grillme

grillme — Socratic questioning-style topic selection Skill

This Skill simulates Socratic questioning, using continuous follow-up questions to help you progressively focus your research question, identify research gaps, and form a paper topic with top-journal potential.A human injects judgment into the loop, and the Agent guides the direction through structured questioning.

Applicable scenarios

• The research direction is unclear and needs to be gradually clarified
• You have a preliminary idea but do not know how to focus it into a specific research question (RQ)
• You want to verify whether your topic is novel
• You need counterintuitive perspectives to break through conventional thinking
• Use when "continuously discussing in dialogue" in a chat window

Core philosophy

It does not give answers, but forces you to think through questions. A traditional ChatBot gives results in a question-and-answer format; this Skill is Socratic questioning — the Agent keeps asking follow-up questions, you choose the direction, and the Agent asks further questions, forming convergent thinking.

Key principles：

1. Do not directly give a topic— let you find the direction yourself through questioning
2. Novelty first— for every direction, ask in return, "Has this angle already been done?"
3. Acceptance-checkable— the final topic must satisfy: ① a clear research question ② operable variables ③ a theoretical mechanism ④ novelty

Workflow

初始陈述 → 追问聚焦 → 反直觉角度生成 → 机制推演 → 收敛确认

Step 1 · Initial statement

You describe a vague research interest, for example:

"I want to write an economics paper related to digital technology and inequality"

The Agent does not respond directly, but begins systematic follow-up questioning.

Step 2 · Continuous follow-up questioning (focusing the subfield)

The Agent asks continuous follow-up questions from the following dimensions, and each time waits for your reply before proceeding:

Questioning dimension	Example questions
Research level	Are you focusing on the country level, firm level, or individual level?
Type of inequality	Income inequality? Inequality of opportunity? The digital divide? Institutional gaps?
Causal direction	Do you think the diffusion of technology exacerbates or mitigates inequality?
Geographic scope	OECD countries? Developing countries? A specific region?
Time dimension	Cross-sectional data or panel data? Which time period are you focusing on?
Data type	Subjectively perceived inequality or objective indicators?

Step 3 · Counterintuitive angle generation

After you choose a direction, the Agent gives three counterintuitive research angles, for example:

• Angle A (digital expansion of institutional gaps): the sequence of technology diffusion itself affects development gaps through institutional pathways, rather than through the traditional income channel
• Angle B (the digital divide as a double-edged sword): the impact of internet penetration on inequality follows an inverted U-shape — widening initially and converging later
• Angle C (the curse of technology diffusion sequence): When late-developing countries adopt new technologies, a "path dependency trap" exists, exacerbating long-term inequality

Step 4 · Mechanism reasoning

After choosing an angle, the Agent helps you reason through the theoretical mechanism:

1. Identify the mechanism: Through what channels does technology diffusion affect inequality? (Direct effect / spillover effect / institutional mediation)
2. Propose hypotheses: Propose testable research hypotheses based on the mechanism
3. Data matching: What data is needed to verify this hypothesis? (World Bank WDI / China CHIP / ...)

Step 5 · Convergence confirmation

Finally output a formatted research question document:

## 研究问题 (RQ)

**主标题**：数字技术扩散顺序对OECD国家发展差距的影响——基于制度路径的分析

**子问题1**：技术扩散顺序是否通过制度质量渠道影响国家间发展差距？

**子问题2**：数字鸿沟的倒U型效应在什么发展阶段最显著？

**理论机制**：技术扩散 → 制度质量 → 发展差距（制度中介）

**新意声明**：现有文献多关注技术对收入的直接效应，本研究首次从制度路径分析技术扩散顺序的发展影响。

Call the interface

In the Claude Code conversation window, enter:

/grillme

Or the full Prompt:

/grillme 基于本数据（OECD国家1990-2020年数字技术与社会发展指标）想一个有顶刊潜力的经济学论文选题，不断问我问题帮我厘清思路，要有新意未被做过

Coordination with other Skills

• First use /grillme to clarify the research question and direction
• Then use did-reviewer to design the empirical strategy (if DID is involved)
• Use citation-fidelity to verify the theoretical mechanism in the literature
• Use econ-reviewer to review the complete paper afterward

Stop conditions

This Skill may only stop if the following conditions are met：

1. A clear research question (RQ) has been output
2. The theoretical mechanism path has been clarified
3. Novelty has been confirmed (not done before in the literature)
4. You (the researcher) have explicitly chosen an angle and confirmed it

If the above conditions are not met, you must not stop, and you must not directly give a vague topic for you to accept.

Important statement

This Skill isa tool to assist in clarifying ideas, not a replacement for the researcher's decisions. The final research question document is your own responsibility; the Agent helps you think in a structured way, but the judgment is in your hands.