Audit and improve a prompt

Audit the prompt below before improving it.

Score 1–5 for objective clarity, context, inputs, constraints, output contract, factual safeguards, tool authority and success criteria. Explain each low score. Identify contradictions, hidden assumptions and information the model cannot know. Then produce a revised prompt that is no longer than necessary, uses explicit placeholders and distinguishes requirements from preferences. Finish with a short test set containing a normal case, edge case and failure case.

Original prompt:
[PROMPT]