Generalize browser automation hosting beyond the desktop app (#1878)

* feat(browser): generalize browser automation hosts

Rename the browser automation client capability to a host-neutral payload, advertise supported commands, and route tab-scoped commands by browserId affinity so future browser hosts can register without another protocol shape change.

* fix(browser): harden browser host registration

Reject unusable browser host capabilities, clean up pending requests when a host registration is replaced, and defer list_tabs affinity updates until every host in the aggregation succeeds.

* fix(browser): preserve host tab affinity across reconnects

Keep stranded tab ownership tied to the disconnected host so the same host can reclaim open tabs after reconnect, and tolerate future command names in browser_host capability payloads while retaining only known commands.

packages/protocol/src/messages.browser-automation.test.ts

@@ -1,5 +1,6 @@
 import { describe, expect, test } from "vitest";
 
+import { BROWSER_AUTOMATION_COMMAND_NAMES } from "./browser-automation/rpc-schemas.js";
 import { CLIENT_CAPS } from "./client-capabilities.js";
 import {
   MutableDaemonConfigPatchSchema,
@@ -12,7 +13,7 @@ import {
 describe("browser automation protocol integration", () => {
   const browserId = "11111111-1111-4111-8111-111111111111";
 
-  test("desktop automation capability parses in hello without narrowing old clients", () => {
+  test("browser host capability parses supported commands in hello", () => {
     expect(
       WSHelloMessageSchema.parse({
         type: "hello",
@@ -20,13 +21,86 @@ describe("browser automation protocol integration", () => {
         clientType: "mobile",
         protocolVersion: 1,
         capabilities: {
-          [CLIENT_CAPS.desktopBrowserAutomation]: true,
+          [CLIENT_CAPS.browserHost]: {
+            supportedCommands: [...BROWSER_AUTOMATION_COMMAND_NAMES],
+            hostKind: "desktop app",
+          },
         },
       }).capabilities,
     ).toMatchObject({
-      [CLIENT_CAPS.desktopBrowserAutomation]: true,
+      [CLIENT_CAPS.browserHost]: {
+        supportedCommands: [...BROWSER_AUTOMATION_COMMAND_NAMES],
+        hostKind: "desktop app",
+      },
     });
+  });
 
+  test("browser host capability requires at least one supported command", () => {
+    expect(() =>
+      WSHelloMessageSchema.parse({
+        type: "hello",
+        clientId: "client-1",
+        clientType: "mobile",
+        protocolVersion: 1,
+        capabilities: {
+          [CLIENT_CAPS.browserHost]: {
+            supportedCommands: [],
+            hostKind: "desktop app",
+          },
+        },
+      }),
+    ).toThrow();
+
+    expect(() =>
+      WSHelloMessageSchema.parse({
+        type: "hello",
+        clientId: "client-2",
+        clientType: "mobile",
+        protocolVersion: 1,
+        capabilities: {
+          [CLIENT_CAPS.browserHost]: {},
+        },
+      }),
+    ).toThrow();
+
+    expect(() =>
+      WSHelloMessageSchema.parse({
+        type: "hello",
+        clientId: "client-3",
+        clientType: "mobile",
+        protocolVersion: 1,
+        capabilities: {
+          [CLIENT_CAPS.browserHost]: {
+            supportedCommands: ["future_command"],
+          },
+        },
+      }),
+    ).toThrow();
+  });
+
+  test("browser host capability ignores unknown future commands when known commands remain", () => {
+    expect(
+      WSHelloMessageSchema.parse({
+        type: "hello",
+        clientId: "client-1",
+        clientType: "mobile",
+        protocolVersion: 1,
+        capabilities: {
+          [CLIENT_CAPS.browserHost]: {
+            supportedCommands: ["list_tabs", "future_command", "list_tabs"],
+            hostKind: "desktop app",
+          },
+        },
+      }).capabilities,
+    ).toMatchObject({
+      [CLIENT_CAPS.browserHost]: {
+        supportedCommands: ["list_tabs"],
+        hostKind: "desktop app",
+      },
+    });
+  });
+
+  test("hello remains valid when no browser host capability is advertised", () => {
     expect(
       WSHelloMessageSchema.parse({
         type: "hello",
@@ -37,7 +111,7 @@ describe("browser automation protocol integration", () => {
     ).toBeUndefined();
   });
 
-  test("daemon to desktop execute request is an outbound session message", () => {
+  test("daemon to browser host execute request is an outbound session message", () => {
     const parsed = SessionOutboundMessageSchema.parse({
       type: "browser.automation.execute.request",
       requestId: "req-1",
@@ -47,7 +121,7 @@ describe("browser automation protocol integration", () => {
     expect(parsed.type).toBe("browser.automation.execute.request");
   });
 
-  test("desktop to daemon execute response is an inbound session message", () => {
+  test("browser host to daemon execute response is an inbound session message", () => {
     const parsed = SessionInboundMessageSchema.parse({
       type: "browser.automation.execute.response",
       payload: {