Get Prompt

​

Overview

The prompts/get method retrieves a specific prompt template by name and processes it with provided arguments. The result contains structured messages ready for language model consumption.

​

Prerequisites

1. Session Initialization: Complete initialize → notifications/initialized flow
2. Authentication: Valid OAuth token or PAT with mcp.read scope
3. Session ID: Include Mcp-Session-Id header from initialization
4. Prompt Discovery: Use prompts/list to discover available prompts and their arguments

​

Request Format

{
  "jsonrpc": "2.0",
  "id": "prompts-get-123",
  "method": "prompts/get",
  "params": {
    "name": "deal_analysis",
    "arguments": {
      "context_type": "deal",
      "context_id": "12345"
    }
  }
}

​

Request Parameters

​

Response Format

​

Example Responses

{
  "jsonrpc": "2.0",
  "id": "prompts-get-123",
  "result": {
    "description": "Deal analysis prompt with contextual data",
    "messages": [
      {
        "role": "user",
        "content": {
          "type": "text",
          "text": "Please analyze deal #12345 and provide recommendations for improving the conversion probability. Consider the current stage, timeline, and stakeholder engagement.\n\nDeal Context:\n- ID: 12345\n- Current Stage: Negotiation\n- Value: $50,000\n- Timeline: 30 days remaining\n- Last Activity: 2 days ago"
        }
      }
    ]
  }
}

​

Argument Processing

​

Required Arguments

All arguments marked as required: true in the prompt definition must be provided:

// From prompts/list response
const prompt = {
  name: "deal_analysis",
  arguments: [
    { name: "context_type", required: true },
    { name: "context_id", required: true },
  ],
};

// Must provide all required arguments
const request = {
  name: "deal_analysis",
  arguments: {
    context_type: "deal", // Required
    context_id: "12345", // Required
  },
};

​

Optional Arguments

Optional arguments can be omitted and may have default values:

// Optional arguments in prompt definition
const prompt = {
  name: "sales_summary",
  arguments: [
    { name: "person_id", required: true },
    { name: "time_period", required: false }, // Optional
  ],
};

// Can omit optional arguments
const request = {
  name: "sales_summary",
  arguments: {
    person_id: "789",
    // time_period omitted - will use default
  },
};

​

Argument Validation

Arguments are validated for:

• Presence: Required arguments must be provided
• Type: String values expected for most arguments
• Context: References to valid CRM entities (deals, contacts, etc.)

​

Message Structure

​

Role-Based Messages

Messages follow conversational structure:

• user: Instructions or questions for the language model
• assistant: Pre-filled responses or examples

​

Content Types

​

Template Processing

Our implementation processes templates with provided arguments:

1. Argument Substitution: Replace placeholders with actual values
2. Context Loading: Fetch relevant data from CRM/database
3. Message Assembly: Construct structured messages
4. Validation: Ensure all required data is available

​

Content Validation Patterns

Based on our test implementation, message content is validated as follows:

// Content can be string or object
const content = message["content"];
assert(typeof content === "string" || typeof content === "object");

if (typeof content === "object") {
  // Object content must have type and appropriate fields
  assert("type" in content);
  assert(["text", "image", "audio"].includes(content.type));

  if (content.type === "text") {
    assert("text" in content);
    assert(typeof content.text === "string");
  }
}

Valid Content Examples:

// String content (simple format)
"content": "Please analyze this deal and provide recommendations"

// Object content (structured format)
"content": {
  "type": "text",
  "text": "Please analyze this deal and provide recommendations"
}

​

OAuth and Permissions

​

Prompt Access Control

• User-specific: Users see only prompts they have permission to access
• Database-driven: Prompt permissions stored and validated in database
• Session-bound: Permissions validated per session and prompt request

​

Context Data Access

• Reference Validation: Context IDs validated against user’s accessible data
• Permission Checking: User must have access to referenced CRM entities
• Data Filtering: Only authorized data included in prompt context

​

Error Handling

​

Common Error Codes

​

Argument Validation Errors

// Missing required argument
{
  code: -32602,
  message: "Invalid params",
  data: {
    parameter: "context_id",
    error: "Missing required argument",
    prompt_name: "deal_analysis"
  }
}

// Invalid context reference
{
  code: -32602,
  message: "Invalid params",
  data: {
    parameter: "context_id",
    error: "Invalid context reference",
    context_id: "999999"
  }
}

​

Implementation Notes

​

Request ID Requirements

• Must be unique within session
• Cannot be null (enforced by implementation)
• Used for error correlation and response matching

​

Session Validation

• Mcp-Session-Id header required after initialization
• Session must be active and authenticated
• Invalid session returns authentication error

​

Template Caching

• Prompt templates cached for performance
• Argument processing optimized for common patterns
• Context data fetched efficiently from database

​

Performance Considerations

• Complex prompts may take longer to process
• Context data fetching optimized per prompt type
• Results not cached (dynamic content with arguments)

This method provides the core functionality for retrieving personalized, context-aware prompt templates that can be directly used with language models for various CRM and business analysis tasks.