Files
reverse-engineering-anything/tests/boundary/mcp/electronMcp.test.ts
T
de4579e993 fix(observation): keep incomplete cancellation observations (#5644)
* fix(observation): keep incomplete cancellation observations

Service projection treated a retained observation that failed the completed
schema as a malformed completed result. Cancellation became unreadable_output
and the partial was dropped. Incomplete observations now stay on the owner
error, including the exact abort reason.

* fix(observation): drop non-JSON incomplete observations

Keep a retained partial only when projectAnalysisError can serialize it.
Clear a non-JSON partial and record an AnalysisOutputError as a secondary
failure so cancellation and timeout stay the owner error.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(observation): validate incomplete partials without copying

jsonValueSchema.safeParse cloned a JSON-safe partial that projection
copies again. Use jsonValueValidationIssue so the gate only checks.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* test(ida): expect non-retryable unresolved procedure lookups

Main stops treating invalid_request as retryable. An unchanged missing
procedure stays invalid_request and is not marked retryable.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* test(observation): share incomplete-observation assertions

The new non-JSON cases repeated the same owner-error checks and pushed
duplication over the 3.5% limit. One expectation helper keeps those
checks without a second copy.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* test(observation): keep the assertion helper out of the package

The helper imports Vitest. Rename it to a fixture so the production
build excludes it and dist does not import a dev dependency.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(observation): clear non-JSON partials on cleanup primaries

Projection can return the cleanup primary after retry. Clearing only
the wrapper left that primary's partial in place, so projectAnalysisError
threw. The gate now clears every non-JSON partial in the chain.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(observation): freeze retained incomplete observations

A JSON partial stayed a live provider object, so a later mutation could
make projection throw. The gate now keeps one frozen snapshot and still
drops a non-JSON partial without copying it.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* refactor(observation): share web runtime partial retention

Observe and inspect used the same promote-or-keep path. One helper
owns that sequence so the two operations cannot drift.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(observation): snapshot a shared partial once

A cleanup wrapper and its primary can hold the same partial object.
The gate now reuses the first frozen snapshot for that object.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* fix(observation): parse a kept partial once

jsonValueSchema.safeParse validates and copies only a JSON partial.
The gate freezes that copy and no longer walks the same value first.

Co-authored-by: morluto <morluto@users.noreply.github.com>

* test(observation): share repeated observation fixtures

The incomplete-observation tests repeated the same provider setup and
an empty Electron coverage tail, which pushed duplication over 3.5%.
One coverage fixture and shared assertions bring the tree back under
the threshold.

Co-authored-by: morluto <morluto@users.noreply.github.com>

---------

Co-authored-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: morluto <morluto@users.noreply.github.com>
2026-10-21 04:17:46 +08:00

694 lines
24 KiB
TypeScript

import { parseEvidence } from "../../../src/domain/evidence.js";
import { ok as resultOk } from "../../../src/domain/result.js";
import { rm, symlink, writeFile } from "node:fs/promises";
import { basename, join } from "node:path";
import { pathToFileURL } from "node:url";
import { Client, InMemoryTransport } from "@modelcontextprotocol/client";
import { Ajv2020 } from "ajv/dist/2020.js";
import { afterEach, expect, it, vi } from "vitest";
import { z } from "zod";
import { createTestTempDirectory } from "../../fixtures/temporaryDirectory.js";
import { createStrippedAsarAddon } from "../../fixtures/strippedAsarAddon.js";
import { createTestBinarySession } from "../../fixtures/binarySession.js";
import type { ElectronActiveObservationPort } from "../../../src/application/javascript/ElectronActiveObservationPort.js";
import type { ElectronObservationPort } from "../../../src/application/javascript/ElectronObservationPort.js";
import { analyzeJavaScriptApplication } from "../../support/javascriptApplicationScope.js";
import { CdpElectronProvider } from "../../../src/browser/CdpElectronProvider.js";
import { createServer } from "../../../src/server/createServer.js";
import { observed } from "../../fixtures/analysisExecution.js";
import {
startFakeCdpBrowser,
type FakeCdpBrowser,
} from "../../fixtures/fakeCdpBrowser.js";
import { writeElectronBoundaryFixture } from "../../fixtures/electronBoundaryApplication.js";
import { createElectronActiveObservationFixtureResult } from "../../../src/domain/javascript/electronActiveObservation.fixture.js";
import { parseMcpToolError } from "../../fixtures/mcpToolError.js";
import { AsarArtifactReader } from "../../../src/artifacts/AsarArtifactReader.js";
import { toolAvailabilityInventory } from "../../../src/contracts/toolOutputSchemaPrimitives.js";
import { AnalysisCancelledError } from "../../../src/domain/analysisErrorCore.js";
import { emptyElectronPageCoverage } from "../../../src/domain/javascript/electronPageInspection.fixture.js";
import type { ElectronPageInspection } from "../../../src/domain/javascript/electronObservation.js";
import { err } from "../../../src/domain/result.js";
const browsers: FakeCdpBrowser[] = [];
const resources: Array<{ close(): Promise<unknown> }> = [];
const temporary: string[] = [];
it("retains failed JavaScript inventory cleanup through server shutdown retries", async () => {
const { archive } = await createStrippedAsarAddon();
const session = createTestBinarySession(() => ({
execute: () => Promise.resolve(observed(null)),
close: () => Promise.resolve(resultOk(null)),
}));
const server = createServer({ kind: "session", session });
const client = new Client({ name: "javascript-cleanup-owner", version: "1" });
const [clientTransport, serverTransport] =
InMemoryTransport.createLinkedPair();
const originalClose = AsarArtifactReader.prototype.close;
const owners = new Set<AsarArtifactReader>();
let blocked = true;
const close = vi
.spyOn(AsarArtifactReader.prototype, "close")
.mockImplementation(function (this: AsarArtifactReader) {
owners.add(this);
if (blocked) return Promise.reject(new Error("snapshot cleanup blocked"));
return originalClose.call(this);
});
try {
await server.connect(serverTransport);
await client.connect(clientTransport);
const response = await client.callTool({
name: "analyze_javascript_application",
arguments: {
input_path: archive,
integrity_policy: "record-and-continue",
},
});
expect(parseMcpToolError(response)).toMatchObject({
error: {
code: "cleanup_incomplete",
details: {
resources: [expect.stringContaining(archive)],
partial_observation: { kind: "artifact-inventory" },
},
},
});
expect(owners.size).toBe(1);
await expect(server.close()).rejects.toThrow();
blocked = false;
await server.close();
expect(owners.size).toBe(1);
expect(close.mock.calls.length).toBeGreaterThanOrEqual(3);
} finally {
blocked = false;
await Promise.all([client.close(), server.close(), session.close()]);
close.mockRestore();
}
});
afterEach(async () => {
await Promise.all(resources.splice(0).map(async (item) => item.close()));
await Promise.all(browsers.splice(0).map(async (browser) => browser.close()));
await Promise.all(
temporary
.splice(0)
.map(async (path) => rm(path, { recursive: true, force: true })),
);
});
it("records completed Electron facts after final detach as server Evidence", async () => {
const inspection: ElectronPageInspection = {
browser: {
product: "Chromium",
protocol_version: "1.3",
revision: "123",
user_agent: "fixture",
js_version: "V8",
},
target: {
target_id: "target-1",
type: "page",
title: "fixture",
file_path: "/tmp/index.html",
attached: true,
},
...emptyElectronPageCoverage(1),
};
const provider: ElectronObservationPort = {
identity: () => ({
id: "test-electron-observation",
name: "Test Electron observation provider",
version: "1",
}),
listTargets: async () => {
throw new Error("unexpected target listing");
},
inspectPage: async () =>
err(
new AnalysisCancelledError("inspect_electron_page", {
partialObservation: {
kind: "electron-page-observation",
inspection,
cleanup: "transport-closed",
},
}),
),
};
const session = createTestBinarySession(() => ({
execute: () => Promise.resolve(observed(null)),
close: () => Promise.resolve(resultOk(null)),
}));
const server = createServer(
{ kind: "session", session },
{
electronObservation: provider,
},
);
const client = new Client({ name: "electron-final-detach", version: "1" });
const [clientTransport, serverTransport] =
InMemoryTransport.createLinkedPair();
resources.push(client, server, session);
await server.connect(serverTransport);
await client.connect(clientTransport);
const response = await client.callTool({
name: "inspect_electron_page",
arguments: {
cdp_endpoint: "http://127.0.0.1:9222",
target_id: "target-1",
observation_ms: 0,
include_script_sources: false,
},
});
const { error } = parseMcpToolError(response);
expect(error.details).toMatchObject({
partial_observation: { kind: "retained-evidence" },
});
const reference = z
.object({
partial_observation: z.object({
kind: z.literal("retained-evidence"),
evidence_id: z.string(),
}),
})
.parse(error.details).partial_observation;
expect(session.evidenceById(reference.evidence_id)).toMatchObject({
operation: "inspect_electron_page",
normalized_result: { target: { target_id: "target-1" } },
limitations: expect.arrayContaining([
expect.stringContaining("closed its CDP transport"),
]),
});
await expect(client.ping()).resolves.toEqual({});
});
it("lets MCP JavaScript analysis continue only with recorded untrusted ASAR bytes", async () => {
const { archive, addon } = await createStrippedAsarAddon();
const session = createTestBinarySession(() => ({
execute: () => Promise.resolve(observed(null)),
close: () => Promise.resolve(resultOk(null)),
}));
const server = createServer({ kind: "session", session });
const client = new Client({ name: "asar-integrity-policy", version: "1" });
const [clientTransport, serverTransport] =
InMemoryTransport.createLinkedPair();
resources.push(client, server, session);
await server.connect(serverTransport);
await client.connect(clientTransport);
const rejected = await client.callTool({
name: "analyze_javascript_application",
arguments: { input_path: archive },
});
expect(parseMcpToolError(rejected)).toMatchObject({
error: {
code: "artifact_integrity_mismatch",
remediation: {
action: expect.stringContaining("integrity_policy=record-and-continue"),
},
},
});
const continued = await client.callTool({
name: "analyze_javascript_application",
arguments: {
input_path: archive,
integrity_policy: "record-and-continue",
},
});
expect(continued.isError).not.toBe(true);
expect(continued.structuredContent).toMatchObject({
parameters: { integrity_policy: "record-and-continue" },
normalized_result: {
integrity_contradictions: [
{ logical_path: addon, trust: "observed-untrusted" },
],
graph: { coverage: { status: "partial" } },
},
});
});
it("advertises owned JavaScript analysis effects without foreign target execution", async () => {
const session = createTestBinarySession(() => ({
execute: () => Promise.resolve(observed(null)),
close: () => Promise.resolve(resultOk(null)),
}));
const server = createServer({ kind: "session", session });
const client = new Client({
name: "javascript-analysis-effects",
version: "1",
});
const [clientTransport, serverTransport] =
InMemoryTransport.createLinkedPair();
resources.push(client, server, session);
await server.connect(serverTransport);
await client.connect(clientTransport);
const status = await client.callTool({
name: "binary_session",
arguments: {},
});
expect(status.isError).not.toBe(true);
const { result } = z
.object({
result: z.object({ tool_availability: toolAvailabilityInventory }),
})
.parse(status.structuredContent);
const analysisAvailability = result.tool_availability.groups.find(
({ names }) => names.includes("analyze_javascript_application"),
);
expect(analysisAvailability?.effects.toSorted()).toEqual([
"idempotent",
"launchesProcess",
"mutatesSession",
"writesFilesystem",
]);
expect(analysisAvailability?.annotations.toSorted()).toEqual([
"idempotent",
"open_world",
]);
});
it("exposes endpoint-scoped Electron discovery and inspection as Evidence", async () => {
const root = await createTestTempDirectory("rea-electron-mcp-");
temporary.push(root);
await writeFile(join(root, "index.html"), "<script src='app.js'></script>");
await writeFile(
join(root, "app.js"),
"export const observed = 'source-secret';",
);
await writeFile(
join(root, "package.json"),
JSON.stringify({ name: "runtime-fixture", renderer: "index.html" }),
);
await writeFile(join(root, "worker.js"), "self.onmessage = () => {};\n");
const browser = await startFakeCdpBrowser({
electronFileUrl: pathToFileURL(join(root, "index.html")).href,
commandEvents: ({ method }) =>
method === "Debugger.enable"
? [
{
method: "Debugger.scriptParsed",
sessionId: "session-1",
params: {
scriptId: "partial-script",
url: pathToFileURL(join(root, "app.js")).href,
executionContextId: 1,
},
},
]
: undefined,
});
browsers.push(browser);
const session = createTestBinarySession(() => ({
execute: () => Promise.resolve(observed(null)),
close: () => Promise.resolve(resultOk(null)),
}));
const server = createServer(
{ kind: "session", session },
{
electronObservation: new CdpElectronProvider(),
availabilityPolicy: () => ({
processCaptureEnabled: false,
investigationInputRoots: 1,
}),
},
);
const client = new Client({ name: "electron-mcp-test", version: "1" });
const [clientTransport, serverTransport] =
InMemoryTransport.createLinkedPair();
resources.push(client, server, session);
await server.connect(serverTransport);
await client.connect(clientTransport);
const listed = await client.callTool({
name: "list_electron_targets",
arguments: {
cdp_endpoint: browser.endpoint,
},
});
expect(listed.isError).not.toBe(true);
expect(listed.structuredContent).toMatchObject({
normalized_result: {
targets: [{ target_id: "electron-page" }],
},
});
const inspected = await client.callTool({
name: "inspect_electron_page",
arguments: {
cdp_endpoint: browser.endpoint,
target_id: "electron-page",
observation_ms: 0,
include_script_sources: true,
},
});
const analyzed = await client.callTool({
name: "analyze_javascript_application",
arguments: { input_path: root },
});
expect(analyzed.isError).not.toBe(true);
expect(analyzed.structuredContent).toMatchObject({
normalized_result: {
graph: { nodes: expect.any(Array), edges: expect.any(Array) },
},
});
const analysisText = analyzed.content.find(({ type }) => type === "text");
expect(analysisText).toBeDefined();
if (analysisText?.type !== "text")
throw new TypeError("Missing JavaScript analysis text projection");
expect(analysisText.text).toContain('"nodes":[');
const reconciled = await client.callTool({
name: "reconcile_javascript_runtime",
arguments: {
static_layers: [
{
role: "application",
analysis: evidenceFor(analyzed.structuredContent),
},
],
runtime_observations: [evidenceFor(inspected.structuredContent)],
},
});
expect(reconciled.isError).not.toBe(true);
expect(reconciled.structuredContent).toMatchObject({
normalized_result: {
summary: { runtime_scripts: 1, matched: expect.any(Number) },
source_map_authority: { used_for_primary_matching: false },
},
});
expect(inspected.isError).not.toBe(true);
expect(inspected.structuredContent).toMatchObject({
normalized_result: {
target: { file_path: expect.stringMatching(/[\\/]index\.html$/u) },
scripts: {
items: [
expect.objectContaining({
frame_id: "frame-main",
file_path: expect.stringMatching(/[\\/]app\.js$/u),
cdp_hash: null,
length: null,
is_module: null,
}),
],
},
workers: [
expect.objectContaining({
target_id: "electron-worker",
opener_target_id: "electron-page",
file_path: expect.stringMatching(/[\\/]worker\.js$/u),
}),
],
},
});
}, 20_000);
it("runs active Electron scenarios with selected paths and inferred working directory", async () => {
const root = await createTestTempDirectory("rea-electron-active-mcp-");
temporary.push(root);
const applicationPath = join(root, "main.js");
await writeFile(applicationPath, "module.exports = {};\n");
const aliasedRoot = join(root, "..", `${basename(root)}-alias`);
// A junction needs no symlink privilege on Windows.
await symlink(
root,
aliasedRoot,
process.platform === "win32" ? "junction" : "dir",
);
temporary.push(aliasedRoot);
const workingDirectory = await createTestTempDirectory(
"rea-electron-working-directory-",
);
temporary.push(workingDirectory);
const capturedInputs: unknown[] = [];
const activeResult =
createElectronActiveObservationFixtureResult(applicationPath);
const capturedEvent = activeResult.timeline.events[0];
if (capturedEvent === undefined) throw new Error("IPC fixture was empty");
const gaps = ["custom-apply", "listener-chain"].map((event, index) => ({
...capturedEvent,
sequence: index + 2,
kind: "ipc-observation-gap" as const,
event,
channel: event === "custom-apply" ? "custom" : "opaque-chain",
direction: null,
receiver: null,
target: "ipcMain.on",
argument_shapes: [],
result_shape: null,
}));
activeResult.timeline.events.push(...gaps);
activeResult.timeline.observed = 3;
activeResult.retention.retained = 3;
activeResult.retention.estimated_retained_bytes = 1_536;
activeResult.retention.observed_runtime = 2;
activeResult.coverage.observed_event_families.push("ipc-observation-gap");
const provider: ElectronActiveObservationPort = {
identity: () => ({
id: "test-electron-active",
name: "Test Electron active provider",
version: "1",
}),
capture: async (input) => {
capturedInputs.push(input);
return { ok: true, value: activeResult };
},
};
const session = createTestBinarySession(() => ({
execute: () => Promise.resolve(observed(null)),
close: () => Promise.resolve(resultOk(null)),
}));
const server = createServer(
{ kind: "session", session },
{
electronActiveObservation: provider,
availabilityPolicy: () => ({
processCaptureEnabled: false,
investigationInputRoots: 0,
}),
},
);
const client = new Client({ name: "electron-active-mcp-test", version: "1" });
const [clientTransport, serverTransport] =
InMemoryTransport.createLinkedPair();
resources.push(client, server, session);
await server.connect(serverTransport);
await client.connect(clientTransport);
const advertised = (await client.listTools()).tools.find(
({ name }) => name === "capture_electron_scenario",
);
if (advertised?.outputSchema === undefined)
throw new Error("Active Electron output schema was not advertised");
const validateOutput = new Ajv2020({
strict: false,
validateFormats: false,
}).compile(z.record(z.string(), z.unknown()).parse(advertised.outputSchema));
const captured = await client.callTool({
name: "capture_electron_scenario",
arguments: {
executable_path: process.execPath,
application_path: join(aliasedRoot, "main.js"),
application_root: workingDirectory,
args: ["--token", "super-secret"],
actions: [
{ step_id: "submit", kind: "click", selector: "#submit-secret" },
],
},
});
expect(captured.isError, JSON.stringify(captured)).not.toBe(true);
expect(
validateOutput(captured.structuredContent),
JSON.stringify(validateOutput.errors),
).toBe(true);
expect(captured.structuredContent).toMatchObject({
normalized_result: {
application: { process_ownership: "provider-owned" },
ipc: { events: [{ channel: "readiness:echo" }] },
timeline: {
events: [
{ kind: "main-handler-invocation" },
{ kind: "ipc-observation-gap", channel: "custom" },
{ kind: "ipc-observation-gap", channel: "opaque-chain" },
],
},
coverage: {
status: "partial_attach",
pre_capture_activity: "unavailable",
},
},
});
expect(capturedInputs).toHaveLength(1);
expect(capturedInputs[0]).toMatchObject({
application_path: applicationPath,
application_root: workingDirectory,
});
const inferredRoot = await client.callTool({
name: "capture_electron_scenario",
arguments: {
executable_path: process.execPath,
application_path: applicationPath,
actions: [],
},
});
expect(inferredRoot.isError, JSON.stringify(inferredRoot)).not.toBe(true);
expect(capturedInputs[1]).toMatchObject({
application_path: applicationPath,
application_root: root,
});
expect(JSON.stringify(captured.structuredContent)).toContain("submit-secret");
expect(JSON.stringify(captured.structuredContent)).toContain("super-secret");
expect(captured.structuredContent).toMatchObject({
predicate_type: "rea.electron-active-scenario",
operation: "capture_electron_scenario",
parameters: {
args: ["--token", "super-secret"],
actions: [
{
step_id: "submit",
kind: "click",
selector: "#submit-secret",
window_index: 0,
},
],
},
});
}, 20_000);
it("exposes the target-free static JavaScript application workflow", async () => {
const root = await createTestTempDirectory("rea-electron-static-mcp-");
temporary.push(root);
await writeElectronBoundaryFixture(root);
const session = createTestBinarySession(() => ({
execute: () => Promise.resolve(observed(null)),
close: () => Promise.resolve(resultOk(null)),
}));
const server = createServer(
{ kind: "session", session },
{
availabilityPolicy: () => ({
processCaptureEnabled: false,
investigationInputRoots: 1,
}),
},
);
const client = new Client({
name: "electron-static-mcp-test",
version: "1",
});
const [clientTransport, serverTransport] =
InMemoryTransport.createLinkedPair();
resources.push(client, server, session);
await server.connect(serverTransport);
await client.connect(clientTransport);
const progress: number[] = [];
const analyzed = await client.callTool(
{
name: "analyze_javascript_application",
arguments: { input_path: root },
},
{ onprogress: (update) => progress.push(update.progress) },
);
expect(analyzed.isError).not.toBe(true);
expect(progress[0]).toBe(1);
expect(progress.at(-1)).toBeGreaterThan(1);
progress.forEach((value, index) => {
expect(value).toBeGreaterThan(progress[index - 1] ?? 0);
});
const projected = z
.object({
evidence_id: z.string(),
normalized_result: z.object({
input_path: z.string(),
summary: z.object({
browser_windows: z.number(),
context_bridge_apis: z.number(),
ipc: z.object({ paired_renderer_transmissions: z.number() }),
}),
graph: z.object({
graph_id: z.string(),
nodes: z.array(z.unknown()),
edges: z.array(z.unknown()),
}),
semantic_graph: z.object({
graph_id: z.string(),
nodes: z.array(z.unknown()),
relations: z.array(z.unknown()),
}),
}),
operation: z.literal("analyze_javascript_application"),
parameters: z.object({ format: z.string() }),
})
.parse(analyzed.structuredContent);
expect(projected.normalized_result).toMatchObject({
input_path: expect.stringMatching(/rea-electron-static-mcp-/u),
summary: {
browser_windows: 3,
context_bridge_apis: 2,
ipc: { paired_renderer_transmissions: 4 },
},
});
expect(projected.normalized_result.graph.nodes.length).toBeGreaterThan(0);
expect(
projected.normalized_result.semantic_graph.nodes.length,
).toBeGreaterThan(0);
expect(projected.operation).toBe("analyze_javascript_application");
const direct = await analyzeJavaScriptApplication({ input_path: root });
if (!direct.ok) throw direct.error;
expect(projected.evidence_id).toBe(direct.value.evidence_id);
});
const evidenceFor = (value: unknown) => {
return parseEvidence(value);
};
it("reports every unresolved Electron path without starting a provider", async () => {
const root = await createTestTempDirectory("rea-electron-invalid-paths-");
const capture = vi.fn();
const session = createTestBinarySession(() => {
throw new Error("must not start");
});
const server = createServer(
{ kind: "session", session },
{
electronActiveObservation: {
capture,
identity: () => ({ id: "test", name: "test", version: "1" }),
},
},
);
const client = new Client({ name: "electron-path-test", version: "1" });
const [clientTransport, serverTransport] =
InMemoryTransport.createLinkedPair();
resources.push(client, server, session);
await server.connect(serverTransport);
await client.connect(clientTransport);
const fields = ["executable_path", "application_path", "application_root"];
const result = await client.callTool({
name: "capture_electron_scenario",
arguments: {
...Object.fromEntries(fields.map((field) => [field, join(root, field)])),
actions: [],
},
});
expect(result.isError).toBe(true);
expect(parseMcpToolError(result)).toMatchObject({
error: {
code: "invalid_request",
details: {
issues: fields.map((field) => ({
path: [field],
reason: "invalid_value",
message: expect.stringContaining("ENOENT"),
})),
},
},
});
expect(capture).not.toHaveBeenCalled();
await client.ping();
});