cannaiq/backend/archive/check-leaks.ts

import puppeteer from 'puppeteer-extra';
import StealthPlugin from 'puppeteer-extra-plugin-stealth';
import { Pool } from 'pg';

puppeteer.use(StealthPlugin());

const pool = new Pool({
  connectionString: 'postgresql://sail:password@localhost:5432/dutchie_menus'
});

async function check() {
  let browser;

  try {
    const proxyResult = await pool.query(`SELECT host, port, protocol FROM proxies ORDER BY RANDOM() LIMIT 1`);
    const proxy = proxyResult.rows[0];

    browser = await puppeteer.launch({
      headless: true,
      args: ['--no-sandbox', '--disable-setuid-sandbox', `--proxy-server=${proxy.protocol}://${proxy.host}:${proxy.port}`]
    });

    const page = await browser.newPage();
    await page.setUserAgent('Mozilla/5.0 (compatible; Googlebot/2.1; +http://www.google.com/bot.html)');

    console.log('🔍 CHECKING FOR DATA LEAKS\n');

    await page.goto('https://curaleaf.com/stores/curaleaf-dispensary-phoenix-airport/brands', {
      waitUntil: 'networkidle2',
      timeout: 60000
    });

    await page.waitForTimeout(5000);

    // Check what browser exposes
    const browserData = await page.evaluate(() => ({
      // Automation detection
      webdriver: navigator.webdriver,
      hasHeadlessUA: /headless/i.test(navigator.userAgent),

      // User agent
      userAgent: navigator.userAgent,

      // Chrome detection
      hasChrome: typeof (window as any).chrome !== 'undefined',
      chromeKeys: (window as any).chrome ? Object.keys((window as any).chrome) : [],

      // Permissions
      permissions: navigator.permissions ? 'exists' : 'missing',

      // Languages
      languages: navigator.languages,
      language: navigator.language,

      // Plugins
      pluginCount: navigator.plugins.length,

      // Platform
      platform: navigator.platform,

      // Screen
      screenWidth: screen.width,
      screenHeight: screen.height,

      // JavaScript working?
      jsWorking: true,

      // Page content
      title: document.title,
      bodyLength: document.body.innerHTML.length,
      hasReactRoot: document.getElementById('__next') !== null,
      scriptTags: document.querySelectorAll('script').length
    }));

    console.log('📋 BROWSER FINGERPRINT:');
    console.log('─'.repeat(60));
    console.log('navigator.webdriver:', browserData.webdriver, browserData.webdriver ? '❌ LEAKED!' : '✅');
    console.log('navigator.userAgent:', browserData.userAgent);
    console.log('Has "headless" in UA:', browserData.hasHeadlessUA, browserData.hasHeadlessUA ? '❌' : '✅');
    console.log('window.chrome exists:', browserData.hasChrome, browserData.hasChrome ? '✅' : '❌ SUSPICIOUS');
    console.log('Chrome keys:', browserData.chromeKeys.join(', '));
    console.log('Languages:', browserData.languages);
    console.log('Platform:', browserData.platform);
    console.log('Plugins:', browserData.pluginCount);

    console.log('\n📄 PAGE STATE:');
    console.log('─'.repeat(60));
    console.log('JavaScript executing:', browserData.jsWorking ? '✅ YES' : '❌ NO');
    console.log('Page title:', `"${browserData.title}"`);
    console.log('Body HTML size:', browserData.bodyLength, 'chars');
    console.log('React root exists:', browserData.hasReactRoot ? '✅' : '❌');
    console.log('Script tags:', browserData.scriptTags);

    if (browserData.bodyLength < 1000) {
      console.log('\n⚠️  PROBLEM: Body too small! JS likely failed to load/execute');
    }

    if (!browserData.title) {
      console.log('⚠️  PROBLEM: No page title! Page didn\'t render');
    }

  } catch (error: any) {
    console.error('❌', error.message);
  } finally {
    if (browser) await browser.close();
    await pool.end();
  }
}

check();