Skip captcha for visitors with a valid idb_* cookie

[konard]

Summary

• Add hasIdbCookie() helper in Home.tsx that checks document.cookie for any idb_* prefix.
• When an idb_* cookie is present, the SmartCaptcha widget is neither shown nor loaded, and the captcha token is not required on form submit.
• In telegram-notify.php, captcha verification is skipped when any idb_* cookie is present in $_COOKIE.
• Updated tests/cta-captcha.test.mjs with a new test that validates the bypass behaviour and updated the existing test to match the new effect guard condition.

How to reproduce the original issue

With VITE_SMARTCAPTCHA_CLIENT_KEY set, the captcha widget always appeared for every visitor once they started typing their contact details — even returning users who already had Yandex identity (idb_*) cookies.

Test plan

• npm test — all 7 tests pass
• Manual: set an idb_test=1 cookie in the browser, open the CTA form, start typing in the contact field → captcha widget should not appear
• Manual: without any idb_* cookie → captcha widget should appear as before

Fixes #124

[konard]

Working session summary

Done. Here's what was implemented:

Frontend (src/pages/Home.tsx):

• Added hasIdbCookie() — checks document.cookie for any idb_* prefix
• Added idbCookieFound state, initialized once on mount
• The effect that loads/renders the SmartCaptcha widget early-returns when idbCookieFound is true
• The submit handler skips the captcha token requirement when idbCookieFound
• The captcha container is not rendered when idbCookieFound

Backend (public/telegram-notify.php):

• Added $hasIdbCookie check via preg_grep('/^idb_/', array_keys($_COOKIE))
• Captcha verification is skipped when $hasIdbCookie is true

Tests (tests/cta-captcha.test.mjs):

• Updated existing test to match the new effect guard condition
• Added new test verifying the hasIdbCookie helper and the bypass logic

All 7 tests pass. PR: #125

---

This summary was automatically extracted from the AI working session output.

[konard]

🤖 Solution Draft Log

This log file contains the complete execution trace of the AI solution draft process.

💰 Cost: $0.827789

📊 Context and tokens usage:

Claude Sonnet 4.6:

• peak request: 6.5K / 1M (1%) input tokens, 8.9K / 64K (14%) output tokens

Total: (42 new + 32.8K cache writes + 1.3M cache reads) input tokens, 8.9K output tokens, $0.650244 cost

Claude Haiku 4.5:

• (894 new + 67.6K cache writes + 642.6K cache reads) input tokens, 5.6K / 64K (9%) output tokens

Total: (894 new + 67.6K cache writes + 642.6K cache reads) input tokens, 5.6K output tokens, $0.177545 cost

🤖 Models used:

• Tool: Anthropic Claude Code
• Requested: sonnet
• Main model: Claude Sonnet 4.6 (claude-sonnet-4-6)
• Additional models:
  • Claude Haiku 4.5 (claude-haiku-4-5-20251001)

📎 Log file uploaded as Gist (1818KB)

• View complete solution draft log

---

Now working session is ended, feel free to review and add any feedback on the solution draft.

[konard]

✅ Ready to merge

This pull request is now ready to be merged:

• No CI/CD checks are configured for this repository
• No merge conflicts
• No pending changes

---

Monitored by hive-mind with --auto-restart-until-mergeable flag