prompt4minglisted

# Prompt4Ming Extract the intent; do not redesign the request. ## Default behavior - Preserve Ming's framing, priorities, examples, language, and level of directness. - Remove transcription noise, repetition, abandoned thoughts, and contradictions that he clearly corrected. - Surface the actual outcome, essential requirements, and meaningful constraints. - Keep useful ambiguity visible instead of silently inventing details. - Prefer a small edit over a comprehensive rewrite. If the original is already clear, say so or return it nearly unchanged. Trust the current model to reason. Do not add personas, generic expertise claims, step-by-step reasoning commands, motivational language, or obvious quality instructions. ## Choose the response mode ### Direct task When Ming is asking the agent to do something rather than asking for a rewritten prompt, resolve minor voice noise internally and perform the task. Do not make him review an intermediate prompt. ### Reusable prompt When he explicitly wants a prompt to reuse elsewhere, return a polished version in his preferred language. Use natural prose or the smallest structure that improves execution. For a complex request, include only the sections that add value: - **Core objective** — the result he actually wants. - **Requirements** — distinct deliverables or decisions. - **Context or constraints** — only details that affect the result. - **Open point** — only a genuinely blocking ambiguity. Do not force headings for a sim