feat(agent): replace the runtime npx for browser agent chrome devtool mcp with pre-built bundle (#22213)

Co-authored-by: Gaurav Ghosh <gaghosh@google.com>
Co-authored-by: Gaurav <39389231+gsquared94@users.noreply.github.com>

packages/core/src/agents/browser/browser-tools-manifest.json

@@ -0,0 +1,22 @@
+{
+  "description": "Explicitly promoted tools from chrome-devtools-mcp for the gemini-cli browser agent.",
+  "targetVersion": "0.19.0",
+  "exclude": [
+    {
+      "name": "lighthouse",
+      "reason": "3.5 MB pre-built bundle — not needed for gemini-cli browser agent's core tasks."
+    },
+    {
+      "name": "performance",
+      "reason": "Depends on chrome-devtools-frontend TraceEngine (~800 KB) — not needed for core tasks."
+    },
+    {
+      "name": "screencast",
+      "reason": "Requires ffmpeg at runtime — not a common browser agent use case and adds external dependency."
+    },
+    {
+      "name": "extensions",
+      "reason": "Extension management not relevant for the gemini-cli browser agent's current scope."
+    }
+  ]
+}

packages/core/src/agents/browser/browserAgentFactory.test.ts

@@ -24,6 +24,7 @@ const mockBrowserManager = {
     { name: 'click', description: 'Click element' },
     { name: 'fill', description: 'Fill form field' },
     { name: 'navigate_page', description: 'Navigate to URL' },
+    { name: 'type_text', description: 'Type text into an element' },
     // Visual tools (from --experimental-vision)
     { name: 'click_at', description: 'Click at coordinates' },
   ]),
@@ -70,6 +71,7 @@ describe('browserAgentFactory', () => {
       { name: 'click', description: 'Click element' },
       { name: 'fill', description: 'Fill form field' },
       { name: 'navigate_page', description: 'Navigate to URL' },
+      { name: 'type_text', description: 'Type text into an element' },
       // Visual tools (from --experimental-vision)
       { name: 'click_at', description: 'Click at coordinates' },
     ]);
@@ -135,7 +137,7 @@ describe('browserAgentFactory', () => {
       );
 
       expect(definition.name).toBe(BROWSER_AGENT_NAME);
-      // 5 MCP tools + 1 type_text composite tool (no analyze_screenshot without visualModel)
+      // 6 MCP tools (no analyze_screenshot without visualModel)
       expect(definition.toolConfig?.tools).toHaveLength(6);
     });
 
@@ -228,7 +230,7 @@ describe('browserAgentFactory', () => {
         mockMessageBus,
       );
 
-      // 5 MCP tools + 1 type_text + 1 analyze_screenshot
+      // 6 MCP tools + 1 analyze_screenshot
       expect(definition.toolConfig?.tools).toHaveLength(7);
       const toolNames =
         definition.toolConfig?.tools
@@ -268,6 +270,7 @@ describe('browserAgentFactory', () => {
         { name: 'close_page', description: 'Close page' },
         { name: 'select_page', description: 'Select page' },
         { name: 'press_key', description: 'Press key' },
+        { name: 'type_text', description: 'Type text into an element' },
         { name: 'hover', description: 'Hover element' },
       ]);
 
@@ -291,7 +294,6 @@ describe('browserAgentFactory', () => {
       expect(toolNames).toContain('click');
       expect(toolNames).toContain('take_snapshot');
       expect(toolNames).toContain('press_key');
-      // Custom composite tool must also be present
       expect(toolNames).toContain('type_text');
       // Total: 9 MCP + 1 type_text (no analyze_screenshot without visualModel)
       expect(definition.toolConfig?.tools).toHaveLength(10);

packages/core/src/agents/browser/browserManager.test.ts

@@ -39,6 +39,7 @@ vi.mock('@modelcontextprotocol/sdk/client/stdio.js', () => ({
 vi.mock('../../utils/debugLogger.js', () => ({
   debugLogger: {
     log: vi.fn(),
+    warn: vi.fn(),
     error: vi.fn(),
   },
 }));
@@ -47,6 +48,20 @@ vi.mock('./automationOverlay.js', () => ({
   injectAutomationOverlay: vi.fn().mockResolvedValue(undefined),
 }));
 
+vi.mock('node:fs', async (importOriginal) => {
+  const actual = await importOriginal<typeof import('node:fs')>();
+  return {
+    ...actual,
+    existsSync: vi.fn((p: string) => {
+      if (p.endsWith('bundled/chrome-devtools-mcp.mjs')) {
+        return false; // Default
+      }
+      return actual.existsSync(p);
+    }),
+  };
+});
+
+import * as fs from 'node:fs';
 import { Client } from '@modelcontextprotocol/sdk/client/index.js';
 import { StdioClientTransport } from '@modelcontextprotocol/sdk/client/stdio.js';
 
@@ -96,6 +111,40 @@ describe('BrowserManager', () => {
     vi.restoreAllMocks();
   });
 
+  describe('MCP bundled path resolution', () => {
+    it('should use bundled path if it exists (handles bundled CLI)', async () => {
+      vi.mocked(fs.existsSync).mockReturnValue(true);
+      const manager = new BrowserManager(mockConfig);
+      await manager.ensureConnection();
+
+      expect(StdioClientTransport).toHaveBeenCalledWith(
+        expect.objectContaining({
+          command: 'node',
+          args: expect.arrayContaining([
+            expect.stringMatching(/bundled\/chrome-devtools-mcp\.mjs$/),
+          ]),
+        }),
+      );
+    });
+
+    it('should fall back to development path if bundled path does not exist', async () => {
+      vi.mocked(fs.existsSync).mockReturnValue(false);
+      const manager = new BrowserManager(mockConfig);
+      await manager.ensureConnection();
+
+      expect(StdioClientTransport).toHaveBeenCalledWith(
+        expect.objectContaining({
+          command: 'node',
+          args: expect.arrayContaining([
+            expect.stringMatching(
+              /(dist\/)?bundled\/chrome-devtools-mcp\.mjs$/,
+            ),
+          ]),
+        }),
+      );
+    });
+  });
+
   describe('getRawMcpClient', () => {
     it('should ensure connection and return raw MCP client', async () => {
       const manager = new BrowserManager(mockConfig);
@@ -222,10 +271,9 @@ describe('BrowserManager', () => {
       // Verify StdioClientTransport was created with correct args
       expect(StdioClientTransport).toHaveBeenCalledWith(
         expect.objectContaining({
-          command: process.platform === 'win32' ? 'npx.cmd' : 'npx',
+          command: 'node',
           args: expect.arrayContaining([
-            '-y',
-            expect.stringMatching(/chrome-devtools-mcp@/),
+            expect.stringMatching(/chrome-devtools-mcp\.mjs$/),
             '--experimental-vision',
           ]),
         }),
@@ -235,6 +283,7 @@ describe('BrowserManager', () => {
         ?.args as string[];
       expect(args).not.toContain('--isolated');
       expect(args).not.toContain('--autoConnect');
+      expect(args).not.toContain('-y');
       // Persistent mode should set the default --userDataDir under ~/.gemini
       expect(args).toContain('--userDataDir');
       const userDataDirIndex = args.indexOf('--userDataDir');
@@ -294,7 +343,7 @@ describe('BrowserManager', () => {
 
       expect(StdioClientTransport).toHaveBeenCalledWith(
         expect.objectContaining({
-          command: process.platform === 'win32' ? 'npx.cmd' : 'npx',
+          command: 'node',
           args: expect.arrayContaining(['--headless']),
         }),
       );
@@ -319,7 +368,7 @@ describe('BrowserManager', () => {
 
       expect(StdioClientTransport).toHaveBeenCalledWith(
         expect.objectContaining({
-          command: process.platform === 'win32' ? 'npx.cmd' : 'npx',
+          command: 'node',
           args: expect.arrayContaining(['--userDataDir', '/path/to/profile']),
         }),
       );

packages/core/src/agents/browser/browserManager.ts

@@ -25,10 +25,12 @@ import type { Config } from '../../config/config.js';
 import { Storage } from '../../config/storage.js';
 import { injectInputBlocker } from './inputBlocker.js';
 import * as path from 'node:path';
+import * as fs from 'node:fs';
+import { fileURLToPath } from 'node:url';
 import { injectAutomationOverlay } from './automationOverlay.js';
 
-// Pin chrome-devtools-mcp version for reproducibility.
-const CHROME_DEVTOOLS_MCP_VERSION = '0.17.1';
+const __filename = fileURLToPath(import.meta.url);
+const __dirname = path.dirname(__filename);
 
 // Default browser profile directory name within ~/.gemini/
 const BROWSER_PROFILE_DIR = 'cli-browser-profile';
@@ -279,7 +281,7 @@ export class BrowserManager {
       this.rawMcpClient = undefined;
     }
 
-    // Close transport (this terminates the npx process and browser)
+    // Close transport (this terminates the browser)
     if (this.mcpTransport) {
       try {
         await this.mcpTransport.close();
@@ -297,8 +299,7 @@ export class BrowserManager {
   /**
    * Connects to chrome-devtools-mcp which manages the browser process.
    *
-   * Spawns npx chrome-devtools-mcp with:
-   * - --isolated: Manages its own browser instance
+   * Spawns node with the bundled chrome-devtools-mcp.mjs.
    * - --experimental-vision: Enables visual tools (click_at, etc.)
    *
    * IMPORTANT: This does NOT use McpClientManager and does NOT register
@@ -323,11 +324,7 @@ export class BrowserManager {
     const browserConfig = this.config.getBrowserAgentConfig();
     const sessionMode = browserConfig.customConfig.sessionMode ?? 'persistent';
 
-    const mcpArgs = [
-      '-y',
-      `chrome-devtools-mcp@${CHROME_DEVTOOLS_MCP_VERSION}`,
-      '--experimental-vision',
-    ];
+    const mcpArgs = ['--experimental-vision'];
 
     // Session mode determines how the browser is managed:
     // - "isolated": Temp profile, cleaned up after session (--isolated)
@@ -373,15 +370,28 @@ export class BrowserManager {
     }
 
     debugLogger.log(
-      `Launching chrome-devtools-mcp (${sessionMode} mode) with args: ${mcpArgs.join(' ')}`,
+      `Launching bundled chrome-devtools-mcp (${sessionMode} mode) with args: ${mcpArgs.join(' ')}`,
     );
 
-    // Create stdio transport to npx chrome-devtools-mcp.
+    // Create stdio transport to the bundled chrome-devtools-mcp.
     // stderr is piped (not inherited) to prevent MCP server banners and
     // warnings from corrupting the UI in alternate buffer mode.
+    let bundleMcpPath = path.resolve(
+      __dirname,
+      'bundled/chrome-devtools-mcp.mjs',
+    );
+    if (!fs.existsSync(bundleMcpPath)) {
+      bundleMcpPath = path.resolve(
+        __dirname,
+        __dirname.includes(`${path.sep}dist${path.sep}`)
+          ? '../../../bundled/chrome-devtools-mcp.mjs'
+          : '../../../dist/bundled/chrome-devtools-mcp.mjs',
+      );
+    }
+
     this.mcpTransport = new StdioClientTransport({
-      command: process.platform === 'win32' ? 'npx.cmd' : 'npx',
-      args: mcpArgs,
+      command: 'node',
+      args: [bundleMcpPath, ...mcpArgs],
       stderr: 'pipe',
     });
 
@@ -492,8 +502,7 @@ export class BrowserManager {
         `Timed out connecting to Chrome: ${message}\n\n` +
           `Possible causes:\n` +
           `  1. Chrome is not installed or not in PATH\n` +
-          `  2. npx cannot download chrome-devtools-mcp (check network/proxy)\n` +
-          `  3. Chrome failed to start (try setting headless: true in settings.json)`,
+          `  2. Chrome failed to start (try setting headless: true in settings.json)`,
       );
     }
 

packages/core/src/agents/browser/mcpToolWrapper.test.ts

@@ -68,18 +68,19 @@ describe('mcpToolWrapper', () => {
       const tools = await createMcpDeclarativeTools(
         mockBrowserManager,
         mockMessageBus,
+        false,
       );
 
-      expect(tools).toHaveLength(3);
+      expect(tools).toHaveLength(2);
       expect(tools[0].name).toBe('take_snapshot');
       expect(tools[1].name).toBe('click');
-      expect(tools[2].name).toBe('type_text');
     });
 
     it('should return tools with correct description', async () => {
       const tools = await createMcpDeclarativeTools(
         mockBrowserManager,
         mockMessageBus,
+        false,
       );
 
       // Descriptions include augmented hints, so we check they contain the original
@@ -93,6 +94,7 @@ describe('mcpToolWrapper', () => {
       const tools = await createMcpDeclarativeTools(
         mockBrowserManager,
         mockMessageBus,
+        false,
       );
 
       const schema = tools[0].schema;
@@ -106,6 +108,7 @@ describe('mcpToolWrapper', () => {
       const tools = await createMcpDeclarativeTools(
         mockBrowserManager,
         mockMessageBus,
+        false,
       );
 
       const invocation = tools[0].build({ verbose: true });
@@ -118,6 +121,7 @@ describe('mcpToolWrapper', () => {
       const tools = await createMcpDeclarativeTools(
         mockBrowserManager,
         mockMessageBus,
+        false,
       );
 
       const invocation = tools[0].build({});
@@ -131,6 +135,7 @@ describe('mcpToolWrapper', () => {
       const tools = await createMcpDeclarativeTools(
         mockBrowserManager,
         mockMessageBus,
+        false,
       );
 
       const invocation = tools[1].build({ uid: 'elem-123' });
@@ -149,6 +154,7 @@ describe('mcpToolWrapper', () => {
       const tools = await createMcpDeclarativeTools(
         mockBrowserManager,
         mockMessageBus,
+        false,
       );
 
       const invocation = tools[0].build({ verbose: true });
@@ -167,6 +173,7 @@ describe('mcpToolWrapper', () => {
       const tools = await createMcpDeclarativeTools(
         mockBrowserManager,
         mockMessageBus,
+        false,
       );
 
       const invocation = tools[1].build({ uid: 'invalid' });
@@ -184,6 +191,7 @@ describe('mcpToolWrapper', () => {
       const tools = await createMcpDeclarativeTools(
         mockBrowserManager,
         mockMessageBus,
+        false,
       );
 
       const invocation = tools[0].build({});

packages/core/src/agents/browser/mcpToolWrapper.ts

@@ -175,144 +175,6 @@ class McpToolInvocation extends BaseToolInvocation<
   }
 }
 
-/**
- * Composite tool invocation that types a full string by calling press_key
- * for each character internally, avoiding N model round-trips.
- */
-class TypeTextInvocation extends BaseToolInvocation<
-  Record<string, unknown>,
-  ToolResult
-> {
-  constructor(
-    private readonly browserManager: BrowserManager,
-    private readonly text: string,
-    private readonly submitKey: string | undefined,
-    messageBus: MessageBus,
-  ) {
-    super({ text, submitKey }, messageBus, 'type_text', 'type_text');
-  }
-
-  getDescription(): string {
-    const preview = `"${this.text.substring(0, 50)}${this.text.length > 50 ? '...' : ''}"`;
-    return this.submitKey
-      ? `type_text: ${preview} + ${this.submitKey}`
-      : `type_text: ${preview}`;
-  }
-
-  protected override async getConfirmationDetails(
-    _abortSignal: AbortSignal,
-  ): Promise<ToolCallConfirmationDetails | false> {
-    if (!this.messageBus) {
-      return false;
-    }
-
-    return {
-      type: 'mcp',
-      title: `Confirm Tool: type_text`,
-      serverName: 'browser-agent',
-      toolName: 'type_text',
-      toolDisplayName: 'type_text',
-      onConfirm: async (outcome: ToolConfirmationOutcome) => {
-        await this.publishPolicyUpdate(outcome);
-      },
-    };
-  }
-
-  override getPolicyUpdateOptions(
-    _outcome: ToolConfirmationOutcome,
-  ): PolicyUpdateOptions | undefined {
-    return {
-      mcpName: 'browser-agent',
-    };
-  }
-
-  override async execute(signal: AbortSignal): Promise<ToolResult> {
-    try {
-      if (signal.aborted) {
-        return {
-          llmContent: 'Error: Operation cancelled before typing started.',
-          returnDisplay: 'Operation cancelled before typing started.',
-          error: { message: 'Operation cancelled' },
-        };
-      }
-
-      await this.typeCharByChar(signal);
-
-      // Optionally press a submit key (Enter, Tab, etc.) after typing
-      if (this.submitKey && !signal.aborted) {
-        const keyResult = await this.browserManager.callTool(
-          'press_key',
-          { key: this.submitKey },
-          signal,
-        );
-        if (keyResult.isError) {
-          const errText = this.extractErrorText(keyResult);
-          debugLogger.warn(
-            `type_text: submitKey("${this.submitKey}") failed: ${errText}`,
-          );
-        }
-      }
-
-      const summary = this.submitKey
-        ? `Successfully typed "${this.text}" and pressed ${this.submitKey}`
-        : `Successfully typed "${this.text}"`;
-
-      return {
-        llmContent: summary,
-        returnDisplay: summary,
-      };
-    } catch (error) {
-      const errorMsg = error instanceof Error ? error.message : String(error);
-
-      // Chrome connection errors are fatal
-      if (errorMsg.includes('Could not connect to Chrome')) {
-        throw error;
-      }
-
-      debugLogger.error(`type_text failed: ${errorMsg}`);
-      return {
-        llmContent: `Error: ${errorMsg}`,
-        returnDisplay: `Error: ${errorMsg}`,
-        error: { message: errorMsg },
-      };
-    }
-  }
-
-  /** Types each character via individual press_key MCP calls. */
-  private async typeCharByChar(signal: AbortSignal): Promise<void> {
-    const chars = [...this.text]; // Handle Unicode correctly
-    for (const char of chars) {
-      if (signal.aborted) return;
-
-      // Map special characters to key names
-      const key = char === ' ' ? 'Space' : char;
-      const result = await this.browserManager.callTool(
-        'press_key',
-        { key },
-        signal,
-      );
-
-      if (result.isError) {
-        debugLogger.warn(
-          `type_text: press_key("${key}") failed: ${this.extractErrorText(result)}`,
-        );
-      }
-    }
-  }
-
-  /** Extract error text from an MCP tool result. */
-  private extractErrorText(result: McpToolCallResult): string {
-    return (
-      result.content
-        ?.filter(
-          (c: { type: string; text?: string }) => c.type === 'text' && c.text,
-        )
-        .map((c: { type: string; text?: string }) => c.text)
-        .join('\n') || 'Unknown error'
-    );
-  }
-}
-
 /**
  * DeclarativeTool wrapper for an MCP tool.
  */
@@ -353,65 +215,6 @@ class McpDeclarativeTool extends DeclarativeTool<
   }
 }
 
-/**
- * DeclarativeTool for the custom type_text composite tool.
- */
-class TypeTextDeclarativeTool extends DeclarativeTool<
-  Record<string, unknown>,
-  ToolResult
-> {
-  constructor(
-    private readonly browserManager: BrowserManager,
-    messageBus: MessageBus,
-  ) {
-    super(
-      'type_text',
-      'type_text',
-      'Types a full text string into the currently focused element. ' +
-        'Much faster than calling press_key for each character individually. ' +
-        'Use this to enter text into form fields, search boxes, spreadsheet cells, or any focused input. ' +
-        'The element must already be focused (e.g., after a click). ' +
-        'Use submitKey to press a key after typing (e.g., submitKey="Enter" to submit a form or confirm a value, submitKey="Tab" to move to the next field).',
-      Kind.Other,
-      {
-        type: 'object',
-        properties: {
-          text: {
-            type: 'string',
-            description: 'The text to type into the focused element.',
-          },
-          submitKey: {
-            type: 'string',
-            description:
-              'Optional key to press after typing (e.g., "Enter", "Tab", "Escape"). ' +
-              'Useful for submitting form fields or moving to the next cell in a spreadsheet.',
-          },
-        },
-        required: ['text'],
-      },
-      messageBus,
-      /* isOutputMarkdown */ true,
-      /* canUpdateOutput */ false,
-    );
-  }
-
-  build(
-    params: Record<string, unknown>,
-  ): ToolInvocation<Record<string, unknown>, ToolResult> {
-    const submitKey =
-      // eslint-disable-next-line no-restricted-syntax
-      typeof params['submitKey'] === 'string' && params['submitKey']
-        ? params['submitKey']
-        : undefined;
-    return new TypeTextInvocation(
-      this.browserManager,
-      String(params['text'] ?? ''),
-      submitKey,
-      this.messageBus,
-    );
-  }
-}
-
 /**
  * Creates DeclarativeTool instances from dynamically discovered MCP tools,
  * plus custom composite tools (like type_text).
@@ -423,13 +226,14 @@ class TypeTextDeclarativeTool extends DeclarativeTool<
  *
  * @param browserManager The browser manager with isolated MCP client
  * @param messageBus Message bus for tool invocations
+ * @param shouldDisableInput Whether input should be disabled for this agent
  * @returns Array of DeclarativeTools that dispatch to the isolated MCP client
  */
 export async function createMcpDeclarativeTools(
   browserManager: BrowserManager,
   messageBus: MessageBus,
   shouldDisableInput: boolean = false,
-): Promise<Array<McpDeclarativeTool | TypeTextDeclarativeTool>> {
+): Promise<McpDeclarativeTool[]> {
   // Get dynamically discovered tools from the MCP server
   const mcpTools = await browserManager.getDiscoveredTools();
 
@@ -438,29 +242,25 @@ export async function createMcpDeclarativeTools(
       (shouldDisableInput ? ' (input blocker enabled)' : ''),
   );
 
-  const tools: Array<McpDeclarativeTool | TypeTextDeclarativeTool> =
-    mcpTools.map((mcpTool) => {
-      const schema = convertMcpToolToFunctionDeclaration(mcpTool);
-      // Augment description with uid-context hints
-      const augmentedDescription = augmentToolDescription(
-        mcpTool.name,
-        mcpTool.description ?? '',
-      );
-      return new McpDeclarativeTool(
-        browserManager,
-        mcpTool.name,
-        augmentedDescription,
-        schema.parametersJsonSchema,
-        messageBus,
-        shouldDisableInput,
-      );
-    });
-
-  // Add custom composite tools
-  tools.push(new TypeTextDeclarativeTool(browserManager, messageBus));
+  const tools: McpDeclarativeTool[] = mcpTools.map((mcpTool) => {
+    const schema = convertMcpToolToFunctionDeclaration(mcpTool);
+    // Augment description with uid-context hints
+    const augmentedDescription = augmentToolDescription(
+      mcpTool.name,
+      mcpTool.description ?? '',
+    );
+    return new McpDeclarativeTool(
+      browserManager,
+      mcpTool.name,
+      augmentedDescription,
+      schema.parametersJsonSchema,
+      messageBus,
+      shouldDisableInput,
+    );
+  });
 
   debugLogger.log(
-    `Total tools registered: ${tools.length} (${mcpTools.length} MCP + 1 custom)`,
+    `Total tools registered: ${tools.length} (${mcpTools.length} MCP)`,
   );
 
   return tools;