🛡️ SentinelTM: Your AI Development Companion

Transform Your IDE into an Intelligent Development Environment
Agent-Oriented Architecture • Context-Aware AI • Zero Tolerance Quality

SentinelTM (st) is a revolutionary Agent Companion that transforms how you build software. Combining enterprise-grade CLI tools with cutting-edge MCP (Model Context Protocol), intelligent workflows, and context-aware AI, it provides a seamless development experience across 20+ AI providers and 23+ IDEs.

🌟 Revolutionary Features (v2.2+)

🤝 Friendly AI Companion

• Personalized Interaction: Configure tone, language, and personality through friendly.yml
• Smart Aliases: Use natural language commands (Persian/English)
• Context Memory: Remembers your preferences and workflows
• Encouragement System: Celebrates your successes and helps with challenges

🎟️ Context Window Monitor

• Real-Time Tracking: Monitor token usage across all AI operations
• Smart Warnings: Get alerts at 85% and 95% usage
• Estimated Turns: Know how many interactions remain
• MCP Integration: Available as sentinel_context_monitor tool

🏗️ High-Level Workflows

• Declarative Operations: From imperative commands to intelligent workflows
• Component Scaffolding: st workflow component UserProfile generates complete structure
• API Endpoints: Auto-generate controllers with tests and validation
• Project-Wide Refactoring: Rename symbols across entire codebase

🧠 Context Enrichment System

• Pattern Detection: Learns from your existing code
• Smart Suggestions: AI generates code matching your project style
• Dependency Analysis: Understands your tech stack
• Convention Extraction: Follows your naming and structure patterns

📜 Named Scripts System

• Reusable Workflows: Define complex operations in scripts.yml
• Security Validation: Whitelist/blacklist commands
• Parameter Substitution: Dynamic script execution
• Confirmation Prompts: For sensitive operations

✨ Core Capabilities

• 🎯 Zero Tolerance Quality: AI-powered code review with structured feedback
• 🔐 Security-First: Sandboxed file system and controlled shell execution
• 🤖 Provider Agnostic: Works with Ollama, OpenAI, Claude, Gemini, Mistral, OpenRouter, and any OpenAI-compatible API
• 🚀 Interactive Setup: Guided configuration with smart defaults
• 📊 Quality Gates: Automated compliance checking with actionable insights
• 🎨 IDE Integration: One-command setup for VS Code, Cursor, Zed, Windsurf, and 19+ more

🚀 Quick Start (2 Minutes)

1️⃣ Install

npm install -g sentineltm-cli@latest

2️⃣ Configure Your AI Provider (Interactive)

SentinelTM provides an interactive setup wizard for seamless configuration:

st set provider

This will guide you through:

1. 📋 Select Provider: Choose from Ollama, OpenAI, Claude, Gemini, Mistral, or OpenRouter
2. 🔑 Enter API Key: Provide your API credentials (or skip for Ollama)
3. 🤖 Select Model: Pick from available models (auto-fetched from the provider)

Example Providers:

Ollama (Local, Free)

# Install Ollama first: https://ollama.ai
ollama pull llama3

# Then configure
st set provider
# → Select: ollama
# → API Key: (leave blank)
# → Model: llama3

OpenRouter (Best for Production)

st set provider
# → Select: openrouter
# → API Key: sk-or-v1-xxxxx
# → Model: deepseek/deepseek-chat or x-ai/grok-2-latest

OpenAI

st set provider
# → Select: openai
# → API Key: sk-xxxxx
# → Model: gpt-4-turbo

3️⃣ Set Up IDE Integration (Interactive)

Configure your development environment with one command:

st set ide

This will:

1. ✅ Select IDEs: Choose from VS Code, Cursor, Zed, Windsurf, Continue, Cline, etc.
2. 🤖 Configure Provider: Select AI provider and model (if not already set)
3. 📜 Apply Rules: Optionally apply Zero Tolerance coding standards

4️⃣ Verify Installation

# Check provider status
st provider status

# Run quality gate
st gate run

📚 Command Reference

🤖 Provider Management

🎨 IDE Configuration

Supported IDEs: VS Code, Cursor, Zed, Windsurf, Continue, Cline, Codex, Claude, Gemini, OpenCode, Roo, Amp, Kilo, Trae, Kiro

✅ Quality Gates

Built-in Checks:

• 🧪 Tests
• 🎨 Linting & Formatting
• 🏗️ Build Verification
• 🔒 Security Scan
• 📁 Project Structure
• 🧹 Code Hygiene
• 🤖 AI Rule Check (enforces Zero Tolerance standards)

📂 File System Operations

🔍 Code Indexing

📊 Dashboard & Reporting

🛠️ Task Management

🏗️ Workflow Commands (v2.2+)

🎟️ Context Monitor (v2.2+)

🤝 Friendly Companion (v2.2+)

Example Aliases (configure in friendly.yml):

aliases:
  وضعیت: "st status"                    # Status
  تست_کامل: "st gate run"              # Full test
  کامپوننت_جدید: "st workflow component $1"  # New component

🎯 Zero Tolerance Contract

SentinelTM enforces strict coding standards through its Zero Tolerance Contract system:

• ✅ Type safety and hints required
• ✅ No hardcoded values (strings, numbers, URLs)
• ✅ Configuration externalized to YAML
• ✅ Modular architecture (max 300 lines per file)
• ✅ PEP8/ESLint compliance
• ✅ Comprehensive error handling
• ✅ Security best practices

See ZERO_TOLERANCE_CONTRACT.md for full details.

🔧 Configuration Files

SentinelTM uses multiple configuration files for maximum flexibility:

Core Configuration (.sentineltm/config/config.json)

{
  "defaults": {
    "provider": "openrouter",
    "model": "deepseek/deepseek-chat"
  },
  "providers": {
    "openrouter": {
      "type": "openai-compatible",
      "baseURL": "https://openrouter.ai/api",
      "apiKey": "sk-or-v1-xxxxx"
    }
  }
}

Friendly Companion (friendly.yml) - v2.2+

personality:
  tone: "enthusiastic_partner"
  use_emojis: true
  user_name: "Your Name"
  
aliases:
  status: "st status"
  test: "st gate run"
  component: "st workflow component $1"

preferences:
  default_component_path: "src/components"
  min_quality_score: 85
  message_language: "english"  # or "persian"

Named Scripts (scripts.yml) - v2.2+

scripts:
  install: "npm install"
  test: "npm test"
  deploy:
    - "npm run build"
    - "npm test"
    - "st gate run"
    - "npm publish"

security:
  allowedCommands: ["npm", "git", "st"]
  requireConfirmation: ["deploy", "publish"]

Project Rules (project_rules.md) - v2.1+

# Project Coding Standards

- All state management must use Zustand
- API calls through `src/api/client.ts` only
- Use react-query for data fetching
- Functional components with TypeScript types

🔌 MCP Tools Integration

SentinelTM provides powerful MCP tools for IDE integration:

Available Tools

Example MCP Usage

// Monitor context window
const stats = await mcp.callTool('sentinel_context_monitor', {
  action: 'getStats'
});
// Returns: { usagePercentage: 65, remainingTurns: 12, warningLevel: 'safe' }

// Record AI operation
await mcp.callTool('sentinel_context_monitor', {
  action: 'recordUsage',
  payload: {
    inputTokens: 1500,
    outputTokens: 800,
    operation: 'code_generation'
  }
});

// Scaffold component
await mcp.callTool('sentinel_task', {
  action: 'createTask',
  payload: {
    title: 'Build UserProfile component',
    priority: 'high'
  }
});

📖 Advanced Guides

• API Integration Guide - Deep dive into provider configuration
• Zero Tolerance Contract - Complete coding standards
• MCP Server Usage - Using SentinelTM as an MCP server

🆘 Troubleshooting

Provider not working?

st provider status    # Check configuration
st provider detect    # Auto-detect available providers

IDE configuration not applying?

st ide list          # Check supported IDEs
st set ide           # Re-run interactive setup

Quality gate failing?

st gate run --min 70  # Lower threshold for testing

🤝 Contributing

We welcome contributions! Please:

1. Fork the repository
2. Create a feature branch
3. Submit a pull request

📄 License

ISC License © 2025 Amin Azmoodeh