bug: unwanted navigation caused by cookie consent logic

Author: mukulboro

src/core/browser-engine.js

@@ -208,9 +208,24 @@ export class BrowserEngine {
           if (!text || text.length > 50) continue; // Skip empty or too long text (unlikely to be a simple button)
 
           const lowerText = text.toLowerCase();
-          if (acceptButtonTexts.some(t => lowerText.includes(t.toLowerCase()))) {
+          const isMatch = acceptButtonTexts.some(t => lowerText.includes(t.toLowerCase()));
+
+          if (isMatch) {
             try {
+              // Safety: If it's a link with an href, it might navigate.
+              // We prefer clicking buttons or elements that don't look like external links.
+              const isLinkWithHref = btn.tagName === 'A' && btn.getAttribute('href') && !btn.getAttribute('href').startsWith('#');
+              const isStrongAccept = /accept all|allow all|agree to all|accept and continue/i.test(lowerText);
+
+              // If it's a link that navigates, only click if it's a VERY strong "Accept All" signal.
+              // Otherwise, we might be clicking a "Cookie Policy" link by mistake.
+              if (isLinkWithHref && !isStrongAccept) {
+                continue;
+              }
+
               btn.click();
+              // One click is often enough for most banners; let's break if we found a strong match
+              if (isStrongAccept) break;
             } catch (e) { }
           }
         }

src/core/mirror-cloner.js

@@ -570,9 +570,39 @@ export class MirrorCloner {
       timeout: this.options.timeout,
     });
 
+    // Capture the baseline URL after potentially initial redirects but BEFORE cookie clicking
+    const baselineUrl = page.url();
+
     // Handle cookie consent banners before further processing
     await this.browserEngine.handleCookieConsent(page);
 
+    // Give a small window for any triggered navigation to start
+    await new Promise(r => setTimeout(r, 1000));
+
+    // Post-cookie click safety check: verify if we navigated away from the original intent
+    let currentUrl = page.url();
+    if (currentUrl !== baselineUrl) {
+      // If we navigated to what looks like a cookie settings/privacy page, try to go back
+      const isCookiePage = /cookie|consent|privacy|settings|policy/i.test(currentUrl);
+      const hostChanged = (new URL(currentUrl).hostname !== new URL(baselineUrl).hostname);
+
+      if (isCookiePage || hostChanged) {
+        if (this.options.debug) {
+          console.log(chalk.yellow(`    ⚠️  Cookie consent caused navigation to: ${currentUrl}. Attempting to backtracking to: ${baselineUrl}`));
+        }
+        try {
+          // Go back or re-navigate
+          await page.goBack({ waitUntil: 'domcontentloaded', timeout: 10000 }).catch(async () => {
+             await page.goto(baselineUrl, { waitUntil: 'domcontentloaded', timeout: 10000 });
+          });
+          // Ensure we are back
+          currentUrl = page.url();
+        } catch (e) {
+          if (this.options.debug) console.log(chalk.dim(`      Backtracking failed: ${e.message}`));
+        }
+      }
+    }
+
     await this.waitForRootReady(page);
     await this.scrollToBottomAndLoad(page);
     await this.waitForImagesSettled(page, 8000);