moma-cli 0.0.1

MOMA CLI - A terminal-based AI coding assistant

moma-cli

A terminal‑based AI coding assistant powered by litellm. It lets you chat with your codebase, automatically read context, and apply changes manually.

Overview

Installation

pip install moma-cli   # or install from the repository

Quick start

# Start an interactive chat (default command)
moma

The CLI will launch a welcome screen and you can start typing messages. Commands are prefixed with /:

Command	Description
/clear	Clear conversation history
/model list	List available models
/model change	Choose a different model
/security	Change security mode (smart/strict/permissive)
/status	Show current session status
/config …	View or edit configuration
/help	Show help panel
/quit or Ctrl‑D	Exit the program

Chat command

You can also invoke the chat directly with arguments:

moma chat "Explain this function"            # send an initial message
moma chat -m gpt-4 "Generate a unit test"    # override model for this session
moma chat -c                               # clear history before starting
moma chat -t 0.7 -max-tokens 500            # set temperature and max tokens

Configuration

Configuration is stored in ~/.moma-cli/config.json. Manage it via the config sub‑command:

# Show all configuration
moma config show

# Set a value (e.g., default model)
moma config set default_model ollama/llama3

# Set your API key
moma config set api_key sk-xxxxxxxxxxxx

# Reset system prompt to default
moma config reset

Security modes

• smart – read‑only shell commands are auto‑executed, mutating commands require confirmation.
• strict – all shell commands require confirmation.
• permissive – all commands run without prompts.

Change the mode with:

moma /security

Example workflow

1. Start chat – moma
2. Ask a question – e.g., How can I improve this function?
3. Review suggestions – the assistant may propose a diff.
4. Apply changes – confirm the diff when prompted.

Enjoy coding with AI assistance!