For months I used my coding agent the same way most people do: write a prompt, read the reply, write the next prompt. I was the loop. I was the thing deciding what came next and noticing when it drifted off course.
Loop engineering is the fix, and I turned it into a Claude Code skill you can run in one command.
Repo: github.com/sxivansx/loop-engineering
What loop engineering actually is
Loop engineering is designing the system that prompts the agent for you, instead of prompting it yourself, turn by turn. You define a goal and a stop condition once. The loop then finds the work, does it, verifies it with a separate agent, writes down what happened, and runs again until the goal is met.
