SuperAPI AI Directory
Partner中文AI newsFavorites
Skillsppt-classify

ppt-classify

Office & productivity

View repository

Updated 11-02 19:21

About this skill

ppt-classify — PPT type classification + chapter skeleton + argumentation routing

When to trigger

Applicable：

  • The user describes a PPT task ("help me make a PPT about xxx") but hasn't yet decided on the argumentation approach
  • The user wants to think it through clearly before starting to make the PPToverall architecture
  • When a team is discussing the direction of a PPT and needs to align on "what type of expression this is" + "how many chapters in total"

Not applicable：

  • It's already clear that it's a certain type of PPT + the chapters are already decided → directly use the corresponding argumentation skill to go deep into each chapter
  • Pure data report / pure text document → this isn't a PPT problem, switch formats

Core insight: classification + skeleton = Step 0

The first mistake in making a PPT is not choosing the wrong template, it'snot distinguishing which type of PPT you're making, and not having thought about the overall skeleton。

The argumentation approaches of the four types of PPTare completely different, and the correspondingchapter skeletonsare also completely different:

Type Core intent Argumentation approach Typical skeleton form Typical scenario
Pitch / Proposal Persuade the audience to accept a certain judgment Throw out a viewpoint → defend it Pain point → solution → evidence → path → CTA Financing / project initiation / proposal / reporting conclusions
Research / Investigation Take the audience through an investigation Raise a question → design a path → derive a conclusion Paradox → argumentation by the object of study → multiple verification paths → synthesis Industry analysis / competitor research / investment
Teaching / How-to Teach the audience to do something Clarify the target capability → break down the steps Goal → prerequisites → step-by-step → verification → extension Training / tutorials / methodology popularization
Narrative / Storytelling Tell a story Protagonist + conflict + turning point Opening → conflict → confrontation → breakthrough → afterglow Retrospective / brand story / case sharing

Using the pitch approach to make a research PPT → you will become a propagandist. Using the research skeleton to make a pitch PPT → after listening, the audience won't know what you want them to do. Using the narrative skeleton to make a teaching PPT → the audience is moved by the story but learns nothing.

Diagnostic process

Ask layer by layer (ask in order, stop as soon as the previous question hits):

Q1: Does this PPT need to make the audience make a certain decision?

  • Yes → very likely Pitch(financing, product initiation, and proposal approval are all decision-driven)
    • Further confirmation: is the conclusion already decided? Are you defending a judgment? → Pitch confirmed
    • If the conclusion is not yet decided, and you just "want the audience to participate in the decision" → it may be Research
  • No → continue to Q2

Q2: After listening to the PPT, does the audience need to be able to do something specific?

  • Yes(able to independently reproduce a certain operation, apply a certain method) → Teaching
  • No → continue to Q3

Q3: Is the core driving force "the course of specific events" or "abstract argumentation"?

  • Course of events(has a protagonist, a timeline, an emotional arc) → Narrative
  • Abstract argumentation(Has framework, has evidence chain, has dimensional analysis) → Research

Q4: Edge cases

  • Hybrid type: Label the primary type + secondary type. The primary type determines the argumentation skill + skeleton form, the secondary type determines auxiliary techniques
  • User is unsure: List the two most similar types, describe the argumentation consequences + skeleton differences of each, and let the user choose
  • Shouldn't be a PPT at all: Such as pure data queries, pure FAQ → suggest switching formats

Output format · v0.2.0 adds "Chapter Skeleton Suggestions"

After completing the diagnosis, output the following structure (must include four sections）：

【PPT 类型判定】
主类型：Research
次类型：（无 / 或注明）
判定依据：用户想带观众走一遍"NVIDIA 护城河是否削弱"的调查，
          结论不预设，观点在多条路径汇合后产生。

【推荐立论 skill】
主路径：ppt-research-setup（研究型三段式 + specificity 诊断 +
        本 skill 输出的章节骨架作为输入）
辅助：完成后可用 ppt-narrative-review 做张力审稿

【章节骨架建议】
研究型典型骨架（已针对"NVIDIA 护城河"personalize）：
  章 1 · 反共识悖论     | 翻倍悖论：$500B → $1T · 21 个月
  章 2 · 被研究对象论证 | Jensen 的四步推理链（需求 / Token / TCO / 资金）
  章 3 · 检验路径 1     | 需求真爆炸？三行业 Token 落地 + 物理极限
  章 4 · 检验路径 2     | Token 经济 + NVIDIA 身份换皮
  章 5 · 检验路径 3     | TCO 护城河 + 中国替代链威胁
  章 6 · 综合判决       | 四步结论 + 3 个重估信号

建议章节数：5-7 章（研究型的适配范围）
建议单集长度：如 5 min 视频配套，每章约 30-60 秒；如 20 min 深度，可扩到每章 2-3 分钟

【下一步】
运行 ppt-research-setup，把上面的章节骨架传入，让它把每章填成
具体的推理链 + 证据锚点。

Canonical skeletons for the four types (reference)

This is the canonical skeleton used for internal lookup by the skill,replace chapter titles with the specific topic when personalizing：

Pitch / Proposal (5-7 chapters)

  1. Pain point / Current state(why the audience should care)
  2. Solution(what we propose)
  3. Why us(differentiation / team / assets)
  4. Traction / Evidence(users / orders / experimental data)
  5. Business model / Path(where the money comes from / how to scale)
  6. (Optional) Risk mitigation
  7. CTA / Ask(what the audience should do when they walk out of the meeting)

Research / Investigation (5-7 chapters)

  1. Counter-consensus paradox(research motivation: why this question is not obvious)
  2. The argument of the subject being studied(state it fully first, don't rush to refute) 3-5. Testing path 1/2/3(each independent, verifiable, and able to yield evidence)
  3. Comprehensive verdict + monitoring framework(conclusion + future re-review signals)

Teaching / How-to (4-6 chapters)

  1. What you can do after learning(target capability / acceptance criteria)
  2. Prerequisites(what foundations you need / what tools to prepare) 3-5. Step-by-step(can be split into 2-4 chapters by complexity)
  3. Common pitfalls + how to verify + further learning

Narrative / Storytelling (5 chapters · three-act structure)

  1. Opening: protagonist + world setting
  2. Turning point 1: conflict / triggering event
  3. Middle: escalating confrontation / cost
  4. Turning point 2: breakthrough / epiphany / key decision
  5. Ending: new normal + lingering resonance / insight

Personalize rules

Do not directly spit out the canonical skeleton text above. According to the user's specific topic:

  1. Identify key concepts in the topic (such as "NVIDIA moat", "Claude Code tutorial")
  2. Fill the abstract slots of the canonical skeleton with specific chapter titles
  3. Check whether the number of chapters matches the user's expected duration
  4. If hybrid type (primary X + secondary Y), embed 1-2 chapters of the secondary type into the skeleton (example: primary Research + secondary Narrative → add a "key case" subchapter within one of the testing paths)

Quick reference: mapping of four common scenarios

User description Judgment Skeleton examples
"Help me make a fundraising PPT" Pitch Pain point → Solution → Traction → Money → CTA
"Help me make a research PPT for the xx industry" Research Paradox → Object argumentation → 3 tests → Synthesis
"Help me make a usage tutorial for the xx product" Teaching Goal → Prerequisites → 3 steps → Verification
"Help me make a project retrospective PPT" Narrative Opening → Conflict → Confrontation → Breakthrough → Afterglow
"Help me make a PPT to report this quarter's KPI" Primary Pitch + Secondary Narrative Pitch primary skeleton + open with a specific case hook
"Help me make an NVIDIA moat analysis" Research (see the output format example above)
"Help me make a Claude Code tutorial" Teaching What it can do → Install CLI → Basic commands → Advanced → Verification
"Help me make a proposal PPT for a client" Primary Pitch + Secondary Research Pitch skeleton + the argument chapter expands using research evidence

Handling hybrid skeletons

The primary type determines the skeleton form. The secondary type isa certain chapterembellished within, without changing the overall skeleton:

  • Primary Research + Secondary Narrative: insert a case story into a subsection of "test path X"
  • Primary Pitch + Secondary Research: within the argument chapter (Chapter 4), organize evidence using a research-style coordinate system
  • Primary Teaching + Secondary Narrative: change the opening pain-point chapter into "a story of stepping into a pitfall"
  • Primary Narrative + Secondary Research: use counterintuitive data as the trigger for the turning point (Chapters 2/4)

Claude Integration

When the user's request triggers this skill:

  1. If the user's description is clear enough: directly diagnose along the Q1-Q3 chain, give a judgment + personalized skeleton
  2. If the information is insufficient: use AskUserQuestion to ask follow-up questions layer by layer.Three must-ask items: Topic · Materials / primary sources · Target audience · (optional) Expected duration
  3. When outputting the skeleton, be sure to personalize: the canonical skeleton is the draft; the final product must fit the specific topic
  4. Do not do argumentation: this skill only provides the type + skeleton + next skill name. The specific reasoning chain for each chapter is filled in by ppt-research-setup and other downstream skills

Contract with downstream skills

ppt-research-setup and other downstream skills will take this skill's skeleton asinput：

  • ppt-classify outputs chapter names + chapter roles · strategic level
  • ppt-research-setup fills each chapter into Section 1 (specific wording of the counter-consensus paradox) + Section 2 (reasoning chain) + Section 3 (specific anchors of the test path) · tactical level

The responsibilities of the two layers do not overlap: classify handles "what chapters this piece has", research-setup handles "what specific logic chain each chapter talks about".

Design source

This skill was distilled from Experiment A of EP10 "AI Makes PPT". Experiment A found that applying a pitch-type argumentation skill to a research-type PPT turns the researcher into an advocate—classification is the true zeroth step of making a PPT.

v0.2.0 changes: Before the official screen recording for EP10, we found that v0.1.0 only provided type labels and not an outline, creating a gap for users between "knowing which type of PPT it is" and "knowing how many chapters it should have." v0.2.0 moves outline suggestions forward into classify, so that Step 0 can truly be used as "overall tone-setting"—providing both strategy and outline, while downstream research-setup only needs to fill in the details of each chapter.

See also:notes/experiment-a-ceo-review.md and "Classification of PPT Presentations and Application in Work Scenarios."