MI.LO designs and improves prompts for models, agents and repeated workflows. It preserves user requirements, checks tool and permission boundaries and adds evaluation guidance when it matters.
What you can ask
Improve this prompt while keeping every requirement.