AI-powered shell assistant that helps you execute commands and file operations through natural language.

aish is an intelligent command-line assistant that uses Large Language Models (LLMs) to understand your natural language requests and execute shell commands and file operations safely and interactively.

• 🤖 AI-Powered: Uses LLMs to understand natural language and execute tasks
• 🔒 Security First: Built-in security features with path validation, operation permissions, and whitelist support
• 🎯 Interactive: Beautiful CLI interface with approval prompts for each operation
• 💬 Conversation Mode: Keep conversations open with -i flag for continuous interaction
• 🧠 Context Aware: Automatically detects workspace context including:
  • Programming languages and versions (Node.js, Python, Java, Rust, Go, PHP, Ruby, .NET)
  • Package managers and build tools
  • Web and backend frameworks
  • Testing frameworks and linters
  • Docker, Kubernetes, and container registries
  • CI/CD platforms (GitHub Actions, GitLab CI, Jenkins, etc.)
  • Cloud providers (AWS, GCP, Azure)
  • Infrastructure as Code tools (Terraform, Pulumi, CDK)
  • Databases and development tools
• 🔧 Flexible: Supports multiple LLM providers (OpenAI, OpenRouter, Custom)
• 📁 File Operations: Read, write, create directories, and list files
• 🚀 Shell Commands: Execute shell commands with real-time output
• ⚙️ Configurable: Easy-to-use configuration system with interactive wizard

git clone https://github.com/sammwyy/aish.git
cd aish
cargo build --release

The binary will be available at target/release/aish.

• Rust 1.70 or later
• An API key from an LLM provider (OpenAI, OpenRouter, etc.)

1. Initialize configuration:

   aish init

   This will guide you through setting up your LLM provider, API key, and preferences.

2. Start using aish:

   aish "list all Python files in the current directory"

3. Try interactive mode:

   aish -i "show me disk usage"

   This starts an interactive session where you can continue the conversation.

4. View configuration:

   aish config

# Execute a simple task
aish "show me disk usage"

# File operations
aish "create a new directory called 'projects'"

# Complex workflows
aish "find all .log files older than 30 days and delete them"

# Auto-approve all operations (use with caution)
aish --accept-all "run tests and commit if they pass"

Interactive mode (-i or --interactive) keeps the conversation open, allowing you to have a continuous dialogue with the AI assistant. The full conversation history is maintained, enabling references to previous commands and results.

# Start interactive mode (prompts for commands)
aish -i

# Start interactive mode with an initial command
aish -i "list all Python files"

# Interactive mode with auto-approve
aish -i --accept-all "run tests"

# Exit interactive mode by typing 'quit' or 'exit', or press Ctrl+C

Features:

• Full conversation history maintained across all prompts
• Type quit or exit to end the session
• Press Ctrl+C to exit at any time
• Empty prompts are ignored (allows you to skip)
• Errors don't terminate the session, allowing you to continue

# View all configuration
aish config

# Get a specific value
aish config llm.model

# Set a configuration value
aish config llm.max_tokens 8192
aish config security.allow_absolute_paths true

# View the current system prompt (with placeholders replaced)
aish showsystem

# Regenerate the system prompt template
aish regen

Configuration is stored in ~/.aish/config.toml and API keys in ~/.aish/tokens.env.

[llm]
provider = "OpenAI"
api_url = "https://api.openai.com/v1"
model = "gpt-4"
max_tokens = 4096

Supported Providers:

• OpenAI
• OpenRouter
• Custom (any OpenAI-compatible API)

[security]
allow_absolute_paths = false
allow_config_path_access = false
blocked_extensions = [".env"]

[security.allowed_operations]
"fs.makedir" = true
"fs.makefile" = true
"fs.writefile" = true
"fs.readfile" = true
"fs.listdir" = true
shell = true

The whitelist is used for auto-execution when --accept-all is enabled. Commands matching whitelist patterns will be automatically approved.

whitelist = [
  "^ls.*",
  "^cat.*\\.txt$",
  "^git status$"
]

Note: The whitelist only applies when using --accept-all. When running normally, all operations require user approval regardless of whitelist status.

Create a .aishignore file in ~/.aish/ to block access to specific file patterns (similar to .gitignore):

*.env
*.key
secrets/*
.git/*
node_modules/*

# Create pro
ect structure
aish "create a Python project with src, tests, and docs folders"

# Read and analyze files
aish "read all Python files in src/ and summarize the code structure"

# Batch operations
aish "rename all .txt files to .md in the current directory"

# System information
aish "show me disk usage and the top 5 largest directories"

# Development tasks
aish "run the tests and if they pass, commit with message 'fix: tests passing'"

# Media processing
aish "convert all .mov files to .mp4 with H.264 encoding"

# Data analysis
aish "analyze server.log, find errors in the last hour, and create a summary"

# Deployment
aish "backup the database, build the Docker image, and push to registry"

# Cleanup
aish "find and delete all .log files older than 30 days"

1. Approval Required: Every operation requires explicit approval (unless whitelisted with --accept-all)
2. Path Validation: Prevents access to config directory and absolute paths by default
3. Extension Blocking: Blocks sensitive file extensions like .env
4. Operation Permissions: Granular control over what operations are allowed
5. Pattern Matching: .aishignore file prevents access to sensitive patterns
6. Whitelist: Auto-approve specific command patterns (only with --accept-all)

Contributions are welcome! Please feel free to submit a Pull Request.

1. Fork the repository
2. Create your feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add some amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

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

Sammwy

• GitHub: @sammwyy

• Built with Rust
• Uses inquire for beautiful CLI prompts
• Uses colored for terminal colors

This tool executes shell commands and file operations based on AI interpretation. Always review proposed operations before approval. Use --accept-all with extreme caution and only with trusted whitelisted commands.