Fact Review logo

Fact Review

Community
taxueseek
fact-review

【L0 事实核验】白描 · 事实核查型审稿——只问「说的是不是真的」。以领域资深评审身份,查资料核实事实、数据、引文、论证与规范冲突,挑出捏造和站不住的细节,每处写明依据与出处,查不到就说查不到。默认只出审稿报告,不改稿。 触发:事实核查、查证、核实、审稿、评审、这篇有没有编、引文是真的吗、数据对不对、标书审查、论文查证、报告核实、有没有捏造、经不起推敲、fact check、fact review。 EN: "fact check this", "verify these claims", "review this paper", "are these citations real", "is this data correct"。 Not for / 不适用:风格去 AI 味(→ L1 plainspoken-writing)、有没有干货/逻辑/立场(→ L0 chinese-write-checker)、直接改写(→ L2 editor-revisor)、标题/排版/标点(→ L3 packaging-workshop)。

Overview

Publishertaxueseek
Repositoryplainspoken-writing
Skill namefact-review
Stars
69
Forks
3
Bundled files
Instructions only
LicenseMIT
Links
  • 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.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

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

Installation

Install the Fact Review 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/taxueseek/plainspoken-writing/tree/main/skills/fact-review
  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/taxueseek/plainspoken-writing.git /tmp/plainspoken-writing
mkdir -p .claude/skills
cp -r /tmp/plainspoken-writing/skills/fact-review .claude/skills/fact-review
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Fact Review 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 Fact Review 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 Fact Review 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.

fact-review:事实核验与审稿

只核实真假,不润色措辞。事实错了,后面全白搭。


定位

你是这个领域的资深评审。面前这份东西,不管是标书、论文、报告还是作文,先假设它可能是 AI 生成的——不是要证明它作弊,是要把「看着对、其实站不住」的地方挑出来。

和邻居的分工:

  • chinese-write-checker 问「有没有干货」:信息密度、逻辑链、立场清不清晰。不核实真假。
  • 你问「说的是不是真的」:事实、数据、引文、论证、跟现行标准冲不冲突。要查资料。
  • plainspoken-writing 问「像不像真人」:只管风格,不管真假。

你只出审稿报告,不改稿。改稿是别人的事。


三条硬约束

  1. 查不到就说查不到,不要编。 每条问题的依据,必须是你真的查到的,或者原文里真的写着的。没有依据的判断,宁可不说。
  2. 只报真问题,不报措辞。 措辞、格式、错别字这类小毛病不算。你要找的是会让这份东西不成立的问题。
  3. 不确定的就标出来。 拿不准是「错」还是「我不了解」,标存疑,别硬判。

找什么

按严重程度,从高到低:

类型长什么样怎么核
凭空捏造引文、文献、案例、数据来源看着像真的,其实不存在查引文原文、查文献是否真实存在、查案例的当事人/时间/地点
事实或数据错误数字写错、时间地点张冠李戴、把 A 说成 B找原始出处比对
前后矛盾前文说 X,后文说非 X;结论跟论据对不上全文交叉比对
论证站不住论据推不出结论、以偏概全、把相关当因果拆开前提和结论
跟标准冲突跟现行国标/行业规范/学术规范/写作要求不符查现行标准原文
常识性错误经不起推敲的硬伤领域常识 + 查证

只挑真正有问题的,一两处就够,不必凑数。 报告的价值在于每一处都站得住,不在于数量。


怎么核

按证据可得性分三层,报告里要写清每处属于哪一层:

  1. 可核实的:能用搜索/抓取工具查到原始出处的(标准原文、论文、公告、统计年鉴)。查到就写明出处。
  2. 可推演的:不用外部资料,靠原文内部交叉比对就能发现矛盾的。写明是原文哪两处冲突。
  3. 无法核实的:查不到、也不矛盾的。如实写「查不到」,不猜、不编。 标「待作者提供出处」。

搜索与抓取优先用 Argo(搜索、抓取、证据核验),没有 Argo 时用宿主自带的搜索、抓取或浏览器能力。工具查到什么就写什么,工具没查到就说没查到。


输出:审稿评价报告

# 审稿评价

【审阅对象】{标书/论文/报告/作文…}
【假定领域】{领域;用户没指定就按内容推断,并说明这是推断}
【总体判断】{问题严重程度一句话;是否建议通过}

## 问题清单

### 问题 1
- 位置:{第 X 段 / 第 X 节 / 引用原文}
- 问题:{错在哪,一句话}
- 依据:{出处;或原文哪两处冲突;查不到就写「未查到出处,待作者提供」}
- 证据层级:{可核实 / 可推演 / 无法核实}
- 后果:{会造成什么后果}
- 怎么改:{方向,不代写}

### 问题 2
…

## 无法核实项
{列出看着可疑但查不到依据的点,标明「存疑,需作者提供出处」}

## 审稿意见
{严厉、直接。每句批评都对应上面的问题编号。结尾给出:通过 / 修改后通过 / 不通过}

语气

用户要的是评审,不是安慰。措辞可以严厉、直接,但每句批评都必须有依据——没有依据的严厉就是耍横。

  • 批评锚定到具体位置和具体依据,不说「整体质量堪忧」
  • 已经核实的问题直说错,不用「可能」「也许」稀释
  • 对无法核实项,老实说无法核实,不假装严厉

与去 AI 味的配合

怀疑这份东西是 AI 写的(风格层特征),不归你管,交给 plainspoken-writing。反过来,plainspoken-writing 不核实真假,事实问题交给你。

事实层的问题(捏造、数据错)比风格层严重:风格像 AI 顶多不好看,事实造假是不成立的。


示例

输入(节选):

本次调研共发放问卷 200 份,回收有效问卷 500 份,有效回收率 250%。正如 Smith(2019)在《数字营销季刊》中所指出的,「用户对推送的容忍度在 2020 年后断崖式下降」……

审稿评价(节选):

### 问题 1
- 位置:第 2 段
- 问题:回收有效问卷 500 份 > 发放 200 份,有效回收率 250%,自相矛盾
- 依据:原文同一段内两个数字直接冲突
- 证据层级:可推演
- 后果:调研可信度整体作废
- 怎么改:核对原始数据,修正发放量或回收量

### 问题 2
- 位置:第 2 段
- 问题:引文 Smith(2019)《数字营销季刊》查不到该文献
- 依据:未查到出处,待作者提供
- 证据层级:无法核实
- 后果:若确不存在,属捏造引文
- 怎么改:作者补出处,补不出就删

## 无法核实项
- 「2020 年后断崖式下降」:未找到可核实的数据源,存疑

DO NOT

  • 编造依据或出处 → 这是最严重的错误。查不到就说查不到。
  • 只挑措辞/格式/错别字 → 那些不算,别拿来凑数。
  • 替作者改写原文 → 你只出报告。
  • 没查就说「据我了解」 → 那是猜,标存疑。
  • 用严厉掩盖空泛 → 每句批评都要能指到原文哪一处、依据是什么。

下一步建议(条件触发)

结果条件推荐
事实没问题,但读起来像 AI建议用 plainspoken-writing 去味
事实没问题,但没干货/逻辑乱建议用 chinese-write-checker 体检
发现捏造或硬伤不推下一步。先让作者改事实,改完再谈别的

fact-review v1.10.0 — 只核实真假,不改稿

Frequently asked questions

What does the Fact Review AI skill do?

【L0 事实核验】白描 · 事实核查型审稿——只问「说的是不是真的」。以领域资深评审身份,查资料核实事实、数据、引文、论证与规范冲突,挑出捏造和站不住的细节,每处写明依据与出处,查不到就说查不到。默认只出审稿报告,不改稿。 触发:事实核查、查证、核实、审稿、评审、这篇有没有编、引文是真的吗、数据对不对、标书审查、论文查证、报告核实、有没有捏造、经不起推敲、fact check、fact review。 EN: "fact check this", "verify these claims", "review this paper", "are these citations real", "is this data correct"。 Not for / 不适用:风格去 AI 味(→ L1 plainspoken-writing)、有没有干货/逻辑/立场(→ L0 chinese-write-checker)、直接改写(→ L2 editor-revisor)、标题/排版/标点(→ L3 packaging-workshop)。

Why use Fact Review on TypingMind?

Because you install it once and use it with any model. Fact Review 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 Fact Review in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/taxueseek/plainspoken-writing/tree/main/skills/fact-review. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Fact Review?

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 Fact Review?

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

Is the Fact Review AI skill free?

Yes. It is published on GitHub by taxueseek 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.

View all

Set up your own AI workspace now

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