Pi is a minimal terminal coding agent. This article explains what it is, how it differs from Claude Code, Codex, OpenCode, and Gemini CLI, and why it went viral.

Sixth installment in the series on software engineering with AI agents. Today we look at the project that GitHub has turned into the de facto standard for specification-driven development: Spec Kit.

So far we have covered methodologies for directing agents: compound engineering (Every), discipline skills (Addy Osmani), a complete operating …

Fifth installment of the series on software engineering with agents. Today it is the project that is taking off right now: Matt Pocock's skills.

I have covered Superpowers (Jesse Vincent), Agent Skills (Addy Osmani), and Compound Engineering (Every). Each brings an important piece of the puzzle. But Matt Pocock's skills …

Fourth installment in the series on agent-driven software engineering. Today I am covering a project that has changed the game: Superpowers, by Jesse Vincent.

With 180,000 stars on GitHub and 16,000 forks, Superpowers is the most popular agent skills framework today. But what makes it special is not …

In the previous two parts of this series we have seen an overview of agent-driven engineering and then Every's compound engineering methodology. Now for another similar approach: Agent Skills, Addy Osmani's project at Google that has nearly 30,000 stars on GitHub and tackles the problem from a completely different …

In the first part of this series I talked about how AI agents have changed the way I build software. But that was just the warm-up. What follows is the vision of a company that has taken this idea to the extreme: Every, the startup from Dan Shipper and Kieran …

After more than 20 years of writing code, the rise of AI agents has changed how I work more than any new framework or language ever could. It is not about the AI writing code for me, but about letting me stay focused on what actually matters: design, architecture, and …

Crit is a tool designed to close the code review loop for AI-generated code, providing an intuitive and efficient web interface.

agent-browser is a tool from Vercel that allows AI agents to navigate the web and perform actions just like a human would, easily integrating with agent frameworks.