Agent Reference

Agent Reference

This page documents the core MuleSoft agents behind the Hushh Agentic Developer APIs.

A2A JSON-RPC 2.0 envelope

{
  "jsonrpc": "2.0",
  "id": "task-124",
  "method": "tasks/send",
  "params": {
    "sessionId": "session-456",
    "message": {
      "role": "user",
      "parts": [
        {
          "type": "text",
          "text": "Provide a detailed JSON profile for Sundar Pichai, email sundar.pichai@example.com, phone +1 6505559001."
        }
      ]
    }
  }
}

Integration flow in Hushh

1. User submits name, email, and phone
2. MuleSoft constructs the JSON-RPC request
3. Agent responds with a userProfile JSON payload
4. DataWeave transforms and stores the profile in Hushh data stores

Error handling

Code	Description	Resolution
400	Invalid JSON-RPC format	Validate the request body schema
401	Unauthorized	Check credentials or rotate keys
500	Internal agent error	Retry with backoff or contact support
TIMEOUT	Agent took too long	Reduce payload size or split requests

OpenAI Public Data Agent

• Purpose: enrich profiles from public signals
• Model: GPT-4.0 mini
• Proxy endpoint: POST /api/a2a/public
• MuleSoft endpoint: https://hushh-open-ai-agent-app-bubqpu.5sc6y6-3.usa-e2.cloudhub.io/public-data-agent

Sample prompt (Sundar Pichai)

Provide a detailed JSON profile for Sundar Pichai, email sundar.pichai@example.com, phone +1 6505559001.

Sample response (trimmed)

{
  "jsonrpc": "2.0",
  "id": "task-124",
  "result": {
    "sessionId": "session-456",
    "status": {
      "state": "completed",
      "message": {
        "role": "agent",
        "parts": [
          {
            "type": "text",
            "text": {
              "userProfile": {
                "fullName": "Sundar Pichai",
                "phone": "+1 6505559001",
                "email": "sundar.pichai@example.com",
                "city": "Mountain View",
                "occupation": "CEO",
                "dietPreference": "Vegetarian"
              }
            }
          }
        ]
      }
    }
  }
}

Gemini Public Data Agent

• Purpose: high-confidence enrichment and gap filling
• Model: gemini-2.5-pro
• Proxy endpoint: POST /api/a2a/gemini
• MuleSoft endpoint: https://hushh-gemini-ai-agent-app-bubqpu.5sc6y6-3.usa-e2.cloudhub.io/public-data-agent

Supabase Profile Creation Agent

• Purpose: create a structured profile from natural language
• Proxy endpoint: POST /api/a2a/hushh-profile
• MuleSoft endpoint: https://hushh-supabase-agent-app-bubqpu.5sc6y6-2.usa-e2.cloudhub.io/supabase-agent

Example prompt:

Create a profile for Sundar Pichai, email sundar.pichai@example.com, phone +1 6505559001.

Supabase Profile Update Agent

• Purpose: update profile fields with natural language
• Proxy endpoint: POST /api/a2a/hushh-profile-update
• MuleSoft endpoint: https://hushh-supabase-agent-app-bubqpu.5sc6y6-2.usa-e2.cloudhub.io/supabase-agent

Example prompt:

Update the city for Sundar Pichai to Mountain View.

Supabase Profile Query Agent

• Purpose: fetch full profiles from Supabase
• Proxy endpoint: POST /api/a2a/hushh
• MuleSoft endpoint: https://hushh-supabase-query-agent-app-bubqpu.5sc6y6-2.usa-e2.cloudhub.io/supabase-query-agent

Example prompt:

Fetch all details for Sundar Pichai with phone +1 6505559001.

Brand User Data Query Agent

• Purpose: retrieve brand preferences and CRM-ready insights
• Model: GPT-4.0 mini
• Proxy endpoint: POST /api/a2a/brand
• MuleSoft endpoint: https://hushh-brand-agent-app-bubqpu.5sc6y6-4.usa-e2.cloudhub.io/crm-agent

This agent acts as the bridge between brand assistants and CRM data, returning a conversational summary plus structured insights.

Example prompt:

Summarize brand preferences and purchase intents for Sundar Pichai.

Example queries

• Fetch all details for Sundar Pichai with phone +1 6505559001.
• Fetch all intents for Sundar Pichai with phone +1 6505559001.
• Fetch all wants and desires for Sundar Pichai with phone +1 6505559001.
• What is the occupation and education level for Sundar Pichai?