Skip to content

MCP Resources & Prompts

Guide to MCP resources for workflow discovery and prompt templates for workflow generation.


Overview

The MCP server exposes resources and prompts that can be accessed directly by Claude AI or other MCP clients.

Resource Types

Static Resources: - n8n://workflows - List all workflows metadata - n8n://execution-stats - Execution statistics summary

Dynamic Resources: - n8n://workflows/{id} - Specific workflow details - n8n://executions/{id} - Specific execution details

Prompt Templates: - 5 workflow generation templates for common patterns


MCP Resources

n8n://workflows

List all workflows with metadata (read-only resource).

URI: n8n://workflows

Query Parameters: - instance (optional) - Instance identifier

Response Format:

{
  workflows: Array<{
    id: string;
    name: string;
    active: boolean;
    createdAt: string;
    updatedAt: string;
    nodes?: number;
    tags?: string[];
  }>;
  totalCount: number;
  generatedAt: string;
}

Equivalent Tool: list_workflows()

Use Case: Quick overview of all workflows without using a tool call.


n8n://execution-stats

Execution statistics summary.

URI: n8n://execution-stats

Response Format:

{
  summary: {
    total: number;
    successful: number;
    failed: number;
    running: number;
  };
  recentExecutions: Execution[];
  byWorkflow: Array<{
    workflowId: number;
    workflowName: string;
    executions: number;
    successRate: number;
  }>;
  timeRange: {
    from: string;
    to: string;
  };
  generatedAt: string;
}

Use Case: Dashboard-style overview of execution health.


n8n://workflows/{id}

Get specific workflow details.

URI Pattern: n8n://workflows/{workflowId}

Parameters: - {workflowId} - Workflow ID - Query: instance (optional)

Equivalent Tool: get_workflow({ id })

Example: n8n://workflows/15?instance=production


n8n://executions/{id}

Get specific execution details.

URI Pattern: n8n://executions/{executionId}

Equivalent Tool: get_execution({ id })

Example: n8n://executions/789


MCP Prompts

Prompt templates for generating common workflow patterns.

1. schedule-triggered-workflow

Generate a workflow with schedule trigger.

Arguments:

{
  workflow_name: string;         // Workflow name
  cron_schedule: string;         // Cron expression (e.g., "0 9 * * *")
  action_description: string;    // What the workflow does
  timezone?: string;             // Timezone (default: UTC)
}

Example Usage:

Use prompt: schedule-triggered-workflow
Arguments:
  workflow_name: "Daily Backup"
  cron_schedule: "0 2 * * *"
  action_description: "Backup database to S3"
  timezone: "America/New_York"

Generated Workflow: - Schedule Trigger node (cron-based) - Action nodes based on description - Error handling nodes


2. http-webhook-workflow

Generate a webhook receiver workflow.

Arguments:

{
  workflow_name: string;         // Workflow name
  webhook_path: string;          // Webhook URL path
  response_mode?: string;        // 'onReceived' | 'lastNode'
  processing_description: string; // What to do with webhook data
}

Example Usage:

Use prompt: http-webhook-workflow
Arguments:
  workflow_name: "Order Processor"
  webhook_path: "process-order"
  response_mode: "lastNode"
  processing_description: "Validate order and save to database"

Generated Workflow: - Webhook Trigger node - Data processing nodes - Response node


3. data-transformation-workflow

Generate data transformation pipeline.

Arguments:

{
  workflow_name: string;            // Workflow name
  source_description: string;       // Data source
  transformation_description: string; // Transformations needed
  destination_description: string;   // Where to send data
}

Example Usage:

Use prompt: data-transformation-workflow
Arguments:
  workflow_name: "CSV to Database"
  source_description: "Read CSV file from S3"
  transformation_description: "Parse CSV, validate, enrich with API data"
  destination_description: "Insert into PostgreSQL"

Generated Workflow: - Source node (file, API, database) - Transformation nodes (Set, Code, Function) - Destination node


4. external-service-integration

Generate service integration workflow.

Arguments:

{
  workflow_name: string;         // Workflow name
  service_name: string;          // Service to integrate (Slack, Gmail, etc.)
  trigger_description: string;   // What triggers the workflow
  action_description: string;    // What action to take
}

Example Usage:

Use prompt: external-service-integration
Arguments:
  workflow_name: "Slack Alert System"
  service_name: "Slack"
  trigger_description: "When error occurs in application"
  action_description: "Send alert to #alerts channel"

Generated Workflow: - Service-specific trigger - Processing nodes - Service action node


5. api-data-polling

Generate API polling workflow.

Arguments:

{
  workflow_name: string;         // Workflow name
  api_endpoint: string;          // API URL to poll
  poll_interval: string;         // How often to poll (cron)
  data_processing: string;       // What to do with API response
}

Example Usage:

Use prompt: api-data-polling
Arguments:
  workflow_name: "Weather Sync"
  api_endpoint: "https://api.weather.com/current"
  poll_interval: "0 */6 * * *"  // Every 6 hours
  data_processing: "Save to database and send notifications if extreme weather"

Generated Workflow: - Schedule Trigger (polling interval) - HTTP Request node (API call) - Processing nodes - Action nodes


Using Prompts

Via MCP Client

// List available prompts
const prompts = await client.listPrompts()

// Get specific prompt
const workflow = await client.getPrompt({
  name: "http-webhook-workflow",
  arguments: {
    workflow_name: "Order Processor",
    webhook_path: "process-order",
    processing_description: "Validate and store order"
  }
})

// Create workflow from template
await client.callTool({
  name: "create_workflow",
  arguments: JSON.parse(workflow.messages[0].content.text)
})

Via Claude Desktop

Request:

Create a schedule-triggered workflow called "Daily Report"
that runs at 9 AM and generates a sales report

Claude recognizes the pattern and may use prompt template internally


Resources vs Tools

Feature Resources Tools
Access Method Direct URI access Tool invocation
Caching Can be cached by client Not cached
Updates Real-time data On-demand
Use Case Quick reference Actions & operations
Performance Faster (cached) Slower (API call)

When to Use Resources: - Quick data lookup - Dashboard views - Cached information

When to Use Tools: - Modify workflows - Execute operations - Get latest data


Next Steps


MCP Protocol

Learn more about MCP resources and prompts at: Model Context Protocol Specification