2 posts tagged with "Claude Code"

We noticed something remarkable while working with Claude in our terminal. Large Language Models are fundamentally passive—they respond to prompts, provide answers, then wait dormant until the next human input. This has been the defining characteristic of every AI interaction since the beginning.

But there's an emergent capability that changes everything. AI agents can now execute sleep commands and schedule their own activity. This transforms them from passive responders into active participants that can monitor, wait, and return with updates autonomously.

AI coding is evolving fast. With Claude Code support for custom commands, it's time to upgrade your workflows. One of the most powerful advanced agentic coding techniques is parallel development with Git worktrees—running multiple Claude agents simultaneously on different branches of your codebase using custom slash commands. Our adoption of this technique is inspired by the benchy repository from this video.

Let's break it down step-by-step so you can replicate this advanced workflow in your own repo using custom Claude Code commands.