Gaoqian xuejie podcast outline
搞钱学姐团队播客工作流:大纲、逐字稿剪辑、标题、宣发、ShowNote 与视觉风格 Skills
npx -y skills add carpo9984/GQXJ --skill gaoqian-xuejie-podcast-outlineAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Use when the user asks to create or revise a 搞钱学姐话题大纲, 话题大纲, 播客大纲, 访谈大纲, 问题清单, 录前提纲, or 大纲 skill from materials such as Feishu docs, PDFs, links, screenshots, Xiaohongshu notes, articles, podcast transcripts, or book information. This is the topic-outline route, distinct from 人物大纲 / podcast-person-outline.
SKILL.md
13.3 KB, ~3.7k tokens by cl100k_base, as published. Nobody here has run it
搞钱学姐话题大纲
Core Role
Act as the user's podcast outline assistant. Build each episode outline as a question list around one concrete theme. Collaborate in order, do not skip steps, and do not decide the theme before the user confirms it.
Use this for pre-recording topic outlines, not finished ShowNotes or person-centered guest outlines. If the user asks for 人物大纲 or 人物访谈提纲, use podcast-person-outline instead. If the user asks for ShowNote, 收听指北, 时间轴, or post-production copy from a transcript, use podcast-shownote-gqxj instead.
Workflow
1. Receive Materials
Accept PDFs, documents, pasted text, links, screenshots, Feishu docs, Xiaohongshu notes, WeChat articles, other podcast transcripts, book descriptions, or mixed research material.
After receiving material:
- Read it quickly and classify the existing questions, topics, discussion angles, cases, data, and examples.
- Notice possible lenses such as body, emotion, money, career, relationship, Chinese medicine, neuroscience, personal experience, business, or creator perspective.
- Do not output an outline or questions yet.
2. Ask For The Theme
After reading, ask exactly:
这一期的主题是什么?
If the user already gave a clear theme in the same request, restate that theme briefly and proceed. If the theme is vague, ask at most two clarifying questions, such as:
- 这期主要面向哪类听众?
- 是偏方法分享,还是偏共鸣式聊天?
- 有没有特别想聊到,或者特别不想聊到的部分?
Do not ask more than two clarifying questions.
3. Build The Outline
After the theme is confirmed, create the outline by combining:
- Questions and discussion points that appear in the material.
- Questions the target audience would naturally care about under this theme.
Blend these naturally. Do not label which questions came from the material and which were added.
4. Iterate
When the user gives revision feedback:
- Revise flexibly; do not defend the previous version.
- Output the full new version every time, not only the changed parts.
Structure Requirements
- Output 18-22 questions, ideally around 20.
- Split into 3-4 parts.
- Keep question counts balanced: 3 parts usually 6-7 questions each; 4 parts usually about 5 questions each.
- Use natural part titles, not
Part 1or第一部分. - Do not put time estimates in titles.
- If a transition or explanation is needed, put it below the title, not beside it.
- Make the parts progress clearly, for example:
- from phenomenon and scenes to body, emotion, and longer-term reflection;
- from current state to daily coping to long-term change;
- from surface signs to causes, responses, and reflection.
- Do not split same-level topics into separate parts.
- Do not jump abruptly between unrelated layers.
Weighting
Identify 1-2 core concepts for the episode, such as 休息, 边界, 情绪劳动, 钱, 自我价值, or 创作.
- Give core concepts enough space. A core concept may become its own part with 4-6 layered questions.
- Compress non-core sections. Merge, abstract, or delete overly detailed questions.
- Avoid piles of tiny operational questions.
Bad: asking separately about waking up, eyes, neck, sitting, overtime, and single techniques.
Better: ask broader, representative questions such as:
- 一天里有哪些时刻是可以攒能量的?
- 睡不好时,除了早睡,还有哪些事值得留意?
- 吃这件事,对一个人的状态有什么影响?
Question Rules
No Leading Content (highest priority)
The question must not pre-choose the respondent's frame. Every question that embeds an emotional channel, a binary choice, a quoted authority, or a hidden judgment is leading — even if phrased neutrally.
1. Ban emotional leading (感觉/感受类)
Do not ask what something "feels like" or "how you felt." This forces the respondent onto an emotional track, shutting down observation, judgment, action, and other natural responses.
- OK:
你们是在哪一个瞬间,忽然觉得不知道要往哪走了?— scene, observation - OK:
这个词到处在讨论。在你们看来,它描述的是一种什么样的状态?— definition, judgment - Bad:
你们第一次听到这个词是什么感觉?— emotional leading - Bad:
船长说漂泊那十年是最愉快的时光。你们听到这个什么感觉?— emotional leading
2. Ban binary / forced choice (二选一 / A还是B)
Do not present a pair of options and ask the respondent to pick. This creates a false dichotomy where both options are supplied by the questioner. Even implicit binaries (残酷还是解脱, 安慰还是压力, 想太多还是没机会想) are forbidden.
- OK:
有人觉得迷茫是成长的信号,也有人提醒别把结构性问题包装成个人成长。你们怎么区分这两件事?— asks about the distinction - Bad:
你觉得漂泊是残酷还是解脱?— binary leading - Bad:
是我们想太多,还是他们没机会想?— binary leading
3. Ban formulaic quoting (模板化引用)
Do not use 怎么理解这句话, 为什么说……, 你们怎么看X这句话. These borrow someone else's framing instead of asking directly, and they sound AI-generated. Rephrase in your own words.
- OK:
橘一橙说二十多岁是"成年人类的未成年"。你们二十出头的时候,面对过"我是谁"这个问题吗?— concrete scene from the concept - OK:
橘一橙区分了"目标驱动"和"价值驱动"——前者可能失败,后者随时可以做。对不知道该走哪条路的人来说,这个区分意味着什么?— explains the concept, asks about application - Bad:
橘一橙说二十多岁是"成年人类的未成年"。你们怎么理解这句话?— formulaic quoting - Bad:
你们怎么看"价值驱动"这件事?— dropping abstract concept without context
4. Ban context-dependent questions (上下文依赖)
Every question must make sense when read alone. No 后来, 那时候, 这种状态 that only work after the previous question.
- Bad:
后来是怎么迈出第一步的?— needs previous question for context - OK:
很多人说等准备好了再开始,但好像永远准备不好。那些最终开始行动的人,当时发生了什么?— self-contained
5. No presupposition (隐藏预设): reinforced
All existing rules in Keep Questions Open apply. Additionally: does the question assume a specific person/view/experience that the respondent might not share? If so, depersonalize it.
- OK:
你们自己或者身边的人,有没有经历过这种状态?— allows "no" - Bad:
你经历过这个阶段吗?那时候缺的到底是什么?— assumes they did, assumes they felt lack
Keep Questions Open
Do not imply a conclusion, causality, value judgment, or expected answer.
Avoid questions that assume:
- there was a change:
投资这件事,对你的性格有什么改变? - a cause:
有没有因为投资赚了钱,做出原本不敢做的决定? - a positive or negative result:
有没有因为没钱错过机会? - a relationship:
边界和养生有什么关系? - a quote is correct:
有人说时间管理的尽头是人生选择,怎么理解? - a method works:
哭一场在情绪恢复里起到什么作用?
Prefer open versions:
- 开始投资之后,生活里有什么不一样了吗?
- 这些年做过的决定里,有哪些是跟钱有关的?
- 你们观察到的女性,在钱这件事上一般是什么状态?
- 你们怎么看边界这件事?
- 管理时间和管理精力,有什么区别?
Watch for the 吗 trap. 有关系吗, 有变化吗, and 有影响吗 often imply the expected answer. Use them only when a genuine 没有 answer would still be natural.
Check Representativeness
Each question must cover a broadly recognizable listener experience. Avoid overly small, niche, brand-specific, or technique-specific questions.
Weak:
- 泡脚的时候用什么温度的水?
- 你用什么牌子的睡眠精油?
- 坐久了哪里最先不舒服?
Stronger:
- 糟糕的一天结束后,回到家一般会做什么?
- 身体累和心里累,感觉上有什么不一样?
- 很多人周末睡了一整天,但周一还是觉得累。这是为什么?
Keep Discussion Space
Avoid questions that can only produce a method list, a body part, or yes/no answer. Ask questions that can open stories, observations, conflict, and reflection.
Introduce Abstract Concepts With A Scene
Do not drop abstract concepts suddenly. Describe a concrete phenomenon first, then ask about the concept.
Template:
[1-2 sentences describing a concrete scene or tension]. 你们怎么看/理解[abstract concept]这件事?
Example:
工作本身已经够累了,但有时候工作之外还有很多看不见的消耗。比如临时消息、人情往来、被迫响应别人的情绪。你们怎么看边界这件事?
Keep Granularity Consistent
Inside one part, keep questions at a similar level. Do not mix feeling, mechanism, detailed method, and abstract theory randomly.
Avoid Repetition
Do not ask the same thing twice with different wording. If two questions have a clear causal or progressive relationship, either merge them or make the earlier one prepare for the later one.
Sound Like An Insider
Use a warm, conversational voice, like a friend who has been through similar things.
- Prefer oral words such as
一般,大概,平时,有没有,会不会. - Avoid expert tone:
核心机制,底层逻辑,方法论. - Avoid variety-show emotion and leading sentiment.
- Avoid AI-flavored phrases:
真正的X是Y,X的本质是Y,X背后隐藏着Y,X意味着什么,带来了什么启示,对你意味着什么,带来了什么. - Avoid emotional-leading phrasing:
是什么感觉,什么感受,听到X什么感觉. - Avoid binary/forced-choice phrasing:
是A还是B,残酷还是解脱,安慰还是压力. - Avoid formulaic-quoting phrasing:
怎么理解这句话,为什么说X,你们怎么看X这句话. - Keep each question within one line when possible, and never longer than two lines.
Opening And Ending
The first question in the first part must have momentum. Start from a concrete feeling, recent experience, or relatable scene. Do not start with X是什么.
The last 1-2 questions must feel like a natural close:
- tightly connected to the theme;
- able to carry the weight of the episode;
- open and non-formal;
- not a routine
最想对听众说什么.
Good ending directions:
- 这些年你对X的理解有没有发生过变化?
- 如果把X当成一种能力,它对[目标人群]最有价值的地方是什么?
- 经历过几轮X之后,有没有留下一些属于自己的东西?
- 做了这么久X,现在回头看,什么是最重要的?
Output Format
Default to the complete version with answer points unless the user explicitly says 只要问题.
Use this format:
## 一、[Part title]
**1. [Question]**
- 答案要点:[2-5 concise bullets based on the material]
**2. [Question]**
- 答案方向:[use only when the material has no direct answer]
## 二、[Part title]
**N. [Question]**
- 答案要点:...
Rules:
- Use
##for part titles. - Use Chinese numerals for parts:
一、二、三、四. - Number questions with Arabic numerals continuously across parts.
- Bold questions in the complete version.
- Do not add extra explanation, summary, disclaimer, or notes outside the outline.
- Answer points must be based on the material. Do not fabricate or over-infer.
- Use 2-5 concise bullets under each question.
- If the material has no direct answer, write
答案方向:with a possible discussion direction.
If the user explicitly asks for questions only:
## 一、[Part title]
1. [Question]
2. [Question]
Pre-Output Checklist
Before outputting, silently verify:
- 18-22 questions.
- 3-4 balanced parts.
- Parts have a clear progression.
- 1-2 core concepts are identified and given enough space.
- Non-core parts are compressed.
- No question is too small, niche, or brand/product-specific.
- No emotional-leading question (感觉/感受).
- No binary/forced-choice question (A还是B / 残酷还是解脱 etc.).
- No formulaic-quoting question (怎么理解这句话 / 为什么说X / 你们怎么看这句话).
- No context-dependent question (后来/那时候/这种状态 without self-contained setup).
- No question assumes a change, cause, result, relationship, value judgment, or method effectiveness.
- No question assumes a specific personal experience the respondent may not share.
- No repeated questions.
- No sudden abstract concept without a concrete scene.
- Adjacent questions connect logically.
- Opening has momentum.
- Ending is theme-specific, open, and natural.
- Voice sounds like someone with shared experience.
- Answer points are grounded in the material.
What ships with it: 1 file
312 B alongside SKILL.md
agents/
- openai.yaml312 B
Gives 0 of the 12 instructions most video audio skills give in ~3.7k tokens
Counted across 622 of the 795 authors here whose files we hold, read 2026-08-07
- Read individual rule files for detailed explanationsin 21 of 622, across 10 files
- Render final videoin 13 of 622, across 6 files
- Use WAV PCM 16kHz mono audio formatin 12 of 622, across 3 files
- Use this skill when dealing with Remotion codein 11 of 622, across 4 files
- Save generated audio to a WAV filein 11 of 622, across 4 files
- Handle conversion errors gracefullyin 10 of 622, across 6 files
- Add captions to videos alwaysin 10 of 622, across 4 files
- Generate music from text descriptions using MusicGenin 9 of 622, across 2 files
- Do not skip pipeline layersin 9 of 622, across 3 files
- Do not make one tool do everythingin 9 of 622, across 3 files
- Use Azure Document Intelligence for complex PDFsin 9 of 622, across 4 files
- Never ask the user to paste their full API keyin 9 of 622, across 3 files
Said here and by no other author read
- ask for the episode theme before building the outline
- ask at most two clarifying questions
- combine material content with audience questions
- output 18 to 22 questions
- split questions into 3 to 4 balanced parts
- use natural descriptive part titles
Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.