演讲稿生成PPT

SKILL效率
效率 未标注权限 许可证不明确
安装这个技能

将演讲稿自动生成PPT,可以自定义PPT风格以及页数

一句话选型结论

按资料字段整理

演讲稿生成PPT是效率类的 Agent Skill,安装方式未标注,权限级别未标注,许可证不明确,有可直接复制的安装命令。技能本质是给模型读的说明书,效果取决于模型与客户端;装完先用一个小任务验证,再放进正式流程。

分类效率 安装方式未标注 权限级别未标注权限 许可证许可证不明确
累计安装 未标注 本周 未标注
GitHub 星标 未标注 仓库 未标注
下载量 未标注 技能市场下载数
权限级别 未标注权限 安装前请确认风险范围
许可证 许可证不明确 未说明

安装

未标注
在支持 Skills 的客户端里执行,装完问一句「你有哪些技能」验证
skillhub install user_db22db11/speech-to-ppt
  1. 1
    确认客户端支持

    确认你的客户端支持 Skills(Claude、Codex、Cursor 等),并升级到较新的版本。

  2. 2
    拿到技能文件

    执行上面的命令安装;也可以把仓库里的 SKILL.md 手动放进客户端的 skills 目录。

  3. 3
    刷新技能列表

    重新打开客户端,或刷新技能列表,让新放进来的技能被加载。

  4. 4
    验证是否生效

    问一句「你有哪些技能」,确认列表里出现它,再拿一个小任务试跑。

技能档案

效率
字段来自技能仓库与公开索引,中文说明由本站标注
作者 / 维护方user_db22db11仓库 未标注 · 版本 未标注
宿主客户端未标注技能可被这些客户端加载
安装方式未标注安装包 community
权限级别未标注权限安装前请确认风险范围
许可证未说明许可证不明确,商用前请自行确认条款
仓库可用性available技能仓库当前是否可访问
最近核验未标注上游最后核验时间 · 数据更新 1791732574
正文体量8.7 KBSKILL.md 原文大小,越大说明说明书写得越细

典型使用场景

3 类
按分类整理的常见用法,具体以 SKILL.md 为准
整理待办与日程

把零散任务收进清单,排出优先级。

批量处理重复任务

把固定动作做成模板,重复执行。

把流程写成清单

把口头流程整理成可复用的步骤。

选择建议

装之前先看这几条
按条目字段给出的判断,不替你决定用不用

值得选的地方

  • 有可直接复制的安装命令,不用自己翻文档找入口。

先想清楚的地方

  • 没有标注权限级别,实际能做什么要装好后确认。
  • 许可证不明确,商用或二次分发前先到仓库确认条款。
  • 没有做站内核验,内容以作者仓库为准。
  • 技能只是说明书,实际效果取决于模型与客户端,装完先用小任务试一遍。

更适合

  • 需要在效率场景里给模型补一套做法的人。
  • 愿意先读一遍 SKILL.md、确认它会让模型做什么的人。

不太适合

  • 对许可证要求很严、必须明确可商用的人。
  • 只想装完就自动变强、不打算看说明书的人。

同类对比

6 个同类
同分类里挑最接近的几条,字段以各自详情页为准
技能权限级别安装方式累计安装宿主客户端安装命令
演讲稿生成PPT当前 未标注权限 未标注 未标注 未标注 有
Nano PDF 文档 联网访问 ClawHub 市场 424.9 万 OpenClaw 有
自动更新程序 浏览器控制 ClawHub 市场 296.4 万 OpenClaw 有
Gemini 助手 联网访问 ClawHub 市场 254.7 万 OpenClaw 有
word DOCX 文档 联网访问 ClawHub 市场 239.6 万 OpenClaw 有
Excel XLSX 前端 联网访问 ClawHub 市场 227.0 万 OpenClaw 有

SKILL.md 全文

8.7 KB
模型实际读到的说明书原文,来自技能仓库

Speech → PPT

Inputs to collect

Collect before designing the deck. Ask only what is missing — do not re-ask what the user already gave.

Input Required? Default if missing
Speech transcript / speaker notes required —
Page count N (hard constraint) required 10
Style preset (one of 7) required 商务
Deck title optional infer from the speech opening
Audience optional 通用
Language optional follow the transcript's language
Output filename optional speech-deck-{slug}.pptx

Style presets (full visual mapping in references/style-presets.md):
商务 · 极简 · 学术 · 演讲型 · 创意 · 科技 · 金融

If the user says "看着办" / "随便" / doesn't pick, ask once with a one-line menu — do not silently default to 商务.

Procedure

Step 1 — Analyze the speech

Read the transcript and extract:

  • Core thesis: one sentence (the single thing the audience should remember)
  • Major sections: 3-5 logical blocks (open / body / close)
  • Key points per section: 2-4 supporting ideas, each ≤ 18 words to stay slide-friendly
  • Strong lines: any quote, stat, or hook worth its own slide
  • Natural transitions: phrases like "接下来", "其次", "最后" that mark section boundaries

If the transcript has no clear structure (e.g. casual Q&A, < 300 words), say so before designing — the user may want to re-paste a more structured version.

Step 2 — Design the page allocation (HARD N pages)

The total slide count must equal N. No over-generation, no under-generation.

Default allocation when N ≥ 5:

Page Type Purpose
1 Cover Title + subtitle + (optional) presenter + date
2 TOC 3-5 section names, derived from the major sections
3 ... N-1 Content One core idea per slide, with at most 1 strong line as a quote/highlight slide
N Closing Key takeaways (3-5) or thank-you + contact

Rules:

  • N < 3: stop and ask the user to pick a larger count. Two-page decks (cover + close) skip too much.
  • N = 3-4: merge TOC into the cover, drop section dividers.
  • N ≥ 15: insert section dividers (one per major section) between the TOC and closing pages. If dividers eat the budget, compress content slides; never drop the cover, TOC, or closing.
  • Long speech, small N: compress by merging adjacent points on one content slide, or promoting a key line into the cover. Never invent content that isn't in the transcript. If essential material is at risk of being lost, surface this to the user before writing the outline (see Failure handling).
  • Short speech, large N: insert section divider pages or data callout slides that visualize the speech's own numbers/quotes (e.g. "3 key stats from the talk"). Do not pad with unrelated stock content.

Step 3 — Write the outline (FIRST DELIVERABLE)

Output the outline as a numbered list, one entry per slide. Each entry has:

[Slide N]  &lt;Type&gt;  &lt;Title&gt;
  - &lt;Bullet 1&gt;
  - &lt;Bullet 2&gt;
  - &lt;optional: visual hint, e.g. &quot;big stat: 73%&quot; / &quot;two-column compare&quot;&gt;

Constraints per slide:

  • Title: ≤ 12 words
  • Bullets: 2-4 per slide, each ≤ 18 words
  • One idea per slide — if a slide needs 5+ bullets, split it
  • No slide is bullet-only — every content slide must plan at least one non-text element (icon, stat callout, two-column layout, comparison, timeline)
  • At least one "highlight" slide if the speech contains a strong line, statistic, or quote worth featuring

Stop here and wait for user approval. Do not start generating the PPTX. The user may say "OK" / "go" / "改一下第 5 页" / "风格换成极简" — handle edits to the outline first, re-render the outline if anything changed, then proceed.

Step 4 — Render the PPT (after approval)

Once the user approves the outline:

  1. Load the pptx skill (/workspace/.skills/pptx/SKILL.md) — it owns all PptxGenJS / layout / QA mechanics. Do not duplicate those rules here.
  2. Pick the theme + style recipe for the chosen preset from references/style-presets.md.
  3. For each slide in the approved outline:
    - Assign it one of the 5 page types from references/slide-types.md (Cover / TOC / Section Divider / Content / Closing).
    - Write slides/slide-XX.js per the pptx skill's From-Scratch Workflow.
  4. Compile → QA → deliver. The pptx skill's QA gate is blocking; do not hand the file to the user until QA passes.
  5. Send the file via &lt;deliver-assets&gt; (do not just print the path).

Output contract

Step 3 deliverable (outline): Markdown numbered list, exactly N entries. Render inline in the chat. No file yet.

Step 4 deliverable (PPTX): One .pptx file at /workspace/&lt;descriptive-name&gt;.pptx, wrapped in &lt;deliver-assets&gt;. The file:

  • has exactly N slides (no more, no less)
  • uses the chosen preset's palette + style recipe consistently
  • passes the pptx skill's QA gate
  • has a descriptive filename (never presentation.pptx)

Failure handling

Situation What to do
User picks N < 3 Stop. Recommend N ≥ 5 and ask them to re-pick.
Transcript is too dense to fit in N Tell the user which key points would be cut, ask whether to (a) increase N, (b) drop those points, or (c) keep all and let some slides be dense. Do not silently drop material.
Transcript is too thin for N Insert section dividers or stat-callout slides. If even that doesn't reach N, ask the user for more material or to lower N.
Transcript has no clear structure Ask the user to either provide an outline or confirm "use my paragraph breaks as sections".
User edits the outline Re-render the affected entries, get re-approval, then render. Do not "fix forward" silently.
User changes style after Step 4 starts Restart from Step 3 with the new style — the page type mix may need to change.
pptx QA fails (e.g. canvas overflow, missing page numbers) Fix and re-verify. Do not deliver a known-broken file.

Examples

Example 1 — standard case

User: "把下面这段 1200 字的演讲稿做成 10 页 PPT,金融风格,标题叫《2026 宏观展望》。[paste transcript]"

Step 3 outline:

[Slide 1]  Cover  2026 宏观展望
  - 主讲:[presenter] · 日期:2026-07-24
[Slide 2]  TOC
  - 01 全球增速 · 02 通胀路径 · 03 政策走向 · 04 资产配置
[Slide 3]  Section Divider  01 全球增速
[Slide 4]  Content (stat callout)  发达经济体放缓至 1.8%
  - 新兴市场仍是 4.2% 的主要拉动
  - visual: 大数字 + 副标
[Slide 5]  Content (comparison)  2025 vs 2026 增速对比
  - 视觉:两列对比卡
[Slide 6]  Section Divider  02 通胀路径
[Slide 7]  Content (data viz)  核心 CPI 回落节奏
  - visual: 折线示意 + 三条 takeaway
[Slide 8]  Content (text + icon)  政策走向与流动性
  - 三条要点 + 图标
[Slide 9]  Content (timeline)  资产配置建议
  - 4 步流程
[Slide 10]  Closing  关键结论
  - 3 条 takeaways

User: "OK,开干。"
→ Render with finance palette + Sharp style, deliver &lt;deliver-assets&gt; file.

Example 2 — outline too thin

User: "这是我的 5 分钟开场致辞草稿,帮我做 8 页极简风格的 PPT。"

Detection: transcript ≈ 200 words, only one section.

Response: 先告诉你一个发现 — 草稿内容撑不满 8 页,硬塞会显得很空。我建议两种走法,你选一个:
- (a) 降到 4-5 页,封面+3-4 张内容+收尾,节奏紧凑
- (b) 保留 8 页,中间插 2-3 张"金句卡"(把你的开场金句放大单独成页),更符合极简风格的留白感

不要直接开干,先对齐再写。

References

  • references/style-presets.md — 7 套风格的 palette + style recipe + 适用场景
  • pptx skill — LAYOUT_16x9 边界、PptxGenJS 模式、QA 流程(必读)
  • pptx skill → references/slide-types.md — 5 种 page type 的版式选择
  • pptx skill → references/design-system.md — 调色板与字体规则

常见问题

5 问
答案由本站依据条目字段整理
演讲稿生成PPT怎么安装?

在支持 Skills 的客户端里执行:skillhub install user_db22db11/speech-to-ppt。安装方式为未标注。装完刷新技能列表或重开客户端。

演讲稿生成PPT支持哪些客户端?

资料没有标注宿主客户端,但只要客户端支持 Skills(Claude、Codex、Cursor 等),一般都能加载。

演讲稿生成PPT权限高吗,安全吗?

权限级别为未标注权限。安装前先看它会读取或写入哪些内容。

演讲稿生成PPT收费吗?

本站不代收费用,也没有标注具体价格。是否收费取决于技能本身和所用的模型或客户端,正式使用前按作者说明确认。

演讲稿生成PPT用什么许可证?

许可证标注为 未说明(许可证不明确)。商用或二次分发前请到仓库确认条款原文。

热门专题

按场景挑技能
全部专题 →

技能内容以作者仓库为准;分类、权限级别与许可证状态的中文说明为中文整理,最近核验 未标注。