Fixed the sign-up metadata extraction failure.
Changed /mnt/d/vcode/codex reg/grokcli-2api-1.9.82/grok-build-auth/xconsole_client/client.py:287 to make the scraper tolerate newer Next.js page shapes:
- Parses all self.__next_f.push(...) payloads instead of one narrow regex.
- Builds router-state fallback from the actual signup_url redirect, so grok-com no longer falls back to cloud-console.
- Finds action IDs from inline HTML or JS chunks, including createServerReference / registerServerReference contexts.
- Gives better diagnostics when action extraction still fails.
Added offline regression coverage in /mnt/d/vcode/codex reg/grokcli-2api-1.9.82/grok-build-auth/tests/test_signup_scraper.py:13.
Verification: python grok-build-auth\tests\test_signup_scraper.py passes. It prints an existing requests dependency warning, but both tests pass. I could not live-verify against accounts.x.ai from this host because direct TCP timed out and Firecrawl had insufficient credits.
Built With
- ets
Log in or sign up for Devpost to join the conversation.