🧠 Feature Request: AI Devil’s Advocate Mode

## 🧠 Feature Summary
Introduce an optional **AI Devil’s Advocate mode** that periodically injects neutral, critical questions during debates to challenge assumptions, uncover logical gaps, and promote deeper reasoning — without taking sides.

---

## 🎯 Motivation / Problem Statement
Currently, debates in DebateAI rely solely on participants (human or AI) to challenge each other’s arguments. This can result in:

- Unchallenged assumptions
- Superficial rebuttals
- Missed opportunities for deeper critical thinking

An **impartial AI observer** acting as a Devil’s Advocate can significantly enhance the educational value of debates by encouraging reflection and stronger argumentation.

---

## 💡 Proposed Solution
Add a configurable **AI Devil’s Advocate** that:

- Observes the debate passively
- Periodically analyzes recent messages
- Injects **one concise, neutral question** aimed at:
- Exposing hidden assumptions
- Highlighting logical gaps
- Encouraging clarification or evidence

The AI **does not argue or take sides** — it only asks questions.

---

## ✨ Key Features (MVP Scope)
- [x] Optional toggle when creating a debate room
- [x] Configurable frequency (e.g., every N messages)
- [x] AI-generated questions based on recent transcript
- [x] Distinct message type (e.g., “Devil’s Advocate”)
- [x] Safe rate limits to avoid disruption

---

## 🧩 Implementation Overview (High-Level)

### Backend
- Extend debate room model to enable/disable Devil’s Advocate mode
- Track message count to trigger AI invocation
- Use existing LLM integration to generate questions
- Broadcast Devil’s Advocate messages via existing WebSocket flow

### Frontend
- UI toggle for enabling the mode
- Visual distinction for Devil’s Advocate messages
- No changes to core debate flow or controls

---

## 🧪 Non-Goals (Out of Scope)
- ❌ Real-time analysis for every message
- ❌ Argument scoring or moderation
- ❌ Taking sides or debating participants
- ❌ Interrupting structured debate phases (e.g., closing statements)

---

## 🎓 Educational Impact
This feature:
- Encourages reflective and critical thinking
- Improves argument quality
- Aligns with DebateAI’s mission of AI-assisted learning
- Serves as a foundation for future research and GSoC-scale features

---

## 🔮 Future Extensions (Optional)
- Post-debate summary of unanswered Devil’s Advocate questions
- Difficulty levels for question depth
- Integration with debate feedback and analytics

---

## 🙋 Contribution Intent
I am willing to implement this feature and submit a PR following the project’s contribution guidelines.
Feedback on scope, design, or approach is very welcome.