TypingMindTypingMindThe best frontend for AI
Finance Explainer logo

Finance Explainer

CommunityPopular
andyhuo520
finance-explainer

「小Lin说」式财经/科技解说视频工作室：选题打分 → 带出处的事实清单（Exa/Tavily/Serper/Firecrawl）→ 按危机型/收购战型/悖论型骨架写有活人感的口播剧本（拍表 episode.json）→ 主播自录音频对齐跳切（或 MiMo 占位）→ 原创数字人出镜（Renoise：Wan 3.0 音频驱动口型 / Grok 1.5 原生说话+后换音色）→ 空镜与梗图（Renoise 便宜模型 + Unsplash/Pexels/Wikimedia）→ Remotion 引擎合成：主播全屏↔缩小↔圆形小窗的「放大缩小」、两代舞台主题（v1 紫色几何 / v2 深色玻璃）、日期贴纸、大数字滚动、术语卡、原文标黄、问题云、竞猜、卡通贴纸独白、章节卡、字幕、BGM 避让与音效。触发：做一期小Lin说那样的视频、财经解说、一口气了解 X、讲讲某个金融事件/收购案/国家经济、数字人口播财经、把这个新闻做成解说视频、复刻财经频道形式。不用于：单支品牌宣传片（用 promo-video）、系列课（用 ai-teardown）、冒充真实主播。

Overview

Publisherandyhuo520
Repositorydigital-human-koubo
Skill namefinance-explainer
Stars
1.6K
Forks
354
Bundled files
39
LicenseMIT
Links
Skill sourceRepository
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 39 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by andyhuo520 on GitHub. Read the source before you install it.

Installation

Install the Finance Explainer AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/andyhuo520/digital-human-koubo/tree/main/skills/finance-explainer
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/andyhuo520/digital-human-koubo.git /tmp/digital-human-koubo
mkdir -p .claude/skills
cp -r /tmp/digital-human-koubo/skills/finance-explainer .claude/skills/finance-explainer
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Finance Explainer in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Finance Explainer on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Finance Explainer is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

财经解说工作室（小Lin说式）

把一个财经/科技话题做成 2–25 分钟的横屏解说片：一个有活人感的主播坐在书房里聊，配图表、文件、空镜和梗，镜头在「主播全屏」和「主播缩成小窗、旁边上图」之间来回切。

完整的频道拆解（160 支元数据、97 支字幕统计、6 支逐帧分类）在 references/channel-teardown.md，网页版报告：https://claude.ai/artifact/EzcUvSJmVqgFJBGpfn3G3Y

红线（每次都适用）

  • **不冒充真人。**数字人只用原创形象：纯文字生成，或者用户本人 / 明确同意的人的照片。不用真实创作者（包括小Lin本人）的脸、改过发型眼镜的同一张脸、或从他们视频里克隆的声音——截图式「授权」、加水印加声明都不改变这一点。用户坚持时，说明理由并给原创方案。
  • **不搬内容。**文案、布景、画面全部原创；只学结构和手法。不用影视剧片段、新闻台画面；真实人物用自由协议照片或文字名牌。
  • **每个数字有出处。**事实清单 facts.md 里每条都带日期 + 来源 URL；剧本里的数字必须能在清单里找到；查不到的写 ⚠ 且不进剧本。拿不准的地方让主播直说「这个我没查到」——这本身就是活人感手法。
  • **生成要花用户的 credits。**Renoise 任何 create_tasks 之前，先给出模型、提示词、参数、数量、estimate_cost 总价和余额，等用户在对话里明确同意。口型视频依赖最终音频：音频没定稿前，不要把整片口型都做掉（会付两次钱）。

流程

0. 准备

  • 检索 key 在 ~/.config/xiaolin-explainer/keys.env（EXA / TAVILY / SERPER / FIRECRAWL / UNSPLASH / 可选 PEXELS、PIXABAY）。只读这个文件，不要扫描系统找密钥。
  • 引擎：bash scripts/setup_engine.sh <项目目录>（复制 templates/engine，node_modules 优先从本机已有 Remotion 项目 cp -c 克隆，失败再 npm install）。

1. 选题（references/topic-and-titles.md）

五项打分卡 ≥3 才做；给 3 个标题候选（「一口气了解 + 大对象」「【硬核】」「带数字」「疑问句 + 竖线钩子」）。确认时长：完整版 15–25 分钟表现最好，演示/跑通先做 2 分钟。

2. 事实清单

python3 scripts/research.py multi "<查询>" 找线索 → 打开原文核对（research.py scrape <url>）→ 写 facts.md：分组、每条「事实（精确数字）— 日期 — 来源 + URL」，最后写「容易讲错的地方」。可派子智能体并行查，但交回后自己抽查关键数字。

3. 剧本 = 拍表 episode.json（references/narrative-playbook.md + episode-schema.md）

  • 选骨架：A 危机/暴雷型（怎么死的）、B 收购/博弈型（怎么打的）、C 悖论型（为什么反常）。
  • 每拍一段口播 + 一个主画面（host: full/side/pip/off，或 stage 图表、bg 空镜）+ 若干叠加卡片，口播里用 [[cue]] 标卡点。
  • 文案指标：成片约 6 字/秒（录音约 4.8）、每分钟 1–2 个设问、3–4 个带参照物的数字、1 个英文术语；至少放 3 个活人感手法（开场个人状态 / 替观众说质疑 / 拟人独白 / 生活比喻 / 坦白没查到 / 竞猜停顿 / 先说答案 / 结尾回扣）。
  • 同时导出录音稿（每拍一段，去掉 cue 标记，写清读法：英文缩写、数字年份）。

4. 声音

  • 首选：主播自录。python3 scripts/align_recording.py 录音.m4a episode.json voice/ → 逐句对齐、口误重来只留最后一遍、拍内长停顿压到 0.16 秒，输出 timings.json。报告里 ⚠（对齐 <80%）的拍要人工听。
  • 占位：python3 scripts/narrate_mimo.py narration.json voice --gap 0.22 --lead 0.3（MiMo 预设音色，语气说明写进 voice.style）。占位就是占位，正式版换真人音频后整条流水线不变。

5. 数字人（references/production.md）

  • 定妆照：Seedream 5.0 Pro 16:9 2k（≈6），书房中景、两侧留白放图表、有专属道具（霓虹名字灯、玩偶）。用户给了原创脸就作参考图，否则纯文字生成。
  • **方案 W（Wan 3.0，音频驱动，便宜）：**参考图 + 该拍音频 → 口型跟着声音走；480p ≈ 6 credits/秒。输出里音频比画面晚 0.23 秒，host.json 的 offset 填 0.23。缺点：眼神偏呆、细节一般。
  • **方案 G（Grok Video 1.5，原生说话，更生动）：**一张图 + 台词写进提示词 → 自带说话声和口型；720p 15 秒 255。之后用 Seed Audio 做「保留节奏换音色」替换成主播声音。适合全屏特写镜头。
  • 小窗（pip）用 3 段 10 秒通用说话循环轮换即可，省掉约 40% 口型费用。
  • 先做 10 秒测试：抽帧看停顿处嘴是否闭合、人物是否走样，再批量。

6. 空镜 / 梗 / 贴纸

  • 空镜：Renoise Wan 3.0 480p 5–10 秒（audioGeneration=false），提示词写「无文字无 logo」；或 scripts/fetch_stock.py <id> "<关键词>" engine/public/assets --kind photo|video（自动记 credits.json，片尾署名）。
  • 梗图：原创短片（如「小机器人一群围住大机器人」），不用影视片段。
  • 卡通贴纸：grok-image 纯绿底 → 抠绿成透明 PNG（脚本见 production.md）。

7. 合成

python3 scripts/build_timeline.py episode.json voice/timings.json <engine>/public --assets assets.json --host host.json --bgm music/bgm.mp3 --badge "…" → npx remotion still Episode out/x.jpg --frame=N 抽帧检查（含过渡中间帧）→ npx remotion render Episode out/raw.mp4 --concurrency=6 → ffmpeg loudnorm=I=-16 + CRF 21 出片。

8. 质检清单

字幕按原文标点断句且 ≤13 字；卡片不压字幕、不出画；主播缩放过渡中间帧无残影；口型片段对齐（offset）；数字与 facts.md 一致；角标/声明在；响度 -16 LUFS；片尾列出素材与数据出处。

目录

  • references/channel-teardown.md — 频道数据与结论
  • references/topic-and-titles.md — 选题打分卡、标题公式
  • references/narrative-playbook.md — 三种骨架、活人感手法、语言习惯、文案指标
  • references/visual-grammar.md — 16 种画面语法 ↔ 引擎组件 ↔ 拍表字段
  • references/episode-schema.md — episode.json / host.json / assets.json 字段
  • references/production.md — 数字人两方案、Renoise 提示词与价格、素材来源与授权、抠图、渲染与坑
  • scripts/ — research.py、fetch_stock.py、narrate_mimo.py、align_recording.py、build_timeline.py、setup_engine.sh；频道分析工具 classify.py / sheet.py / copy_stats.py / vtt2txt.py（拆别的频道时用）
  • templates/engine — Remotion 引擎（Episode 合成、Host、Stage、Overlays、Misc 组件，两代主题）
  • examples/ai-chips — 「中国买不到顶级显卡，为什么还是世界第二 AI 大国」：facts.md、10 分钟拍表、2 分钟拍表、录音稿；examples/svb — 硅谷银行 3.5 分钟拍表

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Finance Explainer AI skill do?

「小Lin说」式财经/科技解说视频工作室：选题打分 → 带出处的事实清单（Exa/Tavily/Serper/Firecrawl）→ 按危机型/收购战型/悖论型骨架写有活人感的口播剧本（拍表 episode.json）→ 主播自录音频对齐跳切（或 MiMo 占位）→ 原创数字人出镜（Renoise：Wan 3.0 音频驱动口型 / Grok 1.5 原生说话+后换音色）→ 空镜与梗图（Renoise 便宜模型 + Unsplash/Pexels/Wikimedia）→ Remotion 引擎合成：主播全屏↔缩小↔圆形小窗的「放大缩小」、两代舞台主题（v1 紫色几何 / v2 深色玻璃）、日期贴纸、大数字滚动、术语卡、原文标黄、问题云、竞猜、卡通贴纸独白、章节卡、字幕、BGM 避让与音效。触发：做一期小Lin说那样的视频、财经解说、一口气了解 X、讲讲某个金融事件/收购案/国家经济、数字人口播财经、把这个新闻做成解说视频、复刻财经频道形式。不用于：单支品牌宣传片（用 promo-video）、系列课（用 ai-teardown）、冒充真实主播。

Why use Finance Explainer on TypingMind?

Because you install it once and use it with any model. Finance Explainer is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Finance Explainer in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/andyhuo520/digital-human-koubo/tree/main/skills/finance-explainer. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Finance Explainer?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Finance Explainer?

As many as you like. As long as a model supports skills, you can use Finance Explainer with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Finance Explainer AI skill free?

Yes. It is published on GitHub by andyhuo520 under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

Related AI Skills

View all
NousResearch logo

Pixel Art

NousResearch
OrganizationPopular

Pixel art w/ era palettes (NES, Game Boy, PICO-8).

53.5K
250.4K
6 files
View
NousResearch logo

Inference Sh Cli

NousResearch
OrganizationPopular

Run 150+ AI apps (image, video, LLM) via inference.sh CLI.

53.5K
250.4K
4 files
View
NousResearch logo

Weights And Biases

NousResearch
OrganizationPopular

W&B: log ML experiments, sweeps, model registry, dashboards.

53.5K
250.4K
3 files
View
NousResearch logo

Nemo Curator

NousResearch
OrganizationPopular

Curate LLM training data: dedupe, filter, PII redaction.

53.5K
250.4K
2 files
View

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇