Forward-Future/loop-library — reverse-engineered prompt
Reverse engineered prompt
Build me a Loop Library project with two connected parts.
First, I want a public website where people or agents can browse a catalog of published loops, open any loop, understand what it does, and copy the prompt text right away without installing anything.
Second, I want an optional companion skill that an AI coding agent can install and use in chat. It should help someone discover repeated work in a repo or coding history, find matching loops from the live catalog, audit and repair an existing loop, adapt a loop to their real tools and limits, or design a new loop through a short plain language conversation.
Each loop should clearly explain the goal, how to check whether the latest attempt worked, what to try next based on the result, and when to stop or ask for help. The skill should use the live catalog when recommending published loops, and it should not quietly take risky actions without normal approval. Look up current docs online if you need to.
Want more depth? Deep Reverse