@apmantza/greedysearch-pi 1.0.6 → 1.0.7

package/extractors/consent.mjs

@@ -1,29 +1,76 @@
-// consent.mjs — auto-dismiss common cookie/consent banners
-// Call dismissConsent(tab, cdpFn) after navigating to any page.
-
-const CONSENT_JS = `
-(function() {
-  // Google consent page (consent.google.com)
-  var g = document.querySelector('#L2AGLb, button[jsname="b3VHJd"], .tHlp8d');
-  if (g) { g.click(); return 'google'; }
-
-  // OneTrust (used by many sites including Stack Overflow)
-  var ot = document.querySelector('#onetrust-accept-btn-handler, .onetrust-accept-btn-handler');
-  if (ot) { ot.click(); return 'onetrust'; }
-
-  // Generic "accept all" / "agree" buttons
-  var btns = Array.from(document.querySelectorAll('button, a[role=button]'));
-  var accept = btns.find(b => /^(accept all|accept cookies|agree|i agree|got it|allow all|allow cookies)$/i.test(b.innerText?.trim()));
-  if (accept) { accept.click(); return 'generic:' + accept.innerText.trim(); }
-
-  return null;
-})()
-`;
-
-export async function dismissConsent(tab, cdp) {
-  const result = await cdp(['eval', tab, CONSENT_JS]).catch(() => null);
-  if (result && result !== 'null') {
-    // Give page a moment to react after dismissal
-    await new Promise(r => setTimeout(r, 1500));
-  }
-}
+// consent.mjs — auto-dismiss common cookie/consent banners and human-verification pages
+// Call dismissConsent(tab, cdpFn) after navigating to any page.
+
+const CONSENT_JS = `
+(function() {
+  // Google consent page (consent.google.com)
+  var g = document.querySelector('#L2AGLb, button[jsname="b3VHJd"], .tHlp8d');
+  if (g) { g.click(); return 'google'; }
+
+  // OneTrust (used by many sites including Stack Overflow)
+  var ot = document.querySelector('#onetrust-accept-btn-handler, .onetrust-accept-btn-handler');
+  if (ot) { ot.click(); return 'onetrust'; }
+
+  // Generic "accept all" / "agree" buttons
+  var btns = Array.from(document.querySelectorAll('button, a[role=button]'));
+  var accept = btns.find(b => /^(accept all|accept cookies|agree|i agree|got it|allow all|allow cookies)$/i.test(b.innerText?.trim()));
+  if (accept) { accept.click(); return 'generic:' + accept.innerText.trim(); }
+
+  return null;
+})()
+`;
+
+// Detect Google's "verify you're human" / unusual traffic page
+const VERIFY_DETECT_JS = `
+(function() {
+  var url = document.location.href;
+  if (url.includes('/sorry/') || url.includes('sorry.google')) return 'sorry-page';
+
+  // Simple click-through verify button (not image CAPTCHA)
+  var btns = Array.from(document.querySelectorAll('button, input[type=submit], a[role=button]'));
+  var verify = btns.find(b => /verify|human|not a robot|continue/i.test(b.innerText?.trim() || b.value || ''));
+  if (verify && !document.querySelector('iframe[src*="recaptcha"]')) {
+    verify.click();
+    return 'clicked-verify:' + (verify.innerText?.trim() || verify.value);
+  }
+
+  // Unchecked reCAPTCHA / Turnstile checkbox (no image challenge)
+  var checkbox = document.querySelector('.recaptcha-checkbox-unchecked, input[type=checkbox][id*="recaptcha"], #cf-stage input[type=checkbox]');
+  if (checkbox) { checkbox.click(); return 'clicked-checkbox'; }
+
+  return null;
+})()
+`;
+
+export async function dismissConsent(tab, cdp) {
+  const result = await cdp(['eval', tab, CONSENT_JS]).catch(() => null);
+  if (result && result !== 'null') {
+    await new Promise(r => setTimeout(r, 1500));
+  }
+}
+
+// Returns 'clear' | 'clicked' | 'needs-human'
+export async function handleVerification(tab, cdp, waitMs = 60000) {
+  const result = await cdp(['eval', tab, VERIFY_DETECT_JS]).catch(() => null);
+
+  if (!result || result === 'null') return 'clear';
+
+  if (result === 'sorry-page') {
+    // Hard CAPTCHA page — wait for user to solve it manually
+    process.stderr.write(`[greedysearch] Google verification required — please solve it in the browser window (waiting up to ${waitMs / 1000}s)...\n`);
+    const deadline = Date.now() + waitMs;
+    while (Date.now() < deadline) {
+      await new Promise(r => setTimeout(r, 2000));
+      const url = await cdp(['eval', tab, 'document.location.href']).catch(() => '');
+      if (!url.includes('/sorry/')) return 'cleared-by-user';
+    }
+    return 'needs-human';
+  }
+
+  if (result.startsWith('clicked-')) {
+    await new Promise(r => setTimeout(r, 2000));
+    return 'clicked';
+  }
+
+  return 'clear';
+}

package/extractors/google-ai.mjs

@@ -13,10 +13,10 @@ import { spawn } from 'child_process';
 import { tmpdir, homedir } from 'os';
 import { join, dirname } from 'path';
 import { fileURLToPath } from 'url';
-import { dismissConsent } from './consent.mjs';
+import { dismissConsent, handleVerification } from './consent.mjs';
 const __dir = dirname(fileURLToPath(import.meta.url));
 
-const CDP = join(dirname(fileURLToPath(import.meta.url)), '..', 'cdp.mjs');
+const CDP = join(homedir(), '.claude', 'skills', 'chrome-cdp', 'scripts', 'cdp.mjs');
 const PAGES_CACHE = `${tmpdir().replace(/\\/g, '/')}/cdp-pages.json`;
 
 const STREAM_POLL_INTERVAL = 600;
@@ -135,6 +135,15 @@ async function main() {
       await new Promise(r => setTimeout(r, 1500));
     }
 
+    // Handle "verify you're human" — auto-click simple buttons, wait for user on hard CAPTCHA
+    const verifyResult = await handleVerification(tab, cdp, 60000);
+    if (verifyResult === 'needs-human') throw new Error('Google verification required — could not be completed automatically');
+    if (verifyResult === 'clicked' || verifyResult === 'cleared-by-user') {
+      // Re-navigate to the search URL after verification
+      await cdp(['nav', tab, url], 35000);
+      await new Promise(r => setTimeout(r, 1500));
+    }
+
     await waitForStreamComplete(tab);
 
     const { answer, sources } = await extractAnswer(tab);

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@apmantza/greedysearch-pi",
-  "version": "1.0.6",
+  "version": "1.0.7",
   "description": "Pi extension: search Perplexity, Bing Copilot, and Google AI simultaneously — synthesized AI answers, not just links",
   "type": "module",
   "keywords": [

package/search.mjs

@@ -27,8 +27,6 @@ import { tmpdir } from 'os';
 import http from 'http';
 
 const __dir = dirname(fileURLToPath(import.meta.url));
-// Pi installs chrome-cdp-skill to ~/.pi/agent/git/...; fall back to Claude Code path
-// Always use the bundled Windows-compatible cdp.mjs
 const CDP = join(__dir, 'cdp.mjs');
 const PAGES_CACHE = `${tmpdir().replace(/\\/g, '/')}/cdp-pages.json`;
 
@@ -184,26 +182,26 @@ function writeOutput(data, outFile) {
   }
 }
 
-const GREEDY_PORT = 9223;
-
-function isGreedySearchChromeUp() {
+function probePort9223(timeoutMs = 3000) {
   return new Promise(resolve => {
-    const req = http.get(`http://localhost:${GREEDY_PORT}/json/version`, res => {
-      resolve(res.statusCode === 200);
+    const req = http.get('http://localhost:9223/json/version', res => {
       res.resume();
+      resolve(res.statusCode === 200);
     });
     req.on('error', () => resolve(false));
-    req.setTimeout(1500, () => { req.destroy(); resolve(false); });
+    req.setTimeout(timeoutMs, () => { req.destroy(); resolve(false); });
   });
 }
 
 async function ensureChrome() {
-  if (await isGreedySearchChromeUp()) return;
-  process.stderr.write('GreedySearch Chrome not running — auto-launching...\n');
-  await new Promise((resolve, reject) => {
-    const proc = spawn('node', [join(__dir, 'launch.mjs')], { stdio: ['ignore', process.stderr, process.stderr] });
-    proc.on('close', code => code === 0 ? resolve() : reject(new Error('launch.mjs failed')));
-  });
+  const ready = await probePort9223();
+  if (!ready) {
+    process.stderr.write('GreedySearch Chrome not running on port 9223 — auto-launching...\n');
+    await new Promise((resolve, reject) => {
+      const proc = spawn('node', [join(__dir, 'launch.mjs')], { stdio: ['ignore', process.stderr, process.stderr] });
+      proc.on('close', code => code === 0 ? resolve() : reject(new Error('launch.mjs failed')));
+    });
+  }
 }
 
 async function main() {
@@ -240,10 +238,10 @@ async function main() {
   if (engine === 'all') {
     await cdp(['list']); // refresh pages cache
 
-    // Assign tabs: reuse existing engine tabs from cache, open new ones only where needed.
-    // Track opened tabs separately so we only close what we created.
+    // Assign tabs: reuse existing engine tabs from cache, open new ones where needed.
+    // Engine tabs are never closed — keeping them alive preserves session cookies and
+    // reduces the chance of verification challenges on subsequent searches.
     const tabs = [];
-    const openedTabs = [];
     let blankReused = false;
 
     for (const e of ALL_ENGINES) {
@@ -253,13 +251,11 @@ async function main() {
       } else if (!blankReused) {
         const tab = await getOrReuseBlankTab();
         tabs.push(tab);
-        openedTabs.push(tab);
         blankReused = true;
       } else {
         await new Promise(r => setTimeout(r, 500));
         const tab = await openNewTab();
         tabs.push(tab);
-        openedTabs.push(tab);
       }
     }
 
@@ -270,9 +266,6 @@ async function main() {
       )
     );
 
-    // Close only tabs we opened (not pre-existing ones)
-    await Promise.allSettled(openedTabs.map(closeTab));
-
     const out = {};
     for (let i = 0; i < results.length; i++) {
       const r = results[i];