Loop/Prompt Library / Loops from builders on X
🔁 Write the test questions first, then improve
Before improving an AI feature, the AI writes a set of test questions that show what 'good' means, then makes one change at a time and keeps only changes that score better.
Use it when: You want to improve an AI prompt, chatbot, or agent, but 'better' is just a feeling right now.
🛑 Stops when: The score reaches your target, or [3] changes in a row don't help, or the round limit is hit.
Copy this loop
I want to improve [AI feature, prompt, or agent]. 'Good' means: [what good answers look like]. First, write [20] test cases with what a good answer must include, mixing easy, hard, and tricky ones. Show them to me and wait for my OK. Run them to get a starting score. Then each round: make ONE change, run all the tests again, and keep the change only if the score goes up without any test getting worse. Never edit the test cases to make the score look better. Keep a log of every change and score. Stop when the score reaches [target], or 3 changes in a row don't help, or after [10] rounds. Show me the best version and the log.
Review every change the agent makes before you keep it. Set a budget or round limit — the prompt includes one.
Found on X · by @ClaudeDevs. Prompt rewritten in plain English by FixKit AI.
More loops like this
One useful thing every night
You give the AI a short mission, and each night it picks one task that moves you closer and leaves it for you to check in the morning.
Two watchers keep your agent healthy
Two separate AI helpers read your main agent's logs, find what broke, and must both agree on a fix before it is used.
Weekly skill cleanup
Once a week, the AI grades every saved skill or saved prompt it uses, removes the dead ones, and merges the repeats.
Overnight drafts that learn from results
Each night the AI drafts tomorrow's post from your content plan, and reads your notes about what worked before it writes the next one.
Split the work, keep only what passes
Splits a big job across several AI helpers working side by side, checks each one's work, and keeps only the work that passes.
Practice a website task until it's reliable
The AI tries a website chore, learns from each failed try, and keeps adjusting until it has steps that work every time.