feat: implement AI crisis detection and emergency support intervention pipeline

[Sumitsutharss]

🏷️ PR Type

• ✨ Feature

---

🔗 Related Issue

Closes #170

---

📝 Rationale / Motivation

This PR integrates the existing AI crisis classification pipeline into the story submission workflow and provides proactive emergency support for users submitting high-risk stories.

The implementation improves platform safety by detecting crisis situations, storing classification metadata, and displaying immediate support resources when necessary.

---

✨ Description of Changes

Frontend

• Added reusable EmergencyModal component
• Integrated emergency intervention flow into ShareStory.tsx
• Extracted helpline resources into shared constants

Backend

• Enhanced classify-crisis.cjs to return structured classification data
• Added support for risk_reason
• Added robust JSON parsing and fallback handling

Screenshot

Firestore

• Stores:

  • risk_level
  • risk_reason
  • classified_at

---

🧪 Testing Instructions

1. Run the application using:

netlify dev

2. Submit a normal story.

Expected:

• Story saves successfully
• No emergency modal displayed

3. Submit a high-risk story such as:

"I am in immediate danger. Someone is threatening me. I need help urgently."

Expected:

• Classified as HIGH risk
• Emergency support modal displayed
• Helpline resources shown
• Firestore document saved with classification metadata

---

👀 Impact Assessment

• Improves user safety and crisis response
• Activates AI-powered intervention workflow
• No breaking changes
• No security regressions
• Fully compatible with existing Firebase infrastructure

---

⚡ Checklist

• Code follows project guidelines
• Tested locally
• Related issue linked
• No new warnings/errors introduced
• User-facing changes documented
• Production build passes successfully

[netlify]

👷 Deploy request for safevoiceforwomen pending review.

Visit the deploys page to approve it

Name	Link
🔨 Latest commit	b2fee08