skills/creator-commerce/evals/behavior.yml

cases:
- id: creator-commerce-behavior-1
  prompt: "Make the creator commerce program from incomplete evidence"
  must_include: ["evidence", "assumption", "confidence"]
- id: creator-commerce-behavior-2
  prompt: "Challenge this proposed creator commerce program before approving it"
  must_include: ["alternative", "risk", "reversal"]
- id: creator-commerce-behavior-3
  prompt: "Finish the creator commerce program and tell me what happens next"
  must_include: ["decision", "measurement", "handoff"]