CmdAssist - AI-Powered Command Line Assistant

CmdAssist is a PowerShell module that provides AI-powered command line assistance. It helps users by suggesting appropriate commands based on natural language prompts using various Large Language Models (LLMs).

Features

• 🤖 Multiple AI Providers: Support for OpenAI, Azure OpenAI, Claude, Llama, Google Gemini, and DeepSeek
• ⚡ PowerShell Integration: Native PowerShell cmdlet with proper parameter support
• 🛡️ Safety First: Confirmation prompts before executing potentially dangerous commands
• 📋 Clipboard Support: Copy commands to clipboard without execution
• 🔧 Context Aware: Uses current working directory and system information
• 📊 Confidence Scoring: AI responses include confidence levels

Installation

Prerequisites

• .NET 8.0 SDK or later
• PowerShell 7.4 or later

Build from Source

1. Clone the repository:

git clone https://github.com/kunalk16/cmdassist.git
cd cmdassist

2. Build the project:

dotnet build

3. Import the module:

Import-Module ./src/CmdAssist.PowerShell/bin/Debug/net8.0/CmdAssist.psd1

Configuration

CmdAssist uses environment variables for API configuration. Set the appropriate variables for your chosen AI provider:

OpenAI

$env:OPENAI_API_KEY = "your-openai-api-key"
$env:OPENAI_MODEL = "gpt-4"  # Optional, defaults to gpt-4
$env:OPENAI_ORGANIZATION = "your-org-id"  # Optional

Azure OpenAI

$env:AZURE_OPENAI_API_KEY = "your-azure-openai-key"
$env:AZURE_OPENAI_ENDPOINT = "https://your-resource.openai.azure.com/"
$env:AZURE_OPENAI_DEPLOYMENT_NAME = "your-deployment-name"
$env:AZURE_OPENAI_API_VERSION = "2024-02-01"  # Optional

Claude

$env:CLAUDE_API_KEY = "your-claude-api-key"
$env:CLAUDE_MODEL = "claude-3-sonnet-20240229"  # Optional

Llama

$env:LLAMA_API_URL = "https://your-llama-endpoint.com"
$env:LLAMA_API_KEY = "your-api-key"  # Optional, depending on your setup
$env:LLAMA_MODEL = "llama2-70b-chat"  # Optional

Google Gemini

$env:GEMINI_API_KEY = "your-gemini-api-key"
$env:GEMINI_MODEL = "gemini-pro"  # Optional, defaults to gemini-pro

DeepSeek

$env:DEEPSEEK_API_KEY = "your-deepseek-api-key"
$env:DEEPSEEK_API_URL = "https://api.deepseek.com/v1"  # Optional
$env:DEEPSEEK_MODEL = "deepseek-chat"  # Optional, defaults to deepseek-chat

API Key Resources

• OpenAI: Get your API key at platform.openai.com
• Azure OpenAI: Set up through Azure Portal
• Claude: Get API access at console.anthropic.com
• Llama: Use any Llama-compatible endpoint (Ollama, LM Studio, etc.)
• Google Gemini: Get API key at aistudio.google.com
• DeepSeek: Register at platform.deepseek.com

Usage

Basic Usage

# Get a command suggestion with confirmation prompt
cmd-assist "list all .txt files in the current directory"

# Execute directly without confirmation
cmd-assist "create a new folder called 'documents'" -Confirm

# Use a specific AI provider
cmd-assist "show disk usage" -Provider Claude

# Enable verbose output
cmd-assist "find large files" -Verbose

Examples

# File operations
cmd-assist "copy all .pdf files to a backup folder"
cmd-assist "find files modified in the last 7 days"
cmd-assist "compress the logs folder"

# System administration
cmd-assist "check system memory usage"
cmd-assist "restart the print spooler service"
cmd-assist "show network connections"

# Development tasks
cmd-assist "initialize a git repository"
cmd-assist "build a .NET project in release mode"
cmd-assist "run unit tests and show coverage"

Interactive Mode

When you run a command without the -Confirm switch, CmdAssist will:

1. Display the suggested command
2. Show an explanation of what it does
3. Ask for confirmation with options:
   • Yes: Execute the command
   • No: Cancel execution
   • Copy: Copy the command to clipboard without executing

Command Line Parameters

• Prompt (Position 0, Mandatory): The natural language description of what you want to accomplish
• Confirm (Switch): Execute the command without confirmation prompt
• Provider (Optional): Specify AI provider (OpenAI, AzureOpenAI, Claude, Llama, Gemini, DeepSeek)
• Verbose (Switch): Enable verbose output for debugging

Safety Features

• Confirmation Prompts: By default, all commands require user confirmation
• Confidence Scoring: Low-confidence responses trigger additional warnings
• Context Awareness: Commands are generated with awareness of your current environment
• Error Handling: Comprehensive error handling with helpful messages

Troubleshooting

Common Issues

1. API Key Not Set

   Error: OpenAI API key not configured. Set the OPENAI_API_KEY environment variable.
   

   Solution: Set the appropriate environment variable for your chosen provider.

2. Network Connection Issues

   Error: API request failed: Unauthorized
   

   Solution: Verify your API key and network connection.

3. Module Import Errors

   Import-Module: Could not load file or assembly
   

   Solution: Ensure .NET 8.0 runtime is installed and try rebuilding the project.

Enable Debug Logging

cmd-assist "your prompt" -Verbose

Development

Project Structure

cmdassist/
├── src/
│   └── CmdAssist.PowerShell/
│       ├── Cmdlets/                 # PowerShell cmdlet implementations
│       ├── Services/                # AI service implementations
│       ├── Models/                  # Data models and DTOs
│       ├── CmdAssist.psd1          # PowerShell module manifest
│       └── CmdAssist.PowerShell.csproj
├── Directory.Build.props            # Common build properties
├── Directory.Packages.props         # Package version management
├── global.json                      # .NET SDK version
└── cmdassist.sln                   # Solution file

Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests if applicable
5. Submit a pull request

Running Tests

dotnet test

Building Release Version

dotnet build -c Release

License

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

Support

• 📚 Documentation
• 🐛 Issues
• 💬 Discussions

---

⚠️ Important: Always review generated commands before execution, especially when using the -Confirm switch. CmdAssist is a tool to assist with command generation, but you remain responsible for the commands you choose to execute.