Open any chat tool and paste: 'Role Writer: draft a 3-bullet onboarding tip for AI newbies. Role Critic: score 1-5 on clarity and demand a rewrite if under 4. Run both.' You're done when you see a score and, if needed, a rewrite (or a pass at 4 or above).
How-to · WFL · WFL.2
X-style: evaluator-optimizer loop in plain English
Evaluator-optimizer: score a draft against criteria, then rewrite once. That is a mini quality loop.
Worked example
Straight from the agent-pattern threads: one model writes, another grades it against a rubric, and it loops until it passes. You can recreate this yourself in a single chat, just by giving it two roles.