MCP Accessibility Scanner

Instructions

Implementation Reference

• src/tools/screenshot.ts:53-87 (handler)

  Executes the screenshot logic: determines file type and name, configures Playwright screenshot options, handles element/fullPage/viewport screenshots using locators, saves to file, adds code and result to response, and adds image if not full page.

  handle: async (tab, params, response) => {
    const fileType = params.type || 'png';
    const fileName = await tab.context.outputFile(params.filename ?? `page-${new Date().toISOString()}.${fileType}`);
    const options: playwright.PageScreenshotOptions = {
      type: fileType,
      quality: fileType === 'png' ? undefined : 90,
      scale: 'css',
      path: fileName,
      ...(params.fullPage !== undefined && { fullPage: params.fullPage })
    };
    const isElementScreenshot = params.element && params.ref;
  
    const screenshotTarget = isElementScreenshot ? params.element : (params.fullPage ? 'full page' : 'viewport');
    response.addCode(`// Screenshot ${screenshotTarget} and save it as ${fileName}`);
  
    // Only get snapshot when element screenshot is needed
    const locator = params.ref ? await tab.refLocator({ element: params.element || '', ref: params.ref }) : null;
  
    if (locator)
      response.addCode(`await page.${await generateLocator(locator)}.screenshot(${javascript.formatObject(options)});`);
    else
      response.addCode(`await page.screenshot(${javascript.formatObject(options)});`);
  
    const buffer = locator ? await locator.screenshot(options) : await tab.page.screenshot(options);
    response.addResult(`Took the ${screenshotTarget} screenshot and saved it as ${fileName}`);
  
    // https://github.com/microsoft/playwright-mcp/issues/817
    // Never return large images to LLM, saving them to the file system is enough.
    if (!params.fullPage) {
      response.addImage({
        contentType: fileType === 'png' ? 'image/png' : 'image/jpeg',
        data: buffer
      });
    }
  }

• src/tools/screenshot.ts:25-41 (schema)

  Zod schema validating input parameters for browser_take_screenshot: type (png/jpeg), filename, element description, ref, fullPage boolean, with refinements ensuring element/ref pair or neither, and no fullPage with elements.

  const screenshotSchema = z.object({
    type: z.enum(['png', 'jpeg']).default('png').describe('Image format for the screenshot. Default is png.'),
    filename: z.string().optional().describe('File name to save the screenshot to. Defaults to `page-{timestamp}.{png|jpeg}` if not specified.'),
    element: z.string().optional().describe('Human-readable element description used to obtain permission to screenshot the element. If not provided, the screenshot will be taken of viewport. If element is provided, ref must be provided too.'),
    ref: z.string().optional().describe('Exact target element reference from the page snapshot. If not provided, the screenshot will be taken of viewport. If ref is provided, element must be provided too.'),
    fullPage: z.boolean().optional().describe('When true, takes a screenshot of the full scrollable page, instead of the currently visible viewport. Cannot be used with element screenshots.'),
  }).refine(data => {
    return !!data.element === !!data.ref;
  }, {
    message: 'Both element and ref must be provided or neither.',
    path: ['ref', 'element']
  }).refine(data => {
    return !(data.fullPage && (data.element || data.ref));
  }, {
    message: 'fullPage cannot be used with element screenshots.',
    path: ['fullPage']
  });

• src/tools/screenshot.ts:43-92 (registration)

  Registers the browser_take_screenshot tool using defineTabTool, specifying name, title, description, input schema, and handler function. Exports it for inclusion in tools list.

  const screenshot = defineTabTool({
    capability: 'core',
    schema: {
      name: 'browser_take_screenshot',
      title: 'Take a screenshot',
      description: `Take a screenshot of the current page. You can't perform actions based on the screenshot, use browser_snapshot for actions.`,
      inputSchema: screenshotSchema,
      type: 'readOnly',
    },
  
    handle: async (tab, params, response) => {
      const fileType = params.type || 'png';
      const fileName = await tab.context.outputFile(params.filename ?? `page-${new Date().toISOString()}.${fileType}`);
      const options: playwright.PageScreenshotOptions = {
        type: fileType,
        quality: fileType === 'png' ? undefined : 90,
        scale: 'css',
        path: fileName,
        ...(params.fullPage !== undefined && { fullPage: params.fullPage })
      };
      const isElementScreenshot = params.element && params.ref;
  
      const screenshotTarget = isElementScreenshot ? params.element : (params.fullPage ? 'full page' : 'viewport');
      response.addCode(`// Screenshot ${screenshotTarget} and save it as ${fileName}`);
  
      // Only get snapshot when element screenshot is needed
      const locator = params.ref ? await tab.refLocator({ element: params.element || '', ref: params.ref }) : null;
  
      if (locator)
        response.addCode(`await page.${await generateLocator(locator)}.screenshot(${javascript.formatObject(options)});`);
      else
        response.addCode(`await page.screenshot(${javascript.formatObject(options)});`);
  
      const buffer = locator ? await locator.screenshot(options) : await tab.page.screenshot(options);
      response.addResult(`Took the ${screenshotTarget} screenshot and saved it as ${fileName}`);
  
      // https://github.com/microsoft/playwright-mcp/issues/817
      // Never return large images to LLM, saving them to the file system is enough.
      if (!params.fullPage) {
        response.addImage({
          contentType: fileType === 'png' ? 'image/png' : 'image/jpeg',
          data: buffer
        });
      }
    }
  });
  
  export default [
    screenshot,
  ];