tools

Miscellaneous Python tools to automate everyday tasks and experiment with ideas.

Inspired by this quote from Simon Willison:

"This is why I care so much about the productivity boost I get from LLMs so much: it’s not about getting work done faster, it’s about being able to ship projects that I wouldn’t have been able to justify spending time on at all."

Why Python is the right tool for the job?

• Python is a nice scripting language that is easy to learn and read.
• It is a powerful language with a huge ecosystem of libraries for all kinds of tasks.
• Lastly, uv has made Python a breeze to work with.

Install

This project uses uv for running scripts.

Usage

It is recommended to use an alias for running scripts. This is an example for Fish from my dotfiles.

Available Tools

fingerprint

Manage fingerprint authentication on Arch Linux.

Commands:

• fingerprint init - Setup PAM configuration for fingerprint authentication (sudo/polkit)
• fingerprint add - Add a new fingerprint for the current user
• fingerprint remove - Remove all fingerprints for the current user
  • --pam - Also remove PAM configuration

Requirements:

• fprintd package must be installed: paru -S fprintd

Example:

# Initial setup
uv run --script fingerprint.py init

# Add a fingerprint
uv run --script fingerprint.py add

# Remove fingerprints (keeping PAM config)
uv run --script fingerprint.py remove

# Remove fingerprints and PAM config
uv run --script fingerprint.py remove --pam

openapi_server

FastAPI test server with diverse endpoints for OpenAPI agent experimentation.

Features:

• RESTful API with users, items, orders, and search endpoints
• Full CRUD operations for resources
• OpenAPI 3.1 specification at /openapi.json
• Interactive docs at /docs

Example:

# Start the server
uv run --script openapi_server.py

# Server runs on http://localhost:8000
# OpenAPI spec: http://localhost:8000/openapi.json
# Interactive docs: http://localhost:8000/docs

openapi_agent

AI agent that can interact with APIs using OpenAPI specifications. Creates dynamic tools from OpenAPI specs and uses Pydantic AI with Mistral to execute API calls.

Commands:

• openapi_agent run <openapi_url> --prompt "<your_prompt>" - Execute a prompt against an API
  • --devstral - Use devstral model instead of default Mistral
• openapi_agent test - Test with local server

Requirements:

• MISTRAL_API_KEY environment variable for Mistral model (optional - uses test model if not set)

Example:

# Test with local server (make sure openapi_server.py is running)
uv run --script openapi_agent.py run localhost:8000/openapi.json --prompt "Get all users and show their names and emails"

# Create a new user and show results
uv run --script openapi_agent.py run localhost:8000/openapi.json --prompt "Create a user named 'John Doe' with email '[email protected]' and show all users"

# Search for items
uv run --script openapi_agent.py run localhost:8000/openapi.json --prompt "Show me all items in the Electronics category"

# Use devstral model
uv run --script openapi_agent.py run localhost:8000/openapi.json --prompt "Get system statistics" --devstral

Environment Variables

General

• Various tools may require specific environment variables - check individual tool documentation.

OpenAPI Agent

• MISTRAL_API_KEY - API key for Mistral AI model (optional - falls back to test model if not set)

Example .env file

# Mistral API key for OpenAPI agent
MISTRAL_API_KEY=your-mistral-api-key-here