_state;
pub mod render_error;
pub mod screenshot;
+269 -71
View File
@@ -1,23 +1,30 @@
//! `bsk record start|stop` — capture user actions in the Agent Window.
use std::fs;
use std::path::PathBuf;
use std::path::{Path, PathBuf};
use std::time::Duration;
use anyhow::Context;
use bsk_protocol::Method;
use bsk_protocol::tools::{
RecordAwaitParams, RecordAwaitResult, RecordStartParams, RecordStartResult, RecordStopParams,
RecordStopResult, RecordedTrace,
RecordStopResult, RecordedTrace, TRACE_VERSION_V3,
};
use bsk_protocol::{ErrorCode, Method};
use clap::{Args, Subcommand};
use crate::cli::TOOL_IPC_TIMEOUT;
mod export;
use crate::cli::business_rpc;
use crate::cli::ensure_daemon::ensure_daemon;
use crate::cli::error::{CliError, Format};
use crate::cli::record_recovery;
use crate::cli::record_state;
use crate::cli::session::{SessionStartOptions, start_session, stop_session};
use export::{
ExportMeta, export_with_recovery, states_dir_for_output, trace_json_path,
validate_record_output,
};
/// Max wait for the user to click 结束 in the browser (24 hours).
const RECORD_AWAIT_TIMEOUT_MS: u32 = 86_400_000;
@@ -50,7 +57,8 @@ pub struct RecordStartArgs {
pub tab_id: Option<i64>,
/// Navigate to this http(s) URL before recording. When omitted, defaults
/// to `https://example.com/`.
/// to `https://example.com/`. If that page does not load, pass `--url`
/// with a site you can open in this browser.
#[arg(long)]
pub url: Option<String>,
@@ -58,15 +66,25 @@ pub struct RecordStartArgs {
#[arg(long)]
pub purpose: Option<String>,
/// Output path for the trace JSON (default `./trace.json`).
#[arg(long, default_value = "trace.json")]
/// Max VOM tokens per page observation file (default 3000).
#[arg(long = "max-page-tokens")]
pub max_page_tokens: Option<u32>,
/// Redact all form values in page observations (`[filled]` only).
#[arg(long = "redact-values")]
pub redact_values: bool,
/// Output directory for the trace bundle (default `./trace`).
/// Writes `<dir>/trace.json` and `<dir>/states/*.txt`.
#[arg(long, default_value = "trace")]
pub output: PathBuf,
}
#[derive(Debug, Clone, Args)]
pub struct RecordStopArgs {
/// Output path for the trace JSON (default `./trace.json`).
#[arg(long, default_value = "trace.json")]
/// Output directory for the trace bundle (default `./trace`).
/// Writes `<dir>/trace.json` and `<dir>/states/*.txt`.
#[arg(long, default_value = "trace")]
pub output: PathBuf,
}
@@ -78,11 +96,7 @@ pub fn dispatch(cmd: RecordCmd, format: Format) -> Result<(), CliError> {
}
fn dispatch_start(args: RecordStartArgs, format: Format) -> Result<(), CliError> {
if record_state::read().is_ok() {
return Err(CliError::Local(anyhow::anyhow!(
"a recording is already in progress; run `bsk record stop` first"
)));
}
prepare_record_start(&args.output)?;
let info = ensure_daemon().context("ensure daemon is running")?;
let session = start_session(
@@ -96,8 +110,12 @@ fn dispatch_start(args: RecordStartArgs, format: Format) -> Result<(), CliError>
let start_params = RecordStartParams {
session_id: session.session_id.clone(),
tab_id: args.tab_id,
url: args.url,
url: args.url.clone(),
purpose: args.purpose.clone(),
max_page_tokens: args.max_page_tokens,
redact_values: Some(args.redact_values),
trace_version: Some(TRACE_VERSION_V3),
supports_tab_switch_steps: Some(true),
};
let start_result = business_rpc::call::<RecordStartParams, RecordStartResult>(
info.sock_path.clone(),
@@ -111,7 +129,7 @@ fn dispatch_start(args: RecordStartArgs, format: Format) -> Result<(), CliError>
Ok(result) => result,
Err(err) => {
let _ = stop_session(info.sock_path, &session.session_id);
return Err(err);
return Err(annotate_default_start_page_error(err, args.url.as_deref()));
}
};
@@ -142,8 +160,8 @@ fn dispatch_start(args: RecordStartArgs, format: Format) -> Result<(), CliError>
// Keep write/render inside a Result so `?` cannot skip session teardown.
let run_result: Result<(), CliError> = match await_result {
Ok(result) => (|| {
write_trace_file(&args.output, &result.trace)?;
render_finish(&result.trace, &args.output, format)
let exported = export_with_recovery(&args.output, &result.trace)?;
render_finish(&result.trace, &args.output, &exported, format)
})(),
Err(err) => Err(err),
};
@@ -157,11 +175,12 @@ fn dispatch_start(args: RecordStartArgs, format: Format) -> Result<(), CliError>
}
fn dispatch_stop(args: RecordStopArgs, format: Format) -> Result<(), CliError> {
let state = record_state::read().map_err(CliError::Local)?;
let info = ensure_daemon().context("ensure daemon is running")?;
let session_id = state.session_id.clone();
validate_record_output(&args.output)?;
if let Ok(state) = record_state::read() {
let info = ensure_daemon().context("ensure daemon is running")?;
let session_id = state.session_id.clone();
let run_result: Result<(), CliError> = (|| {
let params = RecordStopParams {
session_id: session_id.clone(),
};
@@ -172,16 +191,85 @@ fn dispatch_stop(args: RecordStopArgs, format: Format) -> Result<(), CliError> {
Some(params),
TOOL_IPC_TIMEOUT,
)?;
write_trace_file(&args.output, &result.trace)?;
render_stop(&result, &args.output, format)
})();
let session_stop_result = stop_session(info.sock_path, &session_id);
record_state::clear();
let run_result: Result<(), CliError> = (|| {
let exported = export_with_recovery(&args.output, &result.trace)?;
render_stop(&result, &args.output, &exported, format)
})();
run_result?;
session_stop_result?;
Ok(())
let session_stop_result = stop_session(info.sock_path, &session_id);
record_state::clear();
run_result?;
session_stop_result?;
return Ok(());
}
let Some(trace) = record_recovery::load().map_err(CliError::Local)? else {
return Err(CliError::Local(anyhow::anyhow!(
"no recording in progress; run `bsk record start` first"
)));
};
let exported = export_with_recovery(&args.output, &trace)?;
render_finish(&trace, &args.output, &exported, format)
}
/// When `record start` omitted `--url` and the default example.com page
/// never loaded, rewrite the RPC error so the CLI hint points at `--url`.
fn annotate_default_start_page_error(err: CliError, user_url: Option<&str>) -> CliError {
if user_url.is_some() {
return err;
}
let CliError::Rpc {
code,
message,
data,
source,
} = err
else {
return err;
};
if !is_default_start_page_failure(code, &message) {
return CliError::Rpc {
code,
message,
data,
source,
};
}
CliError::Rpc {
code,
message,
data: Some(serde_json::json!({
"reason": crate::cli::render_error::reason::RECORD_START_PAGE_UNREACHABLE
})),
source,
}
}
fn is_default_start_page_failure(code: ErrorCode, message: &str) -> bool {
match code {
ErrorCode::CdpFailed | ErrorCode::Timeout => true,
ErrorCode::InvalidParams => {
message.contains("restricted URL") || message.contains("about:blank")
}
_ => false,
}
}
fn prepare_record_start(output: &Path) -> Result<(), CliError> {
if record_state::read().is_ok() {
return Err(CliError::Local(anyhow::anyhow!(
"a recording is already in progress; run `bsk record stop` first"
)));
}
if record_recovery::exists() {
return Err(CliError::Local(anyhow::anyhow!(
"a previous recording was not exported; run `bsk record stop --output <dir>` to recover it first"
)));
}
validate_record_output(output)
}
fn record_await_ipc_timeout(timeout_ms: u32) -> Duration {
@@ -190,43 +278,66 @@ fn record_await_ipc_timeout(timeout_ms: u32) -> Duration {
.unwrap_or(Duration::from_secs(u64::from(timeout_ms / 1_000) + 15))
}
fn write_trace_file(output: &PathBuf, trace: &RecordedTrace) -> Result<(), CliError> {
let json = serde_json::to_string_pretty(trace)
.context("serialize trace JSON")
.map_err(CliError::Local)?;
if let Some(parent) = output.parent() {
if !parent.as_os_str().is_empty() {
fs::create_dir_all(parent)
.with_context(|| format!("create output directory {}", parent.display()))
.map_err(CliError::Local)?;
}
}
fs::write(output, format!("{json}\n"))
.with_context(|| format!("write trace to {}", output.display()))
.map_err(CliError::Local)?;
Ok(())
}
fn render_finish(trace: &RecordedTrace, output: &PathBuf, format: Format) -> Result<(), CliError> {
fn render_finish(
trace: &RecordedTrace,
output_dir: &PathBuf,
exported: &ExportMeta,
format: Format,
) -> Result<(), CliError> {
let trace_path = trace_json_path(output_dir);
match format {
Format::Json => {
let payload = match trace {
RecordedTrace::V3(t) => serde_json::json!({
"output": output_dir,
"trace_json": trace_path,
"trace_version": exported.trace_version,
"states_dir": exported.states_dir,
"trace": t,
"window_closed": true,
}),
RecordedTrace::V2(t) => serde_json::json!({
"output": output_dir,
"trace_json": trace_path,
"trace_version": exported.trace_version,
"states_dir": null,
"trace": t,
"window_closed": true,
}),
};
println!(
"{}",
serde_json::to_string(&serde_json::json!({
"output": output,
"trace": trace,
"window_closed": true,
}))
.map_err(|e| CliError::Local(anyhow::anyhow!(e)))?
serde_json::to_string(&payload).map_err(|e| CliError::Local(anyhow::anyhow!(e)))?
);
}
Format::Human => {
let step_count = match trace {
RecordedTrace::V2(trace) => trace.steps.len(),
RecordedTrace::V3(trace) => trace.steps.len(),
};
println!("saved {step_count} steps to {}", output.display());
}
Format::Human => match trace {
RecordedTrace::V3(t) => {
let states_dir = exported
.states_dir
.as_ref()
.map(|p| p.display().to_string())
.unwrap_or_else(|| states_dir_for_output(output_dir).display().to_string());
println!(
"saved {} steps to {} and {} states to {}",
t.steps.len(),
trace_path.display(),
t.states.len(),
states_dir
);
}
RecordedTrace::V2(t) => {
if exported.v2_fallback {
eprintln!(
"note: extension returned trace v2 (no page observations); update the BrowserSkill extension for v3 bundles"
);
}
println!(
"saved {} steps to {} (trace v2)",
t.steps.len(),
trace_path.display()
);
}
},
}
Ok(())
}
@@ -234,33 +345,89 @@ fn render_finish(trace: &RecordedTrace, output: &PathBuf, format: Format) -> Res
fn render_stop(
result: &RecordStopResult,
output: &PathBuf,
exported: &ExportMeta,
format: Format,
) -> Result<(), CliError> {
render_finish(&result.trace, output, format)
render_finish(&result.trace, output, exported, format)
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn default_output_is_trace_json() {
let args = RecordStopArgs {
output: PathBuf::from("trace.json"),
};
assert_eq!(args.output, PathBuf::from("trace.json"));
fn with_temp_home<F: FnOnce()>(f: F) {
let _lock = record_recovery::test_env_lock();
let tmp = tempfile::tempdir().unwrap();
unsafe {
std::env::set_var(crate::daemon::paths::BSK_HOME_ENV, tmp.path());
}
f();
record_recovery::clear();
record_state::clear();
unsafe {
std::env::remove_var(crate::daemon::paths::BSK_HOME_ENV);
}
}
#[test]
fn start_args_default_output_is_trace_json() {
fn prepare_record_start_rejects_existing_trace_json_file() {
with_temp_home(|| {
let dir = tempfile::tempdir().unwrap();
let output = dir.path().join("trace.json");
std::fs::write(&output, "{}\n").unwrap();
let err = prepare_record_start(&output).unwrap_err();
let msg = err.to_string();
assert!(
msg.contains("JSON file") || msg.contains("json file"),
"{msg}"
);
assert!(
msg.contains("states/") || msg.contains("--output trace"),
"{msg}"
);
});
}
#[test]
fn prepare_record_start_rejects_unexported_recovery() {
with_temp_home(|| {
crate::daemon::paths::ensure_bsk_home().unwrap();
std::fs::write(
crate::daemon::paths::record_recovery_path().unwrap(),
"{}\n",
)
.unwrap();
let err = prepare_record_start(Path::new("trace")).unwrap_err();
let msg = err.to_string();
assert!(
msg.contains("not exported") || msg.contains("recover"),
"{msg}"
);
});
}
#[test]
fn default_output_is_trace_dir() {
let args = RecordStopArgs {
output: PathBuf::from("trace"),
};
assert_eq!(args.output, PathBuf::from("trace"));
}
#[test]
fn start_args_default_output_is_trace_dir() {
let args = RecordStartArgs {
browser: None,
tab_id: None,
url: None,
purpose: None,
output: PathBuf::from("trace.json"),
max_page_tokens: None,
redact_values: false,
output: PathBuf::from("trace"),
};
assert_eq!(args.output, PathBuf::from("trace.json"));
assert_eq!(args.output, PathBuf::from("trace"));
}
#[test]
@@ -268,4 +435,35 @@ mod tests {
let got = record_await_ipc_timeout(RECORD_AWAIT_TIMEOUT_MS);
assert!(got >= Duration::from_secs(86_400));
}
#[test]
fn annotate_default_start_page_error_adds_url_reason() {
let err = CliError::from_rpc(bsk_protocol::RpcError {
code: ErrorCode::CdpFailed,
message: "Page.navigate rejected: net::ERR_ABORTED".into(),
data: None,
});
let annotated = annotate_default_start_page_error(err, None);
assert_eq!(
crate::cli::render_error::reason_for_data(annotated.data()),
Some(crate::cli::render_error::reason::RECORD_START_PAGE_UNREACHABLE)
);
let info =
crate::cli::render_error::info_for_error(annotated.code().unwrap(), annotated.data());
assert!(info.hint.unwrap().contains("--url"));
}
#[test]
fn annotate_default_start_page_error_skips_when_user_passed_url() {
let err = CliError::from_rpc(bsk_protocol::RpcError {
code: ErrorCode::CdpFailed,
message: "Page.navigate rejected: net::ERR_ABORTED".into(),
data: None,
});
let annotated = annotate_default_start_page_error(err, Some("https://www.example.org/"));
assert_eq!(
crate::cli::render_error::reason_for_data(annotated.data()),
None
);
}
}
+976
View File
@@ -0,0 +1,976 @@
use std::collections::HashSet;
use std::fs::{self, OpenOptions};
use std::io;
use std::path::{Path, PathBuf};
use anyhow::Context;
use bsk_protocol::tools::{RecordedTrace, TRACE_VERSION_V2, TRACE_VERSION_V3, TraceV2, TraceV3};
use serde::Serialize;
use crate::cli::error::CliError;
#[derive(Debug)]
pub(super) struct ExportMeta {
pub(super) states_dir: Option<PathBuf>,
pub(super) trace_version: u32,
/// True when the CLI requested v3 but the extension returned legacy v2.
pub(super) v2_fallback: bool,
}
pub(super) fn export_recorded_trace(
output_dir: &Path,
trace: &RecordedTrace,
) -> Result<ExportMeta, CliError> {
match trace {
RecordedTrace::V3(trace) => {
let states_dir = write_trace_bundle(output_dir, trace)?;
Ok(ExportMeta {
states_dir: Some(states_dir),
trace_version: TRACE_VERSION_V3,
v2_fallback: false,
})
}
RecordedTrace::V2(trace) => {
write_trace_v2(output_dir, trace)?;
Ok(ExportMeta {
states_dir: None,
trace_version: TRACE_VERSION_V2,
v2_fallback: true,
})
}
}
}
/// Save the completed Trace first, then write the bundle. Recovery is
/// removed only after the export commits so a bad `--output` cannot drop
/// a recording the extension already returned.
pub(super) fn export_with_recovery(
output_dir: &Path,
trace: &RecordedTrace,
) -> Result<ExportMeta, CliError> {
let save_result = crate::cli::record_recovery::save(trace);
match export_recorded_trace(output_dir, trace) {
Ok(meta) => {
crate::cli::record_recovery::clear();
Ok(meta)
}
Err(export_err) => {
let export_err = annotate_recovery_hint(export_err, save_result.is_ok());
if let Err(save_err) = save_result {
return Err(match export_err {
CliError::Local(inner) => CliError::Local(
inner.context(format!("also failed to save recovery data: {save_err}")),
),
other => other,
});
}
Err(export_err)
}
}
}
fn annotate_recovery_hint(err: CliError, recovery_saved: bool) -> CliError {
if !recovery_saved {
return err;
}
match err {
CliError::Local(inner) => CliError::Local(inner.context(
"export failed; the recorded trace was saved and can be recovered with `bsk record stop --output <dir>`",
)),
other => other,
}
}
fn looks_like_json_output(path: &Path) -> bool {
path.extension()
.and_then(|ext| ext.to_str())
.is_some_and(|ext| ext.eq_ignore_ascii_case("json"))
}
fn legacy_json_output_error(path: &Path) -> CliError {
CliError::Local(anyhow::anyhow!(
"--output {} is a JSON file path; Trace v3 writes a bundle directory \
(`<dir>/trace.json` and `<dir>/states/`), not a single file. \
Use `--output trace` or another directory.",
path.display()
))
}
/// Reject legacy `--output trace.json` (and any existing non-directory)
/// before a recording starts, so the user is not blocked only after Finish.
pub(super) fn validate_record_output(path: &Path) -> Result<(), CliError> {
match fs::symlink_metadata(path) {
Ok(metadata) if metadata.file_type().is_symlink() => Err(CliError::Local(anyhow::anyhow!(
"--output {} must not be a symlink",
path.display()
))),
Ok(metadata) if metadata.is_dir() => Ok(()),
Ok(_) if looks_like_json_output(path) => Err(legacy_json_output_error(path)),
Ok(_) => Err(CliError::Local(anyhow::anyhow!(
"--output {} is not a directory; Trace v3 writes a bundle directory \
(`<dir>/trace.json` and `<dir>/states/`). Use `--output trace`.",
path.display()
))),
Err(err) if err.kind() == io::ErrorKind::NotFound => {
if looks_like_json_output(path) {
Err(legacy_json_output_error(path))
} else {
Ok(())
}
}
Err(err) => Err(CliError::Local(
anyhow::Error::new(err).context(format!("inspect --output {}", path.display())),
)),
}
}
fn acquire_export_lock(output_dir: &Path) -> Result<std::fs::File, CliError> {
validate_bundle_directory(output_dir, "output bundle directory")?;
fs::create_dir_all(output_dir)
.with_context(|| format!("create output dir {}", output_dir.display()))
.map_err(CliError::Local)?;
let lock_path = output_dir.join(".bsk-record-export.lock");
let lock_file = OpenOptions::new()
.read(true)
.write(true)
.create(true)
.truncate(false)
.open(&lock_path)
.with_context(|| format!("open export lock {}", lock_path.display()))
.map_err(CliError::Local)?;
fs2::FileExt::lock_exclusive(&lock_file)
.with_context(|| format!("lock export bundle {}", output_dir.display()))
.map_err(CliError::Local)?;
Ok(lock_file)
}
pub(super) fn write_trace_v2(output_dir: &Path, trace: &TraceV2) -> Result<(), CliError> {
let trace_path = trace_json_path(output_dir);
let states_dir = states_dir_for_output(output_dir);
let json = serde_json::to_string_pretty(trace)
.context("serialize trace JSON")
.map_err(CliError::Local)?;
let _lock = acquire_export_lock(output_dir)?;
validate_replaceable_file(&trace_path, "trace JSON")?;
let transaction_id = uuid::Uuid::new_v4();
let staging_dir = output_dir.join(format!(".bsk-record-stage-{transaction_id}"));
fs::create_dir_all(&staging_dir)
.with_context(|| format!("create staging dir {}", staging_dir.display()))
.map_err(CliError::Local)?;
let staged_trace = staging_dir.join("trace.json");
let stage_result: Result<(), CliError> = (|| {
fs::write(&staged_trace, format!("{json}\n"))
.with_context(|| format!("write staged trace to {}", staged_trace.display()))
.map_err(CliError::Local)?;
Ok(())
})();
if let Err(err) = stage_result {
let _ = fs::remove_dir_all(&staging_dir);
return Err(err);
}
commit_export_transaction(
&staging_dir,
&states_dir,
&HashSet::new(),
transaction_id,
vec![(
staged_trace,
trace_path,
output_dir.join(format!(".trace.json.{transaction_id}.bak")),
)],
)?;
Ok(())
}
pub(super) fn states_dir_for_output(output_dir: &Path) -> PathBuf {
output_dir.join("states")
}
pub(super) fn trace_json_path(output_dir: &Path) -> PathBuf {
output_dir.join("trace.json")
}
fn is_canonical_state_id(id: &str) -> bool {
let Some(number) = id.strip_prefix('s') else {
return false;
};
let mut digits = number.bytes();
matches!(digits.next(), Some(b'1'..=b'9')) && digits.all(|byte| byte.is_ascii_digit())
}
fn validate_bundle_directory(path: &Path, description: &str) -> Result<(), CliError> {
match fs::symlink_metadata(path) {
Ok(metadata) if metadata.file_type().is_symlink() => Err(CliError::Local(anyhow::anyhow!(
"{description} {} must not be a symlink",
path.display()
))),
Ok(metadata) if metadata.is_dir() => Ok(()),
Ok(_) if looks_like_json_output(path) => Err(legacy_json_output_error(path)),
Ok(_) => Err(CliError::Local(anyhow::anyhow!(
"{description} {} is not a directory; Trace v3 writes `<dir>/trace.json` and `<dir>/states/`. Use `--output trace`.",
path.display()
))),
Err(err) if err.kind() == io::ErrorKind::NotFound => Ok(()),
Err(err) => Err(CliError::Local(
anyhow::Error::new(err).context(format!("inspect {description} {}", path.display())),
)),
}
}
fn validate_replaceable_file(path: &Path, description: &str) -> Result<(), CliError> {
match fs::symlink_metadata(path) {
Ok(metadata) if metadata.file_type().is_file() => Ok(()),
Ok(_) => Err(CliError::Local(anyhow::anyhow!(
"{description} {} is not a regular file",
path.display()
))),
Err(err) if err.kind() == io::ErrorKind::NotFound => Ok(()),
Err(err) => Err(CliError::Local(
anyhow::Error::new(err).context(format!("inspect {description} {}", path.display())),
)),
}
}
fn commit_staged_file(staged: &Path, target: &Path, backup: &Path) -> io::Result<Option<PathBuf>> {
let old_file = match fs::symlink_metadata(target) {
Ok(_) => {
fs::rename(target, backup)?;
Some(backup.to_path_buf())
}
Err(err) if err.kind() == io::ErrorKind::NotFound => None,
Err(err) => return Err(err),
};
if let Err(err) = fs::rename(staged, target) {
if let Some(old_file) = old_file.as_deref() {
if let Err(restore_err) = fs::rename(old_file, target) {
return Err(io::Error::new(
restore_err.kind(),
format!(
"{err}; additionally failed to restore {}: {restore_err}",
target.display()
),
));
}
}
return Err(err);
}
Ok(old_file)
}
fn rollback_installed_files(installed: &[(PathBuf, Option<PathBuf>)]) -> io::Result<()> {
let mut rollback_error = None;
for (target, backup) in installed.iter().rev() {
match fs::remove_file(target) {
Ok(()) => {}
Err(err) if err.kind() == io::ErrorKind::NotFound => {}
Err(err) => {
rollback_error.get_or_insert(err);
continue;
}
}
if let Some(backup) = backup {
if let Err(err) = fs::rename(backup, target) {
rollback_error.get_or_insert(err);
}
}
}
match rollback_error {
Some(err) => Err(err),
None => Ok(()),
}
}
fn is_generated_state_name(name: &str) -> bool {
name.strip_suffix(".txt").is_some_and(is_canonical_state_id)
}
fn restore_staged_stale_states(staged: &[(PathBuf, PathBuf)]) -> io::Result<()> {
let mut restore_error = None;
for (original, backup) in staged.iter().rev() {
if let Err(err) = fs::rename(backup, original) {
restore_error.get_or_insert(err);
}
}
match restore_error {
Some(err) => Err(err),
None => Ok(()),
}
}
/// Move stale generated states aside so replacement and cleanup form one
/// rollback boundary. User-owned files, including arbitrary non-`sN.txt`, stay.
fn stage_stale_states(
states_dir: &Path,
keep: &HashSet<String>,
transaction_id: uuid::Uuid,
) -> Result<Vec<(PathBuf, PathBuf)>, CliError> {
let entries = match fs::read_dir(states_dir) {
Ok(entries) => entries,
Err(err) if err.kind() == io::ErrorKind::NotFound => return Ok(Vec::new()),
Err(err) => {
return Err(CliError::Local(
anyhow::Error::new(err)
.context(format!("read states dir {}", states_dir.display())),
));
}
};
let mut staged = Vec::new();
for entry in entries {
let entry = match entry {
Ok(entry) => entry,
Err(err) => {
let restore_result = restore_staged_stale_states(&staged);
let message = match restore_result {
Ok(()) => format!("read stale state entry: {err}"),
Err(restore_err) => {
format!("read stale state entry: {err}; restore also failed: {restore_err}")
}
};
return Err(CliError::Local(anyhow::anyhow!(message)));
}
};
let name = entry.file_name();
let Some(name) = name.to_str() else { continue };
if !is_generated_state_name(name) || keep.contains(name) {
continue;
}
let original = entry.path();
let file_type = match entry.file_type() {
Ok(file_type) => file_type,
Err(err) => {
let restore_result = restore_staged_stale_states(&staged);
let message = match restore_result {
Ok(()) => format!("inspect stale state {}: {err}", original.display()),
Err(restore_err) => format!(
"inspect stale state {}: {err}; restore also failed: {restore_err}",
original.display()
),
};
return Err(CliError::Local(anyhow::anyhow!(message)));
}
};
if file_type.is_dir() {
let restore_result = restore_staged_stale_states(&staged);
let message = match restore_result {
Ok(()) => format!("stale state {} is a directory", original.display()),
Err(restore_err) => format!(
"stale state {} is a directory; restore also failed: {restore_err}",
original.display()
),
};
return Err(CliError::Local(anyhow::anyhow!(message)));
}
let backup = states_dir.join(format!(".{name}.{transaction_id}.stale"));
if let Err(err) = fs::rename(&original, &backup) {
let restore_result = restore_staged_stale_states(&staged);
let message = match restore_result {
Ok(()) => format!("stage stale state {}: {err}", original.display()),
Err(restore_err) => format!(
"stage stale state {}: {err}; restore also failed: {restore_err}",
original.display()
),
};
return Err(CliError::Local(anyhow::anyhow!(message)));
}
staged.push((original, backup));
}
Ok(staged)
}
fn commit_export_transaction(
staging_dir: &Path,
states_dir: &Path,
keep_states: &HashSet<String>,
transaction_id: uuid::Uuid,
replacements: Vec<(PathBuf, PathBuf, PathBuf)>,
) -> Result<(), CliError> {
let mut installed = Vec::with_capacity(replacements.len());
for (staged, target, backup) in replacements {
match commit_staged_file(&staged, &target, &backup) {
Ok(old_file) => installed.push((target, old_file)),
Err(err) => {
let rollback_result = rollback_installed_files(&installed);
let _ = fs::remove_dir_all(staging_dir);
let message = match rollback_result {
Ok(()) => format!("commit replacement {}: {err}", target.display()),
Err(rollback_err) => format!(
"commit replacement {}: {err}; rollback also failed: {rollback_err}",
target.display()
),
};
return Err(CliError::Local(anyhow::anyhow!(message)));
}
}
}
let stale_states = match stage_stale_states(states_dir, keep_states, transaction_id) {
Ok(stale) => stale,
Err(err) => {
let rollback_result = rollback_installed_files(&installed);
let _ = fs::remove_dir_all(staging_dir);
let message = match rollback_result {
Ok(()) => err.to_string(),
Err(rollback_err) => {
format!("{err}; replacement rollback also failed: {rollback_err}")
}
};
return Err(CliError::Local(anyhow::anyhow!(message)));
}
};
for (_, backup) in &installed {
if let Some(backup) = backup {
let _ = fs::remove_file(backup);
}
}
for (_, backup) in stale_states {
let _ = fs::remove_file(backup);
}
let _ = fs::remove_dir_all(staging_dir);
Ok(())
}
#[derive(Serialize)]
struct BundleState<'a> {
id: &'a str,
url: &'a str,
#[serde(skip_serializing_if = "Option::is_none")]
title: Option<&'a str>,
page: String,
#[serde(skip_serializing_if = "std::ops::Not::not")]
truncated: bool,
}
#[derive(Serialize)]
struct BundleTrace<'a> {
version: u32,
#[serde(skip_serializing_if = "Option::is_none")]
purpose: Option<&'a str>,
#[serde(skip_serializing_if = "Option::is_none")]
started_at: Option<&'a str>,
recorded_at: &'a str,
stopped_by: &'a bsk_protocol::tools::StopReason,
entry: &'a bsk_protocol::tools::TraceEntry,
recorder: &'a bsk_protocol::tools::RecorderInfo,
states: Vec<BundleState<'a>>,
steps: &'a [bsk_protocol::tools::StepV3],
}
pub(super) fn write_trace_bundle(output_dir: &Path, trace: &TraceV3) -> Result<PathBuf, CliError> {
let states_dir = states_dir_for_output(output_dir);
let trace_path = trace_json_path(output_dir);
if trace.version != TRACE_VERSION_V3 {
return Err(CliError::Local(anyhow::anyhow!(
"trace version {} is not supported (expected {})",
trace.version,
TRACE_VERSION_V3
)));
}
let mut state_ids = HashSet::with_capacity(trace.states.len());
for state in &trace.states {
if !is_canonical_state_id(&state.id) {
return Err(CliError::Local(anyhow::anyhow!(
"state id {:?} is not canonical (expected s1, s2, ...)",
state.id
)));
}
if !state_ids.insert(state.id.as_str()) {
return Err(CliError::Local(anyhow::anyhow!(
"duplicate state id {:?}",
state.id
)));
}
}
let mut disk_states = Vec::with_capacity(trace.states.len());
let mut state_writes = Vec::new();
let mut keep = HashSet::new();
for state in &trace.states {
let filename = format!("{}.txt", state.id);
let state_path = states_dir.join(&filename);
if state_path.parent() != Some(states_dir.as_path()) {
return Err(CliError::Local(anyhow::anyhow!(
"state path {} escapes states directory {}",
state_path.display(),
states_dir.display()
)));
}
keep.insert(filename.clone());
state_writes.push((filename.clone(), state.body.as_str()));
disk_states.push(BundleState {
id: &state.id,
url: &state.url,
title: state.title.as_deref(),
page: filename,
truncated: state.truncated,
});
}
let disk_trace = BundleTrace {
version: trace.version,
purpose: trace.purpose.as_deref(),
started_at: trace.started_at.as_deref(),
recorded_at: &trace.recorded_at,
stopped_by: &trace.stopped_by,
entry: &trace.entry,
recorder: &trace.recorder,
states: disk_states,
steps: &trace.steps,
};
let json = serde_json::to_string_pretty(&disk_trace)
.context("serialize trace JSON")
.map_err(CliError::Local)?;
validate_bundle_directory(output_dir, "output bundle directory")?;
fs::create_dir_all(output_dir)
.with_context(|| format!("create output dir {}", output_dir.display()))
.map_err(CliError::Local)?;
validate_bundle_directory(&states_dir, "states directory")?;
let _lock = acquire_export_lock(output_dir)?;
fs::create_dir_all(&states_dir)
.with_context(|| format!("create states dir {}", states_dir.display()))
.map_err(CliError::Local)?;
for (filename, _) in &state_writes {
validate_replaceable_file(&states_dir.join(filename), "state observation")?;
}
validate_replaceable_file(&trace_path, "trace JSON")?;
let transaction_id = uuid::Uuid::new_v4();
let staging_dir = output_dir.join(format!(".bsk-record-stage-{transaction_id}"));
let staging_states = staging_dir.join("states");
fs::create_dir_all(&staging_states)
.with_context(|| format!("create staging dir {}", staging_states.display()))
.map_err(CliError::Local)?;
let stage_result: Result<(), CliError> = (|| {
for (filename, body) in &state_writes {
let staged_path = staging_states.join(filename);
fs::write(&staged_path, body)
.with_context(|| {
format!("write staged state observation {}", staged_path.display())
})
.map_err(CliError::Local)?;
}
let staged_trace = staging_dir.join("trace.json");
fs::write(&staged_trace, format!("{json}\n"))
.with_context(|| format!("write staged trace to {}", staged_trace.display()))
.map_err(CliError::Local)?;
Ok(())
})();
if let Err(err) = stage_result {
let _ = fs::remove_dir_all(&staging_dir);
return Err(err);
}
let mut replacements: Vec<(PathBuf, PathBuf, PathBuf)> =
Vec::with_capacity(state_writes.len() + 1);
for (filename, _) in &state_writes {
replacements.push((
staging_states.join(filename),
states_dir.join(filename),
states_dir.join(format!(".{filename}.{transaction_id}.bak")),
));
}
replacements.push((
staging_dir.join("trace.json"),
trace_path.clone(),
output_dir.join(format!(".trace.json.{transaction_id}.bak")),
));
commit_export_transaction(
&staging_dir,
&states_dir,
&keep,
transaction_id,
replacements,
)?;
Ok(states_dir)
}
#[cfg(test)]
mod tests {
use std::fs;
use super::*;
use bsk_protocol::tools::{
NavigationCause, PageRefV2 as PageRef, RecorderInfo, StepCommonV3 as StepCommon,
StepResultV3 as StepResult, StepV3 as Step, StopReason, TRACE_VERSION_V2, TRACE_VERSION_V3,
TraceEntry, TraceStateV3, TraceV2, TraceV3, VOM_FORMAT_VERSION,
};
fn sample_trace(state_id: &str, body: &str) -> TraceV3 {
TraceV3 {
version: TRACE_VERSION_V3,
recorded_at: "2026-08-10T02:12:55.360Z".into(),
stopped_by: StopReason::UserFinish,
entry: TraceEntry {
start_url: "https://example.com/".into(),
},
recorder: RecorderInfo {
bsk: "0.1.10".into(),
vom: VOM_FORMAT_VERSION,
},
states: vec![TraceStateV3 {
id: state_id.into(),
url: "https://example.com/".into(),
title: Some("Example".into()),
body: body.into(),
truncated: false,
}],
steps: vec![Step::Navigate {
common: StepCommon {
id: 1,
state: state_id.into(),
result: StepResult {
state: state_id.into(),
},
},
to: "https://example.com/".into(),
cause: NavigationCause::UserTyped,
}],
purpose: None,
started_at: None,
}
}
#[test]
fn write_trace_bundle_writes_dir_layout() {
let dir = tempfile::tempdir().unwrap();
let output = dir.path().join("flow");
let trace = sample_trace("s1", "@vom 1\nL1 page");
let states_dir = write_trace_bundle(&output, &trace).unwrap();
let trace_path = output.join("trace.json");
assert!(trace_path.exists());
assert_eq!(states_dir, output.join("states"));
assert!(states_dir.join("s1.txt").exists());
let disk: serde_json::Value =
serde_json::from_str(&fs::read_to_string(&trace_path).unwrap()).unwrap();
assert_eq!(disk["states"][0]["page"], "s1.txt");
assert!(disk["states"][0].get("body").is_none());
assert_eq!(
fs::read_to_string(states_dir.join("s1.txt")).unwrap(),
"@vom 1\nL1 page"
);
// Re-exporting drops states from the previous run but keeps the
// directory in place for the concurrent writer.
fs::write(states_dir.join("s9.txt"), "stale").unwrap();
fs::write(states_dir.join("notes.md"), "keep me").unwrap();
fs::write(states_dir.join("reference.txt"), "user file").unwrap();
write_trace_bundle(&output, &trace).unwrap();
assert!(!states_dir.join("s9.txt").exists());
assert!(states_dir.join("s1.txt").exists());
assert!(states_dir.join("notes.md").exists());
assert_eq!(
fs::read_to_string(states_dir.join("reference.txt")).unwrap(),
"user file"
);
}
#[test]
fn write_trace_bundle_rejects_traversal_state_id() {
let dir = tempfile::tempdir().unwrap();
let output = dir.path().join("flow");
let trace = sample_trace("../escaped", "@vom 1\nL1 page");
let err = write_trace_bundle(&output, &trace).unwrap_err();
assert!(err.to_string().contains("state id"));
assert!(!output.join("escaped.txt").exists());
assert!(!output.join("trace.json").exists());
}
#[test]
fn write_trace_bundle_rejects_duplicate_state_id() {
let dir = tempfile::tempdir().unwrap();
let output = dir.path().join("flow");
let mut trace = sample_trace("s1", "@vom 1\nL1 first");
trace.states.push(TraceStateV3 {
body: "@vom 1\nL1 duplicate".into(),
..trace.states[0].clone()
});
let err = write_trace_bundle(&output, &trace).unwrap_err();
assert!(err.to_string().contains("duplicate state id"));
assert!(!output.exists());
}
#[test]
fn unsupported_version_preserves_existing_bundle() {
let dir = tempfile::tempdir().unwrap();
let output = dir.path().join("flow");
write_trace_bundle(&output, &sample_trace("s1", "original")).unwrap();
let original_json = fs::read(output.join("trace.json")).unwrap();
let mut replacement = sample_trace("s2", "replacement");
replacement.version += 1;
let err = write_trace_bundle(&output, &replacement).unwrap_err();
assert!(err.to_string().contains("not supported"));
assert_eq!(fs::read(output.join("trace.json")).unwrap(), original_json);
assert_eq!(
fs::read_to_string(output.join("states/s1.txt")).unwrap(),
"original"
);
assert!(!output.join("states/s2.txt").exists());
}
#[test]
fn state_replacement_failure_preserves_existing_bundle() {
let dir = tempfile::tempdir().unwrap();
let output = dir.path().join("flow");
write_trace_bundle(&output, &sample_trace("s1", "original")).unwrap();
let original_json = fs::read(output.join("trace.json")).unwrap();
fs::write(output.join("states/notes.md"), "keep me").unwrap();
fs::create_dir(output.join("states/s2.txt")).unwrap();
let err = write_trace_bundle(&output, &sample_trace("s2", "replacement")).unwrap_err();
assert!(err.to_string().contains("state"));
assert_eq!(fs::read(output.join("trace.json")).unwrap(), original_json);
assert_eq!(
fs::read_to_string(output.join("states/s1.txt")).unwrap(),
"original"
);
assert_eq!(
fs::read_to_string(output.join("states/notes.md")).unwrap(),
"keep me"
);
}
#[test]
fn stale_state_cleanup_failure_rolls_back_replacement() {
let dir = tempfile::tempdir().unwrap();
let output = dir.path().join("flow");
write_trace_bundle(&output, &sample_trace("s1", "original")).unwrap();
let original_json = fs::read(output.join("trace.json")).unwrap();
fs::create_dir(output.join("states/s9.txt")).unwrap();
let err = write_trace_bundle(&output, &sample_trace("s2", "replacement")).unwrap_err();
assert!(err.to_string().contains("stale state"));
assert_eq!(fs::read(output.join("trace.json")).unwrap(), original_json);
assert_eq!(
fs::read_to_string(output.join("states/s1.txt")).unwrap(),
"original"
);
assert!(!output.join("states/s2.txt").exists());
assert!(output.join("states/s9.txt").is_dir());
}
#[cfg(unix)]
#[test]
fn write_trace_bundle_rejects_symlinked_states_directory() {
use std::os::unix::fs::symlink;
let dir = tempfile::tempdir().unwrap();
let output = dir.path().join("flow");
let external_states = dir.path().join("external-states");
fs::create_dir_all(&output).unwrap();
fs::create_dir_all(&external_states).unwrap();
fs::write(external_states.join("s9.txt"), "external").unwrap();
symlink(&external_states, output.join("states")).unwrap();
let err = write_trace_bundle(&output, &sample_trace("s1", "replacement")).unwrap_err();
assert!(err.to_string().contains("symlink"));
assert_eq!(
fs::read_to_string(external_states.join("s9.txt")).unwrap(),
"external"
);
assert!(!external_states.join("s1.txt").exists());
assert!(!output.join("trace.json").exists());
}
#[cfg(unix)]
#[test]
fn write_trace_bundle_rejects_symlinked_output_directory() {
use std::os::unix::fs::symlink;
let dir = tempfile::tempdir().unwrap();
let external_output = dir.path().join("external-output");
let output = dir.path().join("flow");
fs::create_dir_all(&external_output).unwrap();
fs::write(external_output.join("notes.md"), "external").unwrap();
symlink(&external_output, &output).unwrap();
let err = write_trace_bundle(&output, &sample_trace("s1", "replacement")).unwrap_err();
assert!(err.to_string().contains("symlink"));
assert_eq!(
fs::read_to_string(external_output.join("notes.md")).unwrap(),
"external"
);
assert!(!external_output.join("trace.json").exists());
assert!(!external_output.join("states").exists());
}
#[test]
fn write_trace_v2_writes_single_json_and_prunes_stale_states() {
let dir = tempfile::tempdir().unwrap();
let output = dir.path().join("flow");
fs::create_dir_all(output.join("states")).unwrap();
fs::write(output.join("states/s1.txt"), "stale").unwrap();
let trace = TraceV2 {
recorded_at: "2026-08-10T02:12:55.360Z".into(),
started_at: None,
purpose: None,
entry: bsk_protocol::tools::TraceEntry {
start_url: "https://example.com/".into(),
},
pages: vec![PageRef {
id: "p1".into(),
url: "https://example.com/".into(),
title: None,
}],
steps: vec![],
};
write_trace_v2(&output, &trace).unwrap();
assert!(output.join("trace.json").exists());
assert!(!output.join("states/s1.txt").exists());
}
#[test]
fn write_trace_v2_stale_cleanup_failure_rolls_back_trace() {
let dir = tempfile::tempdir().unwrap();
let output = dir.path().join("flow");
write_trace_bundle(&output, &sample_trace("s1", "original")).unwrap();
let original_json = fs::read(output.join("trace.json")).unwrap();
fs::create_dir(output.join("states/s9.txt")).unwrap();
let trace = TraceV2 {
recorded_at: "2026-08-10T02:12:55.360Z".into(),
started_at: None,
purpose: None,
entry: bsk_protocol::tools::TraceEntry {
start_url: "https://example.com/".into(),
},
pages: vec![],
steps: vec![],
};
let err = write_trace_v2(&output, &trace).unwrap_err();
assert!(err.to_string().contains("stale state"));
assert_eq!(fs::read(output.join("trace.json")).unwrap(), original_json);
assert_eq!(
fs::read_to_string(output.join("states/s1.txt")).unwrap(),
"original"
);
assert!(output.join("states/s9.txt").is_dir());
}
#[test]
fn export_recorded_trace_accepts_v2() {
let dir = tempfile::tempdir().unwrap();
let output = dir.path().join("flow");
let trace = RecordedTrace::V2(TraceV2 {
recorded_at: "2026-08-10T02:12:55.360Z".into(),
started_at: None,
purpose: None,
entry: bsk_protocol::tools::TraceEntry {
start_url: "https://example.com/".into(),
},
pages: vec![],
steps: vec![],
});
let exported = export_recorded_trace(&output, &trace).unwrap();
assert_eq!(exported.trace_version, TRACE_VERSION_V2);
assert!(exported.states_dir.is_none());
assert!(output.join("trace.json").exists());
}
#[test]
fn existing_trace_json_file_is_rejected_with_bundle_usage() {
let dir = tempfile::tempdir().unwrap();
let output = dir.path().join("trace.json");
fs::write(&output, "{}\n").unwrap();
let err = validate_record_output(&output).unwrap_err();
let msg = err.to_string();
assert!(
msg.contains("JSON file") || msg.contains("json file"),
"should name the legacy JSON file usage: {msg}"
);
assert!(
msg.contains("states/") || msg.contains("--output trace"),
"should hint at the new bundle directory usage: {msg}"
);
assert_eq!(fs::read_to_string(&output).unwrap(), "{}\n");
}
#[test]
fn missing_trace_json_path_is_rejected_as_legacy_output() {
let err = validate_record_output(Path::new("trace.json")).unwrap_err();
let msg = err.to_string();
assert!(
msg.contains("JSON file") || msg.contains("json file"),
"{msg}"
);
assert!(
msg.contains("states/") || msg.contains("--output trace"),
"{msg}"
);
}
#[test]
fn export_failure_keeps_recovery_for_retry() {
let _lock = crate::cli::record_recovery::test_env_lock();
let tmp = tempfile::tempdir().unwrap();
unsafe {
std::env::set_var(crate::daemon::paths::BSK_HOME_ENV, tmp.path());
}
let dir = tempfile::tempdir().unwrap();
let bad_output = dir.path().join("trace.json");
fs::write(&bad_output, "{}\n").unwrap();
let trace = RecordedTrace::V3(sample_trace("s1", "@vom 1\nL1 page"));
let err = export_with_recovery(&bad_output, &trace).unwrap_err();
let msg = err.to_string();
assert!(
msg.contains("recover") || msg.contains("record stop"),
"export failure should tell the user how to recover: {msg}"
);
assert!(
crate::cli::record_recovery::exists(),
"recovery data must survive a failed bundle export"
);
let recovered = crate::cli::record_recovery::load()
.unwrap()
.expect("recovery file");
assert_eq!(recovered, trace);
assert_eq!(fs::read_to_string(&bad_output).unwrap(), "{}\n");
let good_output = dir.path().join("bundle");
export_with_recovery(&good_output, &recovered).unwrap();
assert!(good_output.join("trace.json").exists());
assert!(good_output.join("states/s1.txt").exists());
assert!(
!crate::cli::record_recovery::exists(),
"successful export must delete recovery data"
);
unsafe {
std::env::remove_var(crate::daemon::paths::BSK_HOME_ENV);
}
}
}
+112
View File
@@ -0,0 +1,112 @@
//! Persist a completed Trace so a failed bundle export can be retried.
use std::fs;
use anyhow::{Context, Result};
use bsk_protocol::tools::RecordedTrace;
use crate::daemon::paths;
pub fn save(trace: &RecordedTrace) -> Result<()> {
let path = paths::record_recovery_path()?;
paths::ensure_bsk_home()?;
let json = serde_json::to_string_pretty(trace).context("serialize record recovery")?;
let tmp = path.with_extension("tmp");
fs::write(&tmp, format!("{json}\n")).with_context(|| format!("write {}", tmp.display()))?;
fs::rename(&tmp, &path).with_context(|| format!("commit {}", path.display()))?;
Ok(())
}
pub fn load() -> Result<Option<RecordedTrace>> {
let path = paths::record_recovery_path()?;
if !path.exists() {
return Ok(None);
}
let raw = fs::read_to_string(&path).with_context(|| format!("read {}", path.display()))?;
let trace = serde_json::from_str(&raw).with_context(|| format!("parse {}", path.display()))?;
Ok(Some(trace))
}
pub fn exists() -> bool {
paths::record_recovery_path()
.map(|path| path.exists())
.unwrap_or(false)
}
pub fn clear() {
if let Ok(path) = paths::record_recovery_path() {
let _ = fs::remove_file(path);
}
}
#[cfg(test)]
pub(crate) fn test_env_lock() -> std::sync::MutexGuard<'static, ()> {
crate::daemon::paths::test_env_lock()
}
#[cfg(test)]
mod tests {
use super::*;
use bsk_protocol::tools::{
RecorderInfo, StopReason, TRACE_VERSION_V3, TraceEntry, TraceStateV3, TraceV3,
VOM_FORMAT_VERSION,
};
fn sample_trace() -> RecordedTrace {
RecordedTrace::V3(TraceV3 {
version: TRACE_VERSION_V3,
recorded_at: "2026-08-10T02:12:55.360Z".into(),
stopped_by: StopReason::UserFinish,
entry: TraceEntry {
start_url: "https://example.com/".into(),
},
recorder: RecorderInfo {
bsk: "0.1.10".into(),
vom: VOM_FORMAT_VERSION,
},
states: vec![TraceStateV3 {
id: "s1".into(),
url: "https://example.com/".into(),
title: Some("Example".into()),
body: "@vom 1\nL1 page".into(),
truncated: false,
}],
steps: vec![],
purpose: None,
started_at: None,
})
}
fn with_temp_home<F: FnOnce()>(f: F) {
let _lock = test_env_lock();
let tmp = tempfile::tempdir().unwrap();
unsafe {
std::env::set_var(crate::daemon::paths::BSK_HOME_ENV, tmp.path());
}
f();
clear();
unsafe {
std::env::remove_var(crate::daemon::paths::BSK_HOME_ENV);
}
}
#[test]
fn save_load_round_trips() {
with_temp_home(|| {
let trace = sample_trace();
save(&trace).unwrap();
assert!(exists());
assert_eq!(load().unwrap(), Some(trace));
});
}
#[test]
fn clear_removes_recovery_file() {
with_temp_home(|| {
save(&sample_trace()).unwrap();
clear();
assert!(!exists());
assert_eq!(load().unwrap(), None);
});
}
}
+1 -7
View File
@@ -59,15 +59,9 @@ fn started_at_unix_ms() -> String {
#[cfg(test)]
mod tests {
use super::*;
use std::sync::Mutex;
fn env_guard() -> &'static Mutex<()> {
static GUARD: Mutex<()> = Mutex::new(());
&GUARD
}
fn with_temp_home<F: FnOnce()>(f: F) {
let _lock = env_guard().lock().unwrap_or_else(|e| e.into_inner());
let _lock = crate::daemon::paths::test_env_lock();
let tmp = tempfile::tempdir().unwrap();
unsafe {
std::env::set_var(crate::daemon::paths::BSK_HOME_ENV, tmp.path());
+27
View File
@@ -50,6 +50,7 @@ pub mod reason {
pub const BORROW_CONFLICT: &str = "borrow_conflict";
pub const SCREENSHOT_CAPTURE_FAILED: &str = "screenshot_capture_failed";
pub const SESSION_BUSY: &str = crate::rpc_reason::SESSION_BUSY;
pub const RECORD_START_PAGE_UNREACHABLE: &str = "record_start_page_unreachable";
}
/// Per-code rendering metadata. The `summary` is what the CLI prints
@@ -269,6 +270,16 @@ pub fn info_for_error(code: ErrorCode, data: Option<&serde_json::Value>) -> Rend
),
exit_code: base.exit_code,
},
(
ErrorCode::CdpFailed | ErrorCode::Timeout | ErrorCode::InvalidParams,
reason::RECORD_START_PAGE_UNREACHABLE,
) => RenderInfo {
summary: "default start page https://example.com/ could not be opened",
hint: Some(
"that site may be unreachable from your network. Pass --url with a page you can open, e.g. `bsk record start --url https://<site>/`",
),
exit_code: base.exit_code,
},
_ => base,
}
}
@@ -378,6 +389,22 @@ mod tests {
assert_eq!(info.exit_code, 4);
}
#[test]
fn record_start_page_unreachable_overrides_cdp_failed_copy() {
let data = serde_json::json!({ "reason": reason::RECORD_START_PAGE_UNREACHABLE });
let info = info_for_error(ErrorCode::CdpFailed, Some(&data));
assert_eq!(
info.summary,
"default start page https://example.com/ could not be opened"
);
assert!(
info.hint.unwrap().contains("--url"),
"expected --url hint, got {:?}",
info.hint
);
assert_eq!(info.exit_code, 3);
}
#[test]
fn screenshot_capture_failed_overrides_cdp_failed_copy() {
let data = serde_json::json!({ "reason": reason::SCREENSHOT_CAPTURE_FAILED });
+21 -9
View File
@@ -100,6 +100,14 @@ pub fn record_session_path() -> Result<PathBuf> {
Ok(bsk_home()?.join("record-session.json"))
}
/// Path to a completed Trace saved before bundle export (`record-recovery.json`).
///
/// Kept until export succeeds so a failed `--output` write cannot drop the
/// recording the extension already returned.
pub fn record_recovery_path() -> Result<PathBuf> {
Ok(bsk_home()?.join("record-recovery.json"))
}
/// Windows named-pipe name. Include the resolved `BSK_HOME` path in the
/// token so test homes and custom installs do not share a predictable
/// per-username pipe.
@@ -119,6 +127,13 @@ pub fn pipe_name() -> String {
render_pipe_name(&user, &home)
}
/// Process-wide lock for tests that mutate `BSK_HOME`.
#[cfg(test)]
pub(crate) fn test_env_lock() -> std::sync::MutexGuard<'static, ()> {
static GUARD: std::sync::Mutex<()> = std::sync::Mutex::new(());
GUARD.lock().unwrap_or_else(|e| e.into_inner())
}
/// Render the Windows named-pipe name for a user/home pair. Kept
/// platform-agnostic so unit tests on any host can pin the output
/// character set.
@@ -136,19 +151,12 @@ fn render_pipe_name(user: &str, home: &str) -> String {
#[cfg(test)]
mod tests {
use super::*;
use std::sync::Mutex;
use tempfile::TempDir;
/// Serialise tests that mutate the global `BSK_HOME` env var.
fn env_guard() -> &'static Mutex<()> {
static GUARD: Mutex<()> = Mutex::new(());
&GUARD
}
fn with_temp_home<F: FnOnce(&Path)>(f: F) {
let _lock = env_guard().lock().unwrap_or_else(|e| e.into_inner());
let _lock = test_env_lock();
let tmp = TempDir::new().unwrap();
// SAFETY: serialised by env_guard above.
// SAFETY: serialised by test_env_lock above.
unsafe {
std::env::set_var(BSK_HOME_ENV, tmp.path().join("bsk"));
}
@@ -187,6 +195,10 @@ mod tests {
record_session_path().unwrap(),
home.join("record-session.json")
);
assert_eq!(
record_recovery_path().unwrap(),
home.join("record-recovery.json")
);
});
}
+2 -2
View File
@@ -26,7 +26,7 @@ use crate::daemon::{
browsers::{BROWSER_LIVENESS_TICK, BROWSER_LIVENESS_TIMEOUT, EXTENSION_CONNECT_WAIT},
info as daemon_info, ipc, lockfile, paths,
sessions::{StopSessionError, forget_session, stop_session},
state::DaemonState,
state::{DaemonState, PROTOCOL_VERSION},
ws,
};
@@ -297,7 +297,7 @@ pub fn run_foreground(cfg: DaemonConfig) -> Result<()> {
ws_port,
sock_path: sock_path.clone(),
daemon_version: env!("CARGO_PKG_VERSION"),
protocol_version: "1.0",
protocol_version: PROTOCOL_VERSION,
};
let handler = ipc::full_handler(status, Arc::clone(&state));
+1 -1
View File
@@ -16,7 +16,7 @@ use super::start::DaemonConfig;
use super::ws::WsHandle;
pub const DAEMON_VERSION: &str = env!("CARGO_PKG_VERSION");
pub const PROTOCOL_VERSION: &str = "1.0";
pub const PROTOCOL_VERSION: &str = "1.1";
/// Lowest **protocol** version peers must speak (e.g. `"1.0"`).
pub const MIN_COMPATIBLE_PROTOCOL: &str = "1.0";
/// Legacy app-semver floor used only when `HandshakeResult.min_compatible_peer`
+10 -10
View File
@@ -108,12 +108,12 @@ async fn send_handshake_with_floors(
async fn handshake_ok_when_protocol_matches() {
let (handle, _sock) = spawn_daemon().await;
let mut ws = open_ws(handle.ws_addr()).await;
let resp = send_handshake(&mut ws, "1.0", env!("CARGO_PKG_VERSION")).await;
let resp = send_handshake(&mut ws, "1.1", env!("CARGO_PKG_VERSION")).await;
let result: HandshakeResult = match resp.body {
ResponseBody::Ok(v) => serde_json::from_value(v).unwrap(),
ResponseBody::Err(e) => panic!("expected ok handshake, got {e:?}"),
};
assert_eq!(result.protocol_version, "1.0");
assert_eq!(result.protocol_version, "1.1");
assert_eq!(
result
.min_compatible_peer
@@ -135,7 +135,7 @@ async fn handshake_ok_when_app_versions_differ_but_protocol_matches() {
let (handle, _sock) = spawn_daemon().await;
let mut ws = open_ws(handle.ws_addr()).await;
let resp =
send_handshake_with_floors(&mut ws, "1.0", "9.9.9", Some("0.0.0"), Some("1.0")).await;
send_handshake_with_floors(&mut ws, "1.1", "9.9.9", Some("0.0.0"), Some("1.1")).await;
match resp.body {
ResponseBody::Ok(_) => {}
other => panic!("expected ok when protocol matches, got {other:?}"),
@@ -149,10 +149,10 @@ async fn handshake_skew_when_protocol_minor_differs() {
let mut ws = open_ws(handle.ws_addr()).await;
let resp = send_handshake_with_floors(
&mut ws,
"1.1",
"1.2",
env!("CARGO_PKG_VERSION"),
Some("0.0.0"),
Some("1.0"),
Some("1.1"),
)
.await;
match resp.body {
@@ -210,7 +210,7 @@ async fn handshake_legacy_ext_without_protocol_floor_still_ok() {
let mut ws = open_ws(handle.ws_addr()).await;
let resp = send_handshake_with_floors(
&mut ws,
"1.0",
"1.1",
env!("CARGO_PKG_VERSION"),
Some("0.1.0"),
None,
@@ -235,7 +235,7 @@ async fn status_surfaces_version_skew_for_skewed_browser() {
browser_name: "chrome".into(),
browser_version: "131.0".into(),
extension_version: "9.9.9".into(),
extension_protocol_version: "1.1".into(),
extension_protocol_version: "1.2".into(),
label: "Older".into(),
sink: bsk::daemon::browsers::BrowserSink { tx },
pending: Mutex::new(bsk::daemon::browsers::Pending::default()),
@@ -263,8 +263,8 @@ async fn status_surfaces_version_skew_for_skewed_browser() {
.iter()
.find(|s| s.instance_id == "skew-only-test")
.expect("status must list our skew client");
assert_eq!(skew.client_protocol_version, "1.1");
assert_eq!(skew.server_protocol_version, "1.0");
assert_eq!(skew.client_protocol_version, "1.2");
assert_eq!(skew.server_protocol_version, "1.1");
assert_eq!(skew.client_version, "9.9.9");
let entry = status
.browsers
@@ -281,7 +281,7 @@ async fn handshake_rejects_when_local_below_peer_min_compatible_protocol() {
let mut ws = open_ws(handle.ws_addr()).await;
let resp = send_handshake_with_floors(
&mut ws,
"1.0",
"1.1",
env!("CARGO_PKG_VERSION"),
Some("0.0.0"),
Some("99.0.0"),
@@ -0,0 +1,327 @@
use std::path::{Path, PathBuf};
use std::sync::Arc;
use std::sync::atomic::{AtomicUsize, Ordering};
use std::time::Duration;
use bsk::daemon::info::DaemonInfo;
use bsk::daemon::paths::BSK_HOME_ENV;
use bsk::daemon::{self, DaemonConfig};
use bsk::ipc_client::IpcClient;
use bsk_protocol::system::{HandshakeParams, HandshakeResult};
use bsk_protocol::tools::{
RecordStopResult, RecordedTrace, RecorderInfo, SessionStartParams, SessionStartResult,
StopReason, TraceEntry, TraceStateV3, TraceV3, VOM_FORMAT_VERSION,
};
use bsk_protocol::{BrowserPeerInfo, Frame, Method, RequestFrame, ResponseBody, ResponseFrame};
use futures_util::{SinkExt, StreamExt};
use serde_json::json;
use tokio::process::Command;
use tokio_tungstenite::tungstenite::handshake::client::generate_key;
use tokio_tungstenite::tungstenite::http::Request;
use tokio_tungstenite::tungstenite::protocol::Message;
const TEST_EXT_ID: &str = "abcdefghijklmnopabcdefghijklmnop";
type Ws =
tokio_tungstenite::WebSocketStream<tokio_tungstenite::MaybeTlsStream<tokio::net::TcpStream>>;
fn bsk_bin() -> PathBuf {
PathBuf::from(env!("CARGO_BIN_EXE_bsk"))
}
async fn connect_extension(addr: std::net::SocketAddr) -> Ws {
let origin = format!("chrome-extension://{TEST_EXT_ID}");
let request = Request::builder()
.method("GET")
.uri(format!("ws://{addr}/"))
.header("Host", addr.to_string())
.header("Upgrade", "websocket")
.header("Connection", "Upgrade")
.header("Sec-WebSocket-Version", "13")
.header("Sec-WebSocket-Key", generate_key())
.header("Origin", origin)
.body(())
.unwrap();
tokio_tungstenite::connect_async(request).await.unwrap().0
}
async fn handshake(ws: &mut Ws) {
let request = RequestFrame {
id: "handshake".into(),
method: Method::SystemHandshake,
params: Some(
serde_json::to_value(HandshakeParams {
client: "browser-skill-extension".into(),
version: "0.1.10".parse().unwrap(),
protocol_version: "1.1".into(),
instance_id: TEST_EXT_ID.into(),
browser: BrowserPeerInfo {
name: "chrome".into(),
version: "131.0".into(),
},
min_compatible_peer: Some("0.1.10".parse().unwrap()),
min_compatible_protocol: Some("1.1".into()),
label: "Record export recovery test".into(),
})
.unwrap(),
),
};
ws.send(Message::Text(serde_json::to_string(&request).unwrap()))
.await
.unwrap();
let Message::Text(response) = ws.next().await.unwrap().unwrap() else {
panic!("expected text handshake response");
};
let response: ResponseFrame = serde_json::from_str(&response).unwrap();
let ResponseBody::Ok(value) = response.body else {
panic!("handshake failed");
};
let _: HandshakeResult = serde_json::from_value(value).unwrap();
}
fn sample_trace() -> TraceV3 {
TraceV3 {
version: 3,
purpose: Some("export recovery".into()),
started_at: Some("2026-08-12T10:00:00Z".into()),
recorded_at: "2026-08-12T10:01:00Z".into(),
stopped_by: StopReason::CliStop,
entry: TraceEntry {
start_url: "https://example.com/".into(),
},
recorder: RecorderInfo {
bsk: "0.1.10".into(),
vom: VOM_FORMAT_VERSION,
},
states: vec![TraceStateV3 {
id: "s1".into(),
url: "https://example.com/".into(),
title: Some("Example".into()),
body: "@vom 1\nL1 page".into(),
truncated: false,
}],
steps: vec![],
}
}
fn run_extension(
mut ws: Ws,
record_stop_calls: Arc<AtomicUsize>,
session_stop_calls: Arc<AtomicUsize>,
) {
tokio::spawn(async move {
while let Some(Ok(Message::Text(text))) = ws.next().await {
let Ok(Frame::Request(request)) = serde_json::from_str::<Frame>(&text) else {
continue;
};
let body = match request.method {
Method::ToolSessionStart => {
let _: SessionStartParams =
serde_json::from_value(request.params.clone().unwrap()).unwrap();
ResponseBody::Ok(
serde_json::to_value(SessionStartResult {
agent_window_id: Some(100),
})
.unwrap(),
)
}
Method::ToolRecordStop => {
record_stop_calls.fetch_add(1, Ordering::SeqCst);
ResponseBody::Ok(
serde_json::to_value(RecordStopResult {
trace: RecordedTrace::V3(sample_trace()),
})
.unwrap(),
)
}
Method::ToolSessionStop => {
session_stop_calls.fetch_add(1, Ordering::SeqCst);
ResponseBody::Ok(json!({}))
}
_ => ResponseBody::Err(bsk_protocol::RpcError {
code: bsk_protocol::ErrorCode::ProtocolError,
message: format!("unexpected method {:?}", request.method),
data: None,
}),
};
let response = ResponseFrame {
id: request.id,
body,
};
ws.send(Message::Text(serde_json::to_string(&response).unwrap()))
.await
.unwrap();
}
});
}
async fn start_session(sock: &Path) -> String {
#[derive(serde::Deserialize)]
struct StartResult {
session_id: String,
}
let mut client = IpcClient::connect(sock).await.unwrap();
client
.call::<(), StartResult>(
"session-start",
Method::SessionStart,
None,
Duration::from_secs(5),
)
.await
.unwrap()
.unwrap()
.session_id
}
fn write_record_session(home: &Path, session_id: &str) {
std::fs::create_dir_all(home).unwrap();
std::fs::write(
home.join("record-session.json"),
format!("{{\"session_id\":\"{session_id}\",\"started_at\":\"1\"}}\n"),
)
.unwrap();
}
async fn run_record_stop(home: &Path, output: &Path) -> std::process::Output {
Command::new(bsk_bin())
.arg("record")
.arg("stop")
.arg("--output")
.arg(output)
.env(BSK_HOME_ENV, home)
.output()
.await
.unwrap()
}
async fn setup_daemon(temp: &Path) -> (bsk::daemon::DaemonHandle, PathBuf, PathBuf) {
let home = temp.join("bsk-home");
let sock = temp.join("daemon.sock");
let daemon = daemon::run(DaemonConfig::new(0), Some(sock.clone()))
.await
.unwrap();
bsk::daemon::info::write_to_path(
&DaemonInfo::now(
std::process::id(),
sock.clone(),
daemon.ws_addr().port(),
env!("CARGO_PKG_VERSION"),
),
&home.join("daemon.json"),
)
.unwrap();
(daemon, home, sock)
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn existing_trace_json_file_is_rejected_before_record_stop() {
let temp = tempfile::Builder::new()
.prefix("bsk-record-json-")
.tempdir()
.unwrap();
let (daemon, home, sock) = setup_daemon(temp.path()).await;
let output = temp.path().join("trace.json");
std::fs::write(&output, "{}\n").unwrap();
let mut extension = connect_extension(daemon.ws_addr()).await;
handshake(&mut extension).await;
let record_stop_calls = Arc::new(AtomicUsize::new(0));
let session_stop_calls = Arc::new(AtomicUsize::new(0));
run_extension(
extension,
Arc::clone(&record_stop_calls),
Arc::clone(&session_stop_calls),
);
let session_id = start_session(&sock).await;
write_record_session(&home, &session_id);
let first = run_record_stop(&home, &output).await;
let stderr = String::from_utf8_lossy(&first.stderr);
assert!(
!first.status.success(),
"legacy JSON --output should fail: {stderr}"
);
assert!(
stderr.contains("JSON file") || stderr.contains("bundle directory"),
"{stderr}"
);
let state: serde_json::Value =
serde_json::from_str(&std::fs::read_to_string(home.join("record-session.json")).unwrap())
.unwrap();
assert_eq!(state["session_id"], session_id);
assert_eq!(
record_stop_calls.load(Ordering::SeqCst),
0,
"existing trace.json must be rejected before the extension returns the Trace"
);
assert_eq!(session_stop_calls.load(Ordering::SeqCst), 0);
assert_eq!(std::fs::read_to_string(&output).unwrap(), "{}\n");
daemon.shutdown().await;
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn failed_bundle_export_can_be_recovered() {
let temp = tempfile::Builder::new()
.prefix("bsk-record-recov-")
.tempdir()
.unwrap();
let (daemon, home, sock) = setup_daemon(temp.path()).await;
let output = temp.path().join("trace");
std::fs::create_dir_all(&output).unwrap();
std::fs::write(output.join("states"), "not a directory\n").unwrap();
let mut extension = connect_extension(daemon.ws_addr()).await;
handshake(&mut extension).await;
let record_stop_calls = Arc::new(AtomicUsize::new(0));
let session_stop_calls = Arc::new(AtomicUsize::new(0));
run_extension(
extension,
Arc::clone(&record_stop_calls),
Arc::clone(&session_stop_calls),
);
let session_id = start_session(&sock).await;
write_record_session(&home, &session_id);
let first = run_record_stop(&home, &output).await;
let stderr = String::from_utf8_lossy(&first.stderr);
assert!(
!first.status.success(),
"blocked states path should fail export: {stderr}"
);
assert!(
home.join("record-recovery.json").exists(),
"completed Trace must be saved before the failed export"
);
assert!(
!home.join("record-session.json").exists(),
"session state is cleared after the extension returns the Trace"
);
assert_eq!(record_stop_calls.load(Ordering::SeqCst), 1);
assert_eq!(session_stop_calls.load(Ordering::SeqCst), 1);
std::fs::remove_file(output.join("states")).unwrap();
let second = run_record_stop(&home, &output).await;
assert!(
second.status.success(),
"recovery export failed: {}",
String::from_utf8_lossy(&second.stderr)
);
assert_eq!(
record_stop_calls.load(Ordering::SeqCst),
1,
"recovery must not call the extension again"
);
assert!(output.join("trace.json").exists());
assert!(output.join("states/s1.txt").exists());
assert!(
!home.join("record-recovery.json").exists(),
"successful export must delete recovery data"
);
daemon.shutdown().await;
}
+271
View File
@@ -0,0 +1,271 @@
use std::path::{Path, PathBuf};
use std::sync::Arc;
use std::sync::atomic::{AtomicUsize, Ordering};
use std::time::Duration;
use bsk::cli::record_state;
use bsk::daemon::info::DaemonInfo;
use bsk::daemon::paths::BSK_HOME_ENV;
use bsk::daemon::{self, DaemonConfig};
use bsk::ipc_client::IpcClient;
use bsk_protocol::system::{HandshakeParams, HandshakeResult};
use bsk_protocol::tools::{
RecordStopResult, RecordedTrace, RecorderInfo, SessionStartParams, SessionStartResult,
StopReason, TraceEntry, TraceStateV3, TraceV3, VOM_FORMAT_VERSION,
};
use bsk_protocol::{
BrowserPeerInfo, ErrorCode, Frame, Method, RequestFrame, ResponseBody, ResponseFrame, RpcError,
};
use futures_util::{SinkExt, StreamExt};
use serde_json::json;
use tokio::process::Command;
use tokio_tungstenite::tungstenite::handshake::client::generate_key;
use tokio_tungstenite::tungstenite::http::Request;
use tokio_tungstenite::tungstenite::protocol::Message;
const TEST_EXT_ID: &str = "abcdefghijklmnopabcdefghijklmnop";
type Ws =
tokio_tungstenite::WebSocketStream<tokio_tungstenite::MaybeTlsStream<tokio::net::TcpStream>>;
fn bsk_bin() -> PathBuf {
PathBuf::from(env!("CARGO_BIN_EXE_bsk"))
}
async fn connect_extension(addr: std::net::SocketAddr) -> Ws {
let origin = format!("chrome-extension://{TEST_EXT_ID}");
let request = Request::builder()
.method("GET")
.uri(format!("ws://{addr}/"))
.header("Host", addr.to_string())
.header("Upgrade", "websocket")
.header("Connection", "Upgrade")
.header("Sec-WebSocket-Version", "13")
.header("Sec-WebSocket-Key", generate_key())
.header("Origin", origin)
.body(())
.unwrap();
tokio_tungstenite::connect_async(request).await.unwrap().0
}
async fn handshake(ws: &mut Ws) {
let request = RequestFrame {
id: "handshake".into(),
method: Method::SystemHandshake,
params: Some(
serde_json::to_value(HandshakeParams {
client: "browser-skill-extension".into(),
version: "0.1.10".parse().unwrap(),
protocol_version: "1.1".into(),
instance_id: TEST_EXT_ID.into(),
browser: BrowserPeerInfo {
name: "chrome".into(),
version: "131.0".into(),
},
min_compatible_peer: Some("0.1.10".parse().unwrap()),
min_compatible_protocol: Some("1.1".into()),
label: "Record stop retry test".into(),
})
.unwrap(),
),
};
ws.send(Message::Text(serde_json::to_string(&request).unwrap()))
.await
.unwrap();
let Message::Text(response) = ws.next().await.unwrap().unwrap() else {
panic!("expected text handshake response");
};
let response: ResponseFrame = serde_json::from_str(&response).unwrap();
let ResponseBody::Ok(value) = response.body else {
panic!("handshake failed");
};
let _: HandshakeResult = serde_json::from_value(value).unwrap();
}
fn sample_trace() -> TraceV3 {
TraceV3 {
version: 3,
purpose: Some("retry record stop".into()),
started_at: Some("2026-08-12T10:00:00Z".into()),
recorded_at: "2026-08-12T10:01:00Z".into(),
stopped_by: StopReason::CliStop,
entry: TraceEntry {
start_url: "https://example.com/".into(),
},
recorder: RecorderInfo {
bsk: "0.1.10".into(),
vom: VOM_FORMAT_VERSION,
},
states: vec![TraceStateV3 {
id: "s1".into(),
url: "https://example.com/".into(),
title: Some("Example".into()),
body: "@vom 1\nL1 page".into(),
truncated: false,
}],
steps: vec![],
}
}
fn run_extension(
mut ws: Ws,
record_stop_calls: Arc<AtomicUsize>,
session_stop_calls: Arc<AtomicUsize>,
) {
tokio::spawn(async move {
while let Some(Ok(Message::Text(text))) = ws.next().await {
let Ok(Frame::Request(request)) = serde_json::from_str::<Frame>(&text) else {
continue;
};
let body = match request.method {
Method::ToolSessionStart => {
let _: SessionStartParams =
serde_json::from_value(request.params.clone().unwrap()).unwrap();
ResponseBody::Ok(
serde_json::to_value(SessionStartResult {
agent_window_id: Some(100),
})
.unwrap(),
)
}
Method::ToolRecordStop => {
if record_stop_calls.fetch_add(1, Ordering::SeqCst) == 0 {
ResponseBody::Err(RpcError {
code: ErrorCode::ProtocolError,
message: "recording remains active; retry record stop".into(),
data: Some(json!({ "recording": true, "retryable": true })),
})
} else {
ResponseBody::Ok(
serde_json::to_value(RecordStopResult {
trace: RecordedTrace::V3(sample_trace()),
})
.unwrap(),
)
}
}
Method::ToolSessionStop => {
session_stop_calls.fetch_add(1, Ordering::SeqCst);
ResponseBody::Ok(json!({}))
}
_ => ResponseBody::Err(RpcError {
code: ErrorCode::ProtocolError,
message: format!("unexpected method {:?}", request.method),
data: None,
}),
};
let response = ResponseFrame {
id: request.id,
body,
};
ws.send(Message::Text(serde_json::to_string(&response).unwrap()))
.await
.unwrap();
}
});
}
async fn start_session(sock: &Path) -> String {
#[derive(serde::Deserialize)]
struct StartResult {
session_id: String,
}
let mut client = IpcClient::connect(sock).await.unwrap();
client
.call::<(), StartResult>(
"session-start",
Method::SessionStart,
None,
Duration::from_secs(5),
)
.await
.unwrap()
.unwrap()
.session_id
}
async fn run_record_stop(home: &Path, output: &Path) -> std::process::Output {
Command::new(bsk_bin())
.arg("record")
.arg("stop")
.arg("--output")
.arg(output)
.env(BSK_HOME_ENV, home)
.output()
.await
.unwrap()
}
#[tokio::test(flavor = "multi_thread", worker_threads = 2)]
async fn failed_record_stop_keeps_state_and_session_for_retry() {
// Keep the Unix-domain socket below macOS's SUN_LEN even from a deep worktree.
let temp = tempfile::Builder::new()
.prefix("bsk-record-stop-")
.tempdir()
.unwrap();
let home = temp.path().join("bsk-home");
let sock = temp.path().join("daemon.sock");
let output = temp.path().join("trace");
let daemon = daemon::run(DaemonConfig::new(0), Some(sock.clone()))
.await
.unwrap();
bsk::daemon::info::write_to_path(
&DaemonInfo::now(
std::process::id(),
sock.clone(),
daemon.ws_addr().port(),
env!("CARGO_PKG_VERSION"),
),
&home.join("daemon.json"),
)
.unwrap();
let mut extension = connect_extension(daemon.ws_addr()).await;
handshake(&mut extension).await;
let record_stop_calls = Arc::new(AtomicUsize::new(0));
let session_stop_calls = Arc::new(AtomicUsize::new(0));
run_extension(
extension,
Arc::clone(&record_stop_calls),
Arc::clone(&session_stop_calls),
);
let session_id = start_session(&sock).await;
unsafe {
std::env::set_var(BSK_HOME_ENV, &home);
}
record_state::write(&session_id).unwrap();
let first = run_record_stop(&home, &output).await;
assert!(
!first.status.success(),
"first stop unexpectedly succeeded: {}",
String::from_utf8_lossy(&first.stdout)
);
assert_eq!(record_state::read().unwrap().session_id, session_id);
assert_eq!(record_stop_calls.load(Ordering::SeqCst), 1);
assert_eq!(
session_stop_calls.load(Ordering::SeqCst),
0,
"failed record stop must leave the session active"
);
let second = run_record_stop(&home, &output).await;
assert!(
second.status.success(),
"retry failed: {}",
String::from_utf8_lossy(&second.stderr)
);
assert!(record_state::read().is_err());
assert_eq!(record_stop_calls.load(Ordering::SeqCst), 2);
assert_eq!(session_stop_calls.load(Ordering::SeqCst), 1);
assert!(output.join("trace.json").exists());
assert!(output.join("states/s1.txt").exists());
unsafe {
std::env::remove_var(BSK_HOME_ENV);
}
daemon.shutdown().await;
}
+1 -1
View File
@@ -61,7 +61,7 @@ fn bsk_status_json_returns_structured_payload() {
assert!(parsed["pid"].as_u64().unwrap() > 0);
assert!(!parsed["daemon_version"].as_str().unwrap().is_empty());
assert_eq!(parsed["protocol_version"], "1.0");
assert_eq!(parsed["protocol_version"], "1.1");
assert!(parsed["sock_path"].as_str().is_some());
assert_eq!(parsed["browsers"], serde_json::json!([]));
assert_eq!(parsed["sessions"], serde_json::json!([]));
+2 -2
View File
@@ -53,7 +53,7 @@ pub async fn send_handshake(
let params = HandshakeParams {
client: "browser-skill-extension".into(),
version: "0.1.0-dev.0".parse().unwrap(),
protocol_version: "1.0".into(),
protocol_version: "1.1".into(),
instance_id: instance_id.into(),
browser: BrowserPeerInfo {
name: "chrome".into(),
@@ -91,7 +91,7 @@ async fn ws_handshake_registers_browser_in_state() {
let mut ws = connect_ext(handle.ws_addr(), &origin).await;
let result = send_handshake(&mut ws, TEST_EXT_ID).await;
assert_eq!(result.server, "browser-skill-daemon");
assert_eq!(result.protocol_version, "1.0");
assert_eq!(result.protocol_version, "1.1");
let state = handle.state();
let browsers = state.browsers.snapshot();
@@ -480,6 +480,36 @@
}
}
},
{
"description": "Fields shared by every step variant (flattened in JSON).",
"type": "object",
"required": [
"id",
"op",
"result",
"state"
],
"properties": {
"id": {
"type": "integer",
"format": "uint32",
"minimum": 0.0
},
"op": {
"type": "string",
"enum": [
"switch_tab"
]
},
"result": {
"$ref": "#/definitions/StepResultV3"
},
"state": {
"description": "Observation id immediately before this action.",
"type": "string"
}
}
},
{
"description": "Fields shared by every step variant (flattened in JSON).",
"type": "object",
@@ -6,15 +6,36 @@
"session_id"
],
"properties": {
"max_page_tokens": {
"type": [
"integer",
"null"
],
"format": "uint32",
"minimum": 0.0
},
"purpose": {
"type": [
"string",
"null"
]
},
"redact_values": {
"type": [
"boolean",
"null"
]
},
"session_id": {
"type": "string"
},
"supports_tab_switch_steps": {
"description": "Client can decode the v3 `switch_tab` step variant.",
"type": [
"boolean",
"null"
]
},
"tab_id": {
"type": [
"integer",
@@ -22,6 +43,15 @@
],
"format": "int64"
},
"trace_version": {
"description": "Desired trace export format. Omitted means v2; `3` requests v3.",
"type": [
"integer",
"null"
],
"format": "uint32",
"minimum": 0.0
},
"url": {
"type": [
"string",
@@ -480,6 +480,36 @@
}
}
},
{
"description": "Fields shared by every step variant (flattened in JSON).",
"type": "object",
"required": [
"id",
"op",
"result",
"state"
],
"properties": {
"id": {
"type": "integer",
"format": "uint32",
"minimum": 0.0
},
"op": {
"type": "string",
"enum": [
"switch_tab"
]
},
"result": {
"$ref": "#/definitions/StepResultV3"
},
"state": {
"description": "Observation id immediately before this action.",
"type": "string"
}
}
},
{
"description": "Fields shared by every step variant (flattened in JSON).",
"type": "object",
+30
View File
@@ -469,6 +469,36 @@
}
}
},
{
"description": "Fields shared by every step variant (flattened in JSON).",
"type": "object",
"required": [
"id",
"op",
"result",
"state"
],
"properties": {
"id": {
"type": "integer",
"format": "uint32",
"minimum": 0.0
},
"op": {
"type": "string",
"enum": [
"switch_tab"
]
},
"result": {
"$ref": "#/definitions/StepResultV3"
},
"state": {
"description": "Observation id immediately before this action.",
"type": "string"
}
}
},
{
"description": "Fields shared by every step variant (flattened in JSON).",
"type": "object",
@@ -431,6 +431,36 @@
}
}
},
{
"description": "Fields shared by every step variant (flattened in JSON).",
"type": "object",
"required": [
"id",
"op",
"result",
"state"
],
"properties": {
"id": {
"type": "integer",
"format": "uint32",
"minimum": 0.0
},
"op": {
"type": "string",
"enum": [
"switch_tab"
]
},
"result": {
"$ref": "#/definitions/StepResultV3"
},
"state": {
"description": "Observation id immediately before this action.",
"type": "string"
}
}
},
{
"description": "Fields shared by every step variant (flattened in JSON).",
"type": "object",

Some files were not shown because too many files have changed in this diff Show More