Prompt Golf

How this is graded

A real model reads your brief and judges whether an assistant could follow it, judging the meaning rather than the keywords. It cannot be satisfied by dropping in magic words, and it cannot be fooled by rewording the same vagueness.

• · Says who it is for
• · Sets the format, length, or shape
• · Names the concrete subject matter
• · Two different assistants would produce the same thing

The model provider is offline — verdicts fall back to the provisional keyword check until it is back.

Scoring

• · Score = (8 × failed swings) + word count. Lower is better.
• · The score is fixed arithmetic — brevity and fewer swings.
• · The verdict comes from the model: would this actually work?
• · Finish every drawn hole at or under par for bonus XP.

Scorecard