The modern tool builder: CLIs, agents, and PEP 723
Being able to write Python scripts is an incredibly useful skill, allowing you to automate nearly any task imaginable. However, with a bit of extra effort, those brittle scripts can be transformed into robust, reusable tools. In particular, tools that you can run from the command line or that can be run by coding agents. Future-you will thank you for making the extra effort.
script (noun, Computing) A sequence of instructions, often single-purpose and context-dependent, that is interpreted by another program.
tool (noun, Computing) A reusable and flexible piece of software designed to solve problems across different contexts or aid in accomplishing diverse tasks.
Being able to write Python scripts is an incredibly useful skill, allowing you to automate nearly any task imaginable. However, with a bit of extra effort, those brittle scripts can be transformed into robust, reusable tools. In particular, tools that you can run from the command line or that can be run by coding agents. Future-you will thank you for making the extra effort.
In this talk, we will explore why you should graduate from writing scripts to building tools, particularly command-line interfaces (CLIs).
CLIs have been around for over 50 years. Thanks to the rise of coding agents, being able to understand and use the command line is now more important than ever. I'd say there are three reasons for this. First, coding agents rely on CLI tools such as find, grep, and git (and maybe yours!). Second, some coding agents, such as Claude Code, are themselves CLI tools that you can incorporate into your scripts. Third, agents are actually really good at helping you to build your own tools, so if you use them there's really no excuse anymore.
The talk first explains why you should consider turning your scripts into flexible, reusable tools. Then, it covers the steps needed in order to do so. Outline:
- Introduction: The CLI Renaissance (5 mins)
- The Unix Philosophy is the AI Tool Philosophy (5 mins)
- Designing for Humans and Agents (10 mins)
- Zero-Setup Execution with PEP 723 (5 mins)
- Polishing the Interface (5 mins)
Jeroen Janssens, PhD, is Head of Developer Relations at Posit, PBC. His expertise lies in visualizing data, implementing machine learning models, and building solutions using Python, R, JavaScript, and Bash. He’s passionate about open source and sharing knowledge. He’s the author of Python Polars: The Definitive Guide (O’Reilly, 2025) and Data Science at the Command Line (O’Reilly, 2021). Jeroen holds a PhD in machine learning from Tilburg University and an MSc in artificial intelligence from Maastricht University. He lives with his wife and two kids in Rotterdam, the Netherlands.