Decision Intelligence API Endpoints

Page Outline

• Decision Intelligence API Documentation
  • Decision Making
    • 1. Make Decision
  • Model Training (Custom)
    • 2. Train Decision Model
  • Auditing
    • 3. List Decisions
  • Governance
    • 4. Create Guideline

---

Decision Intelligence API Documentation

Base URL: /api/decision-intelligence

Decision Making

1. Make Decision

Make a decision based on agent metadata using the ML engine and guideline refinement.

Endpoint: POST /api/decision-intelligence/decide Status Code: 200 OK

Request Body:

{
  "agent_metadata": [
    {
      "agent_id": "agent_1",
      "metadata": { "cpu": 0.8, "load": 50 }
    },
    {
      "agent_id": "agent_2",
      "metadata": { "cpu": 0.4, "load": 20 }
    }
  ],
  "decision_type": "classification",
  "model_name": "default"
}

Response:

{
  "decision": "scale_up",
  "confidence": 0.85,
  "reasoning": "High CPU load detected across agents."
}

Model Training (Custom)

2. Train Decision Model

Train a custom decision model using historical data.

Endpoint: POST /api/decision-intelligence/train Status Code: 201 Created

Request Body:

{
  "training_data": [...],
  "target": "action",
  "model_type": "random_forest"
}

Auditing

3. List Decisions

List past decisions for auditing purposes.

Endpoint: GET /api/decision-intelligence/decisions

Governance

4. Create Guideline

Define a new operational guideline or prompt for the decision engine.

Endpoint: POST /api/decision-intelligence/guidelines

Request Body:

{
  "condition": "cpu > 0.9",
  "action": "critical_alert",
  "prompt_template": "If system is under heavy load, prioritize stability."
}