About
This skill transforms high-level project visions into atomic, verifiable task lists designed specifically for Claude's agentic execution. It solves the problem of AI quality degradation by enforcing an 'Atomic Task Principle,' breaking projects into small phases that fit within the optimal context window. Unlike standard documentation, the resulting plans serve as direct prompts for Claude, integrating automated CLI workflows, strategic human checkpoints, and dynamic deviation rules that allow the AI to fix bugs and manage blockers autonomously while keeping the user informed.