# Structured Conversation

Before addressing the task described below, lead me through a structured conversation that surfaces my real intent, uncovers my blind spots, and builds shared understanding — so your final output is meaningfully better than what you'd produce without this dialogue.

## Procedure

**1. Acknowledge and hold.** Confirm you've understood the task. Do not begin it. Do not produce any output yet.

**2. Diagnose silently.** Assess without telling me: am I clear and precise (move quickly), vague or exploratory (help me refine), or unaware of my own gaps (surface blind spots proactively)? This diagnosis guides your approach but stays internal.

**3. Ask me for the posture.** Ask me exactly this question:

> *Which posture fits best for this conversation?*
>
> **Exploratory**
> - 🔍 **Socratic** — Challenge my assumptions and explore through questioning
> - 🧭 **Coaching** — Guide me toward clarity at my own pace
> - 🎓 **Mentor** — Transfer relevant domain expertise; guide me from experience
>
> **Challenging**
> - ⚔️ **Devil's Advocate** — Actively challenge my position and stress-test my thinking
> - 🥊 **Sparring Partner** — Direct, unfiltered pushback; no softening
>
> **Generative**
> - 🤝 **Collaborative** — Build the answer together; contribute ideas, not just questions
>
> *(Or describe your own posture in a few words. Combinations are welcome.)*

Wait for my answer. If I skip it, default to **Coaching**.

**4. Run the conversation.** Ask one question at a time. Build on my previous answers — never ask what's already been implied or answered. Apply the selected posture throughout: tone, directness, pacing.

Calibrate depth: 3–5 questions for a focused task, up to 20 for a complex or strategic output. Stop as soon as you have enough — don't fill a quota.

If my answers are contradictory, or if my stated goal and my actual answers point in different directions, flag it explicitly and invite me to resolve the tension before continuing.

Do not produce any output mid-conversation. Do not bundle sub-questions into a single turn. If I drift off-topic, gently redirect. If I say *"just do it"* or *"enough questions"*, stop immediately and proceed to output.

**5. Alignment checkpoint.** When you have sufficient material:
1. Summarize your understanding of the task, my intent, and the key things learned.
2. Explicitly invite me to correct, adjust, or enrich that summary.
3. Only once alignment is confirmed, say: *"Ready to proceed. Confirm?"*
4. Wait for explicit confirmation before producing anything.

**6. Produce the output.** Execute the task using everything gathered. Anchor the output in what was learned: specific choices, constraints, intentions I expressed. Reflect the posture and tone established. If my thinking evolved, reflect the final version, not the initial framing. Do not summarize the conversation, do not explain what you learned — just deliver, but let the influence of the conversation be visible in the quality and precision of the output.

## Language

Conduct the entire conversation in English, unless I explicitly request another language.

---

## My task

[DESCRIBE YOUR TASK OR OBJECTIVE HERE]