AI Toolbox

This repository contains the code for the AI Toolbox Project from my Prompt Engineering course. The entire code for the course lectures can be found here: https://github.com/lm-academy/prompt-engineering-developers

➡️ Course link (with a big discount 🙂): https://www.lauromueller.com/courses/prompt-engineering

Check my other courses:

• 👉 Python for DevOps: Mastering Real-World Automation
• 👉 The Complete Docker and Kubernetes Course: From Zero to Hero
• 👉 Mastering Terraform: From Beginner to Expert
• 👉 The Definitive Helm Course: From Beginner to Master
• 👉 Mastering GitHub Actions: From Beginner to Expert
• 👉 Write better code: 20 code smells and how to get rid of them

Description

AI Toolbox is a lightweight, CLI-first Python utility for using LLMs to help with developer workflows. It provides small, focused commands and a minimal tool registry so LLMs can call local developer tools.

Highlights

• Click-based CLI with top-level commands: hello, commit, and review.
• Commit message generation: analyzes staged diffs and produces Conventional Commits-compliant messages (interactive approval / adjust flow).
• Review pipeline: syntax, logic, persona-based reviews (performance, maintainability, security) and a synthesis/refinement stage.
• Tool Registry + Tool wrappers: tool_registry and tool_utils expose tools such as run_pylint and run_security_scan that the review pipeline can use.
• Git helpers: git_utils provides get_diff and run_commit using GitPython.
• Pluggable LLM backend via litellm and configurable model with --model.

Quick install (development)

Clone and install in editable mode with dev dependencies:

git clone <repo-url>
cd ai-toolbox
python -m venv .venv
source .venv/bin/activate
python -m pip install -e .[dev]

Environment

• The project uses litellm for LLM access. Provide LLM credentials via environment variables or a .env file at the project root. The code calls dotenv.load_dotenv() in the CLI bootstrap.
• Ensure git is installed and available on PATH when running git-related commands.
• Optional tools used by tool_utils: pylint and bandit (installable via dev extras).

Usage examples

Run the CLI module directly:

# Show help
python -m ai_toolbox.main --help

# Generate a greeting
python -m ai_toolbox.main hello

# Generate a commit message (requires staged changes)
python -m ai_toolbox.main commit

# Run a lightweight review of staged changes
python -m ai_toolbox.main review

Console script

When installed (pip install . or via the editable dev install above) a console script named ai-toolbox is exposed and maps to the same Click entry point.

Development notes

• Tests: run pytest -q to execute the unit test suite in tests/.
• Lint & security: pylint and bandit wrappers are available via ai_toolbox.tool_utils.TOOL_REGISTRY and can be executed by the review pipeline.
• Packaging: pyproject.toml defines the package metadata and dependencies.

Documentation

See the docs/ folder for CLI details, per-command docs and development instructions. Start with docs/index.md.

License

MIT — see LICENSE.