Snow-Flow: ServiceNow Development Platform

Snow-Flow is an AI-powered development platform for ServiceNow that provides 180+ MCP tools and multi-agent coordination. It enables developers and administrators to interact with ServiceNow through natural language commands and automated workflows.

Quick Start - One Command to Rule Them All

🚀 The Swarm Command

The simplest way to use Snow-Flow is with the swarm command. Just describe what you want in natural language:

# Basic usage - no flags needed!
snow-flow swarm "Create an incident dashboard with real-time updates"

This single command automatically:

• Spawns multiple specialized AI agents
• Discovers existing ServiceNow artifacts
• Creates complete working solutions
• Deploys directly to your instance
• Handles all technical complexity

More examples:

snow-flow swarm "Build employee onboarding workflow"
snow-flow swarm "Create knowledge base for IT support"
snow-flow swarm "Set up change management process"
snow-flow swarm "Analyze last month's incident patterns"

Core Functionality

Snow-Flow connects directly to ServiceNow instances through OAuth 2.0 authentication and provides:

• Direct API integration with ServiceNow REST endpoints
• Multi-agent system with specialized agents for different tasks
• Natural language interface for ServiceNow operations
• Machine learning capabilities using TensorFlow.js
• Process automation and workflow optimization
• Real-time logging and token tracking for API operations

API Operation Visibility

Snow-Flow displays detailed information during API calls:

⏳ Operation in progress... (2s elapsed, 150 tokens used)
🔄 API Call: GET /api/now/table/incident (10 records)
📊 Tokens used: 450 (in: 100, out: 350)
✅ Operation complete in 3 seconds

MCP operations run in separate processes and their logs are sent to stderr for visibility in the console.

ServiceNow API Integration

Authentication

Snow-Flow uses OAuth 2.0 for secure authentication with ServiceNow:

snow-flow auth login
# Opens ServiceNow OAuth consent screen
# Authenticate with ServiceNow credentials
# Token stored locally in ~/.snow-flow/auth.json

Security

• OAuth 2.0 authentication protocol
• Respects ServiceNow ACLs and user permissions
• Local token storage with encryption
• No external data transmission
• All operations logged in ServiceNow audit trails

Network Architecture

┌─────────────┐    OAuth 2.0 + REST APIs    ┌─────────────────┐
│  Snow-Flow  │ ◄─────────────────────────► │ Your ServiceNow │
│   (Local)   │      HTTPS Only              │    Instance     │
└─────────────┘                              └─────────────────┘

Available Tools

Snow-Flow provides 180+ tools across 17 specialized MCP servers:

Complete Tool Reference

MCP Server Overview

Detailed Tool Listing by Server

1. ATF Testing Tools (servicenow-automation)

2. Knowledge Management (servicenow-knowledge-catalog)

3. Service Catalog (servicenow-knowledge-catalog)

4. Change Management (servicenow-change-virtualagent-pa)

5. Virtual Agent (servicenow-change-virtualagent-pa)

6. Performance Analytics (servicenow-change-virtualagent-pa)

7. Flow Designer (servicenow-flow-workspace-mobile)

8. Agent Workspace (servicenow-flow-workspace-mobile)

9. Mobile Platform (servicenow-flow-workspace-mobile)

10. CMDB & Discovery (servicenow-cmdb-event-hr-csm-devops)

11. Event Management (servicenow-cmdb-event-hr-csm-devops)

12. HR Service Delivery (servicenow-cmdb-event-hr-csm-devops)

13. Customer Service Management (servicenow-cmdb-event-hr-csm-devops)

14. DevOps (servicenow-cmdb-event-hr-csm-devops)

Table Operations

The universal snow_query_table tool works with all ServiceNow tables:

// Query any table with optimized performance
snow_query_table({ 
  table: "incident",  // Works with any table
  query: "state!=7",
  fields: ["number", "short_description", "priority"],
  limit: 100
})

// Analytics mode - retrieve all records with minimal fields
snow_query_table({ 
  table: "incident",
  fields: ["sys_created_on", "priority"]  // Omit limit for all records
})

// Count-only mode for large datasets
snow_query_table({ 
  table: "sc_request",
  include_content: false  // Returns count only
})

// Server-side aggregation
snow_query_table({ 
  table: "change_request",
  group_by: "risk",
  order_by: "-count"
})

Development Tools

• Widget creation and deployment (snow_deploy for new, snow_update for existing)
• Business rule development
• Script include management
• Flow Designer automation
• UI Page creation
• Portal configuration

Service Management

• Incident management
• Change management with CAB scheduling
• Problem management
• Request fulfillment
• Knowledge management

Enterprise APIs

• Performance Analytics (PA) with KPI management
• Virtual Agent configuration
• Agent Workspace customization
• Mobile platform deployment
• CMDB and Discovery
• Event Management
• HR Service Delivery
• Customer Service Management
• DevOps integration

Machine Learning

Snow-Flow includes TensorFlow.js for neural network capabilities:

Incident Classification

Train LSTM networks on historical incident data:

snow-flow ml train-incident-classifier --sample-size 5000 --epochs 100
snow-flow ml classify-incident INC0123456

Change Risk Prediction

Analyze change requests to predict implementation risk:

snow-flow ml train-change-risk --include-failed-changes
snow-flow ml predict-change-risk CHG0123456

Time Series Forecasting

Predict future incident volumes:

snow-flow ml forecast-incidents --days 7 --category network

Anomaly Detection

Detect unusual patterns using autoencoder networks:

snow-flow ml detect-anomalies --metric incident_patterns --sensitivity 0.9

ServiceNow Native ML Integration

Access ServiceNow's native ML capabilities when licensed:

Performance Analytics ML

snow-flow ml performance-analytics --indicator "Incident Resolution Time" --forecast-days 90

Predictive Intelligence

snow-flow ml predictive-intelligence --operation similar_incidents --incident INC0123456

Agent Intelligence

snow-flow ml agent-intelligence --task incident --id INC0123456 --auto-assign

Note: Native ML features require appropriate ServiceNow plugin licenses (PA, PI, Agent Intelligence).

Multi-Agent System

Snow-Flow includes 38 specialized agents that work together:

Agent Types

• Architect: System design and architecture planning
• Developer: Code generation and implementation
• Tester: Test creation and execution
• Analyst: Data analysis and reporting
• Security: Compliance and security validation
• Optimizer: Performance optimization
• Documenter: Documentation generation

Coordination Modes

• Hierarchical: Top-down task delegation
• Mesh: Peer-to-peer collaboration
• Swarm: Distributed problem solving
• Adaptive: Dynamic reorganization based on task requirements

Installation

Prerequisites

• Node.js 18.0 or higher
• npm 8.0 or higher
• ServiceNow instance with admin access
• OAuth application configured in ServiceNow

Install via npm

npm install -g snow-flow

Initial Setup

# Initialize configuration
snow-flow init

# Configure authentication
snow-flow auth login

# Verify connection
snow-flow auth test

Troubleshooting: Claude Code stdio Error

If you encounter a stdio/permissions error when using Snow-Flow with Claude Code:

# 1. Fix Claude permissions
claude --dangerously-skip-permissions

# 2. Activate Claude Code
# Open Claude Code and ensure it's activated with your account

# 3. Login to Claude
# Sign in with your Claude account in the Claude Code interface

# 4. Close Claude Code completely
# Make sure Claude Code is fully closed before proceeding

# 5. Now run Snow-Flow swarm
snow-flow swarm "Your ServiceNow task here"

This resolves the common "stdio" error that occurs when Claude Code doesn't have proper permissions to communicate with MCP servers.

Usage Examples

Swarm Command (Recommended)

# The simplest way - just describe what you want!
snow-flow swarm "Create incident management dashboard"

# No flags needed for basic usage
snow-flow swarm "Build service catalog for laptop requests"

# Advanced usage with optional flags
snow-flow swarm "Optimize incident resolution process" --monitor --parallel

Natural Language Commands

# Create an incident dashboard
snow-flow sparc "Create incident management dashboard with SLA tracking"

# Build a workflow
snow-flow sparc "Create employee onboarding workflow with approval steps"

# Analyze data
snow-flow sparc "Analyze incident resolution patterns from last quarter"

Direct Tool Usage

# Query incidents
snow-flow query --table incident --limit 10

# Create a widget
snow-flow deploy widget --name "SLA Dashboard" --template dashboard

# Generate documentation
snow-flow generate docs --scope "incident_management"

Advanced Swarm Options

# With specific strategy
snow-flow swarm "Build complete HR service portal" \
  --strategy development \
  --mode hierarchical \
  --max-agents 8

# Parallel analysis
snow-flow swarm "Analyze system performance bottlenecks" \
  --strategy analysis \
  --mode mesh \
  --parallel

# With monitoring
snow-flow swarm "Deploy production changes" \
  --monitor \
  --auto-rollback

Configuration

Configuration file location: ~/.snow-flow/.env

Required Settings

SNOW_INSTANCE=your-instance.service-now.com
SNOW_CLIENT_ID=your-oauth-client-id
SNOW_CLIENT_SECRET=your-oauth-client-secret

Optional Settings

# Timeout configuration (milliseconds)
# SNOW_FLOW_INIT_TIMEOUT=30000
# SNOW_FLOW_OPERATION_TIMEOUT=60000
# SNOW_FLOW_MEMORY_TIMEOUT=5000

# Agent configuration
SNOW_FLOW_MAX_AGENTS=10
SNOW_FLOW_AGENT_MODE=hierarchical

# ML configuration
SNOW_FLOW_ML_ENABLED=true
SNOW_FLOW_ML_MODEL_PATH=./models

Project Structure

snow-flow/
├── src/                 # Source code
│   ├── agents/         # Agent implementations
│   ├── mcp/           # MCP server implementations
│   ├── utils/         # Utility functions
│   └── services/      # Service layer
├── dist/               # Compiled JavaScript
├── memory/             # Persistent storage
├── models/             # ML models
└── tests/              # Test suite

Documentation

• API Reference
• Agent Documentation
• MCP Tools Reference
• ML Capabilities
• Configuration Guide

Contributing

Contributions are welcome. Please read the Contributing Guide for details on our code of conduct and the process for submitting pull requests.

Support

• GitHub Issues: github.com/groeimetai/snow-flow/issues
• Documentation: github.com/groeimetai/snow-flow/wiki

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• ServiceNow for the comprehensive REST API
• TensorFlow.js for machine learning capabilities
• Model Context Protocol (MCP) for tool orchestration