Implement AI Signal Validation Integration

[Mathews-25]

Context:
To maintain signal quality and prevent spam, we integrate AI validation (xAI) to analyze signal rationale, detect patterns, and score quality.

Problem:
Build a service that sends signal data to xAI API for quality assessment and uses the score to filter low-quality signals.

What Done Looks Like:

• xAI API integration
• Signal quality scoring (0-100)
• Automatic rejection of low-quality signals (<30 score)
• Async processing to avoid blocking submission
• Fallback when xAI unavailable

Folder Structure:

src/
├── ai-validation/
│   ├── ai-validation.service.ts
│   ├── ai-validation.module.ts
│   ├── dto/
│   │   ├── validation-request.dto.ts
│   │   └── validation-response.dto.ts
│   └── interfaces/
│       └── xai-client.interface.ts

Implementation Guidelines:

• Use xAI Grok API for analysis
• Send signal rationale + asset pair + action
• Request quality score and fraud detection
• Process asynchronously with Bull queue
• Store validation score in signal entity
• Reject signals with score < 30
• Handle API timeouts gracefully (default to manual review)

Validation Criteria:

• Rationale coherence and specificity
• Market data references
• Realistic price targets
• Spam/manipulation detection

Edge Cases:

• xAI API timeout
• Rate limit exceeded
• Invalid API response

Validation:

• High-quality signal passes (score >70)
• Low-quality signal rejected (score <30)
• Async processing doesn't block submission
• Fallback works when API unavailable