loopy

Why we included this project

Most prompt libraries hand you one-shot instructions that stop after the first attempt. Loopy is built on the opposite idea: a loop gives an agent a way to learn from the result and take the next useful step, measuring each attempt, keeping what works, and repeating until it reaches a sensible stopping point instead of guessing. Teams that already run coding agents like Codex, Cursor, or Claude Code get ready-made patterns for the jobs where a first pass rarely is the final answer, such as fixing production errors, raising test coverage, reviewing a product, or keeping documentation current. The companion Loopy skill turns the catalog into an interactive toolkit: an agent can browse the live catalog, audit a loop for weak checks, adapt one to different tools, and craft a new loop through a short conversation, while agents that skip the skill can still read the catalog through its published guides. If you keep writing throwaway prompts for the same kind of work, this is a practical place to start.