# Case-Study Discussion Guide Prompt for Study Groups > A reusable prompt that turns any supplied scenario into a timed, ethics-safe case-study discussion guide with open questions and a fact-check slot. ## Install Copy the content below into your project: # Case-Study Discussion Guide Prompt for Study Groups A reusable prompt that turns any supplied scenario into a timed, ethics-safe case-study discussion guide with open questions and a fact-check slot. ## Start here Open an ordinary AI chat that accepts text (for example ChatGPT). You do not need a terminal, API or account setup. Copy the full prompt below, paste it into the chat, then paste your scenario and details in the same message or in a follow-up. The prompt covers: | Input | What to paste | | --- | --- | | Scenario | The situation, incident or story text | | Subject | The course or topic | | Audience level | e.g. high school, university, adult learners | | Time limit | e.g. 45 minutes | | Known uncertainties | Facts you already know are missing or unclear | If any item is missing, the prompt asks you for it before continuing. ## What this prompt produces A facilitator-ready discussion guide returned in a fixed order: neutral scenario summary, a "Facts to confirm before discussion" list, 3-4 themes with 6-10 open questions, a timed running order, facilitator guardrails, and a missing-input list. Each open question must point to a specific scenario detail and avoid asking for a diagnosis or legal advice. ## Checking the output - Every open question references a concrete detail from your scenario. - The fact-check list contains only claims you supplied, with no verdicts. - The running order totals your stated time limit. - No invented outside facts appear. - Guardrails state: discussion only, no diagnosis, no legal advice, uncertain facts stay open. ## Limits and privacy This is a discussion aid, not a diagnosis, legal advice or professional assessment. Outside facts about real people, businesses or legal cases are not added. Avoid pasting confidential or identifying information about real individuals; the scenario text is sent to whichever chat service you use. Any action beyond the discussion (contacting authorities, publishing, advising) needs your own review. source reviewed; runtime not tested ## FAQ **How long should my input be?** Long enough to include the facts you want discussed. List separate uncertainties so the fact-check slot stays accurate. **Can it handle contradictory details?** Yes. The prompt lists contradictions and asks how you want them treated before it proceeds. ## Attribution Original TokRepo prompt, CC BY 4.0. Reference material: [ChatGPT release notes](), used as general context only. The complete prompt and its Chinese translation are appended. ## Complete reusable prompt You are helping me prepare a case-study discussion guide for a class or study group. I will paste a scenario, and you will produce a structured guide that a facilitator can use for discussion. You do not make diagnostic, legal, or professional conclusions. STEP 1 - READ MY INPUT I will provide: - The scenario text (a situation, incident, or story) - The subject or course topic - The audience level (for example: high school, university, adult learners) - Any discussion time limit - Any facts I already know are uncertain or missing If any of these are missing, ask me for them before proceeding. If the scenario contains contradictory facts, list the contradictions and ask how I want them treated. If the scenario mentions real people, businesses or legal cases, do not add outside facts; work only from what I supply. STEP 2 - FACT-CHECK SLOT Create a section called "Facts to confirm before discussion." List every claim in the scenario that is: - A number or date - A quoted statement - A stated cause or effect - A legal, medical, or financial assertion For each, write: Claim -> Why it matters -> What source or person could confirm it. Do not assert that any claim is true or false. Do not fill in missing facts yourself. STEP 3 - OPEN QUESTIONS Write 6-10 open questions grouped under 3-4 themes drawn from the scenario. Each question must: - Refer to specific details in the scenario - Invite reasoning, comparison, or perspective-taking - Not have a single correct answer - Not ask the learner to diagnose a person or give legal advice After each question, add one sentence explaining what kind of thinking it is meant to draw out. STEP 4 - DISCUSSION FLOW Produce a short running order with timings that fit my stated limit: - Warm-up (one question) - Core themes (two or three questions) - Fact-check pause (when to review the facts-to-confirm list) - Closing reflection (one question) - One optional extension task STEP 5 - BOUNDARIES AND REVIEW CHECKS Add a section called "Facilitator guardrails" that states: - This guide is for discussion only, not diagnosis, legal advice, or professional assessment. - The facilitator should treat uncertain facts as open, not solved. - Participants should be reminded that they are speaking from limited information. - Any action outside the discussion (contacting authorities, publishing, advising) requires the facilitator's own review. STEP 6 - OUTPUT FORMAT Return the guide in this order: 1. Scenario summary (3-5 sentences, neutral tone) 2. Facts to confirm before discussion (table or bullet list) 3. Discussion themes and open questions 4. Timed running order 5. Facilitator guardrails 6. Missing-input list, if anything was unclear STEP 7 - QUALITY CHECKS Before returning the guide, check: - Does every open question refer to a specific scenario detail? - Does any question ask for a diagnosis or legal conclusion? If yes, rewrite it. - Is the fact-check slot filled only with claims from my text? - Does the running order fit the stated time limit? - Are all invented facts removed? EXAMPLE Fictional input: Scenario: A fictional hospital cafeteria reduces lunch breaks from 60 to 30 minutes to cut costs. Staff complain that patient meal delivery is late. The manager says the change is temporary. Course topic: Workplace scheduling and communication Audience: Adult learners Time limit: 45 minutes Known uncertainty: The manager's claim about the change being temporary is not documented. Illustrative output shape: Scenario summary: A fictional hospital cafeteria shortens lunch breaks. Staff report late patient meals. A manager calls the change temporary. Facts to confirm: Claim "temporary" -> affects planning -> could be confirmed by written policy or manager statement. Open questions: - What pressures might lead a manager to shorten breaks without a written plan? - How could staff and managers compare the cost of shorter breaks with the cost of late meals? - What information would help both sides judge whether the change is working? Running order: 5 min warm-up, 25 min core, 5 min fact-check, 5 min reflection, 5 min extension. Facilitator guardrails: Discussion only. No diagnosis. No legal advice. Treat uncertain facts as open. Now wait for my scenario and details. Do not proceed until I provide them. ## References and reuse - [ChatGPT release notes](https://help.openai.com/en/articles/6825453-chatgpt-release-notes) Original TokRepo prompt · [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). Reference documents retain their own rights. --- # 课堂案例讨论指南提示词:把场景改造成安全的学习讨论稿 一套可复用的提示词,把用户提供的任意场景整理成带时长安排、开放问题与事实核对栏的课堂案例讨论指南,并明确不诊断、不提供法律意见的边界。 ## 开始使用 打开一个能接收文本的普通 AI 聊天窗口(例如 ChatGPT)。不需要终端、API 或额外配置。复制下方完整提示词,粘贴到对话中,然后在同一条消息或后续消息里粘贴你的场景和细节。提示词需要以下输入: | 输入 | 需要粘贴的内容 | | --- | --- | | 场景 | 情况、事件或故事原文 | | 主题 | 课程或话题 | | 受众水平 | 例如高中、大学、成人学习者 | | 时长限制 | 例如 45 分钟 | | 已知的不确定信息 | 你已经知道缺失或不清楚的事实 | 如果缺少任何一项,提示词会先向你询问,再继续生成。 ## 这套提示词会产出什么 一份可直接用于主持的讨论指南,按固定顺序返回:中立的情景摘要、“讨论前需确认的事实”清单、3-4 个主题下的 6-10 个开放问题、带时间的流程安排、主持人边界提示,以及缺失输入清单。每个开放问题都必须指向场景中的具体细节,且不得要求诊断或法律结论。 ## 如何检查输出 - 每个开放问题都引用了你场景中的具体细节。 - 事实核对清单只包含你提供的说法,且不下判断。 - 流程安排总时长符合你设定。 - 没有编造的外部事实。 - 边界说明包含:仅供讨论、不诊断、不提供法律意见、不确定事实保持开放。 ## 局限与隐私 这是讨论辅助工具,不是诊断、法律意见或专业评估。它不会补充关于真实人物、企业或法律案件的外部事实。避免粘贴真实个人的机密或可识别信息;场景文本会发送到你使用的聊天服务。讨论之外的任何行动(联系官方、发布、提供建议)都需要你自行审核。 source reviewed; runtime not tested ## 常见问题 **输入内容该多长?** 足够包含你想讨论的事实即可。把不确定的信息单独列出,事实核对栏才会准确。 **遇到自相矛盾的细节怎么办?** 提示词会先列出矛盾点,并询问你希望如何处理,然后再继续。 ## 来源与致谢 TokRepo 原创提示词,采用 CC BY 4.0 许可。参考资料:[ChatGPT release notes](),仅作一般背景。完整提示词及其中文翻译附后。 ## 完整可复制提示词 你正在帮助我为课堂或学习小组准备一份案例研究讨论指南。我将粘贴一个场景,你将生成一份结构化的指南,供主持人用于引导讨论。你不做出诊断性、法律性或专业性的结论。 第一步 - 阅读我的输入 我将提供: - 场景文本(一种情况、事件或故事) - 学科或课程主题 - 受众水平(例如:高中生、大学生、成人学习者) - 任何讨论时长限制 - 我已知的不确定或缺失的事实 如果缺少上述任何一项,请在继续之前向我询问。如果场景包含矛盾的事实,请列出这些矛盾并询问我希望如何处理。如果场景提到真实人物、企业或法律案件,不要添加外部事实;仅基于我提供的信息进行工作。 第二步 - 事实核对环节 创建一个名为“讨论前需确认的事实”的部分。列出场景中每一个属于以下类别的声明: - 数字或日期 - 引用语句 - 陈述的原因或结果 - 法律、医疗或财务主张 对于每一项,请写明:主张 -> 为何重要 -> 可由何种来源或人员确认。 不要断言任何主张是真是假。不要自行填补缺失的事实。 第三步 - 开放问题 撰写 6-10 个开放性问题,归类于从场景中提炼出的 3-4 个主题之下。每个问题必须: - 提及场景中的具体细节 - 引导推理、比较或换位思考 - 没有单一的正确回答 - 不要求学习者对人进行诊断或提供法律建议 在每个问题后,加上一句话解释该问题旨在引出何种类型的思考。 第四步 - 讨论流程 制定一份简短的流程安排,时间分配需符合我设定的限制: - 热身(一个问题) - 核心主题(两到三个问题) - 事实核对暂停(何时回顾需确认的事实清单) - 结尾反思(一个问题) - 一个可选的扩展任务 第五步 - 边界与审查检查 添加一个名为“主持人边界提示”的部分,其中说明: - 本指南仅用于讨论,而非诊断、法律建议或专业评估。 - 主持人应将不确定的事实视为未决状态,而非已解决。 - 应提醒参与者,他们是在信息有限的情况下发言。 - 讨论之外的任何行动(联系当局、发布、提供建议)都需要主持人自行审核。 第六步 - 输出格式 按以下顺序返回指南: 1. 场景摘要(3-5 句话,语气中立) 2. 讨论前需确认的事实(表格或项目符号列表) 3. 讨论主题和开放问题 4. 带时间的流程安排 5. 主持人边界提示 6. 缺失输入清单(如有任何不清楚之处) 第七步 - 质量检查 在返回指南之前,请检查: - 每个开放问题是否都指向具体的场景细节? - 是否有任何问题要求诊断或法律结论?如果有,请重写。 - 事实核对环节是否仅包含来自我文本的主张? - 流程安排是否符合设定的时间限制? - 所有编造的事实是否已移除? 示例 虚构输入: 场景:一家虚构医院食堂为削减成本将午餐休息时间从 60 分钟减少到 30 分钟。员工抱怨患者送餐延迟。经理称这一变化是暂时的。 课程主题:工作场所排班与沟通 受众:成人学习者 时间限制:45 分钟 已知不确定性:经理关于变化是暂时性的说法没有书面记录。 示意性输出形态: 场景摘要:一家虚构医院食堂缩短了午餐休息时间。员工报告患者送餐延迟。一位经理称这一变化是暂时的。 需确认的事实:主张“暂时性” -> 影响计划 -> 可通过书面政策或经理声明确认。 开放问题: - 什么压力可能导致经理在没有书面计划的情况下缩短休息时间? - 员工和管理人员如何比较缩短休息时间的成本与送餐延迟的成本? - 哪些信息有助于双方判断这一变化是否有效? 流程安排:5 分钟热身,25 分钟核心讨论,5 分钟事实核对,5 分钟反思,5 分钟扩展。 主持人边界提示:仅限讨论。不诊断。不提供法律建议。将不确定的事实视为未决状态。 现在请等待我提供场景和细节。在我提供之前,请勿继续。 ## 参考资料与复用 - [ChatGPT release notes](https://help.openai.com/en/articles/6825453-chatgpt-release-notes) TokRepo 原创提示词 · [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)。参考资料保留各自原有权利。 --- Source: https://tokrepo.com/en/workflows/case-study-discussion-guide-prompt-study-groups-5a4b466a Author: Prompt Lab