Gzh Article logo

Gzh Article

CommunityPopular
mateaix
gzh_article

公众号图文创作 / 推文 / 官方号文章 (official account article) — 端到端:选题→搜集→成文→配图→去AI化→公众号内联样式排版→交付/草稿箱。honors user persona & style memory.

Overview

Publishermateaix
Repositorymateclaw
Skill namegzh_article
Stars
1.1K
Forks
337
Bundled files
6
LicenseApache-2.0
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.

  • 6 bundled files

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

  • Open source

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

Installation

Install the Gzh Article 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/mateaix/mateclaw/tree/dev/mateclaw-server/src/main/resources/skills/gzh_article
  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/mateaix/mateclaw.git /tmp/mateclaw
mkdir -p .claude/skills
cp -r /tmp/mateclaw/mateclaw-server/src/main/resources/skills/gzh_article .claude/skills/gzh_article
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Gzh Article 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 Gzh Article 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 Gzh Article 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.

公众号图文创作

把一个选题做成可直接粘进公众号编辑器的图文推文。7 个阶段,每步都接到平台真实工具上。

📌 动笔前务必先读 references/gzh_platform_rules.md —— 那是微信公众号的真实平台规矩(封面尺寸、标题/摘要、编辑器排版、诱导分享/关注红线、群发频次与时机、原创机制),懂了这些才叫"会做公众号",而不只是"会写字"。下面 SOP 的成文/配图/自查各步都以它为准。

开工前:读取共享人设记忆

先用 recall_structured 取回并全程遵守:

  • content_persona — 人设 / 口吻
  • writing_style_gzh — 公众号文风
  • topic_interests — 选题方向
  • banned_words — 禁用 / 敏感词
  • signature_blocks — 固定开场 / 结尾 / 引导关注段

取不到就用中性默认,不要编造。

七步 SOP

1. 选题

用 web_search(freshness=week、language=zh-CN)围绕 topic_interests 找近期热点和角度,count 取 5–8。给用户 3–5 个候选选题(每个带一句话切入角度),让其确认或补充后再往下。

2. 搜集汇总(参考文章)

用户给出参考公众号链接时:

  • 优先用可能存在的 wechat_article_extract 工具直接抽正文(若该工具可用,用 load_skill 或工具列表确认)。
  • 没有该工具,就用 browser_use:先 action=open 打开链接,再 action=snapshot 抓取页面可见正文。
  • 把每篇参考的核心观点、结构、可借鉴角度提炼成要点。

红线:本技能产出原创内容,参考只用于找角度、补事实,并在文末标注引用来源。严禁洗稿 / 搬运 / 逐段改写他人文章。

3. 成文

按公众号结构模板成文(详见 references/gzh_structure.md):

  1. 钩子引言 — 用具体场景 / 反常识数据 / 一个问题抓住读者。
  2. 3–5 个小标题分节 — 每节一个论点,配具体案例或数据,不要空谈。
  3. 金句 — 每节或结尾埋一句可摘录的话。
  4. 结尾行动号召 + 引导关注 — 用 signature_blocks 里的固定收尾。

全程遵守 writing_style_gzh + content_persona。

同时定好标题和摘要(打开率的命门,见 references/gzh_platform_rules.md):

  • 标题:≤30 字为佳(后台上限 64),带具体信息 / 数字 / 悬念 / 情绪;与正文一致,不做标题党、不用"震惊体"、别满屏感叹号(会被限流)。给 2–3 个候选让用户挑。
  • 摘要(digest):手写 ≤120 字,补一句标题没说完的钩子(不填会被系统乱截正文前 54 字)。交付时作为 gzh_publish 的 digest 传入。

4. 配图

用 image_generate(action=generate):

  • 封面头图:aspectRatio=landscape(公众号头图是 2.35:1,约 900×383 / 1080×460)。prompt 写清主题、风格、留白;封面文字要大而少(缩略图很小,一行主标题足矣)。
  • 分享/朋友圈封面:如需单独出,用 aspectRatio=square(1:1,≥500×500)。
  • 关键小节配图:按需为 2–3 个重点小节各生成一张,风格与封面统一。

image_generate 只认 landscape / portrait / square 三种比例,其它比例映射到最近的一个(如 3:4 → portrait)。

5. 去 AI 化

load_skill deai_humanize,然后对全文跑它的"打分→改写→复检"循环,platform=gzh,目标 score ≤ 55。

6. 违禁词 + 平台规则自查

对照 banned_words 和 references/compliance_checklist.md,命中就标注并给出替换建议,不要静默通过。两类都要扫:

  • 广告法:极限词(最/第一/唯一/国家级/100%)、虚假功效、承诺收益、敏感内容、侵权。
  • 微信平台红线(比广告法更容易封号,见 references/gzh_platform_rules.md 第 5 节):诱导分享(集赞/助力/分享解锁)、诱导关注(关注才能看全文)、违规外链/二维码、标题党。引导互动只用话术、不用利诱。

产品教程/操作类文章配真实截图:写"如何用 MateClaw 做 XX"这类教程时,用 capture_screenshot(path) 截真实后台界面(path 为站内相对路径,如 /chat、/channels、/agents、/skills),把返回的图片 URL 以 ![步骤说明](URL) 直接嵌进对应步骤的 Markdown,替代【截图】占位;再整体交给 gzh_package。这样成品里是真实产品截图,不用手动补图。

7. 打包交付(gzh_package —— 在线预览 + 素材下载)

默认用 gzh_package 交付。你只需把成稿正文以 Markdown 形式传入,服务端会转成公众号内联样式 HTML(公众号不认 <style> 块和 class,一律内联),并产出「在线预览链接 + 素材 zip(article.html / article.md / 封面)+ 可粘贴的内联 HTML」:

gzh_package(title="<标题>", markdown="<正文 Markdown,含小标题/列表/引用/```代码块```/表格>",
            coverImageUrl="<第4步封面图链接>", author="<作者/来源>")

⚠️ 不要自己手写整段内联样式 HTML 再塞进 write_file 或 render_html_image(html=...)。大段、转义密集的 HTML 作为单个工具参数在流式传输时会被截断成非法 JSON、导致工具连续失败并触发循环熔断。让 gzh_package 在服务端生成 HTML,你只传紧凑的 Markdown。

把 gzh_package 返回的在线预览链接发给用户预览;满意后:

  • 直接进草稿箱:gzh_publish(action="draft", ...)(先向用户确认,发布是对外不可逆动作)。
  • 或让用户下载素材 zip / 复制内联 HTML,手动粘贴进公众号编辑器。

references/ 里的 gzh_layout*.html 是服务端配色/排版的参考(gzh_package 已内置同款风格);只有当用户明确要高度定制的特殊版式、且篇幅不大时,才手写内联 HTML 并用 render_html_image(html=...) 出预览图,注意控制体量避免截断。

定时 / 批量场景:内容日历 + 合规

长期投产(尤其每日定时)时:

  • 选题前查重(务必在成文前):content_item(action="check_recent", platform="gzh", topic="<选题>")。命中"疑似重复"就换角度或另选。只计已打包/已发布,不会把你这一轮刚产出的算进去。
  • 交付即扫即记(自动,无需手动):gzh_package 交付时会自动跑合规扫描(极限词/诱导/承诺/功效,报告并入返回)并自动记入内容日历。你只需在 gzh_package 里传上 topic="<选题>"(与 title 区分,用于台账去重指纹)。个人 / 品牌禁用词可另调 compliance_scan(text, extraBannedWords="<recall 到的 banned_words,逗号分隔>")。
  • 发布:gzh_publish 进草稿箱时对高危词(极限/诱导/承诺)硬拦截;真正发表后 content_item(action="mark_published", id, externalRef)。

保存自定义模板 / 对话升级技能

当用户对某个自创模板满意、想以后复用时,用 skill_manage 把它存成一个自定义技能(builtin=false 才能写入):

  1. 首次:skill_manage(action="create", name="my_gzh_templates", content="<一份 SKILL.md,说明这是我的公众号模板库>")。
  2. 存模板:skill_manage(action="write_file", name="my_gzh_templates", filePath="references/<模板名>.html", content="<内联样式HTML>")。
  3. 以后:readSkillFile(skillName="my_gzh_templates", filePath="references/<模板名>.html") 取回复用。

本技能 gzh_article 是内置技能,不能直接被编辑(skill_manage 会拒绝写内置技能);自定义模板一律存到用户自己的自定义技能里。每次写入都会过安全扫描。

参考

  • references/gzh_layout.html — 移动端优先的公众号内联样式 HTML 模板(通用)。
  • references/gzh_layout_minimal.html — 极简编辑风模板。
  • references/gzh_layout_business.html — 商务专业风模板。
  • references/gzh_structure.md — 文章结构 + 钩子 / 金句方法论。
  • references/gzh_platform_rules.md — 微信公众号平台规范与实操(封面尺寸、标题/摘要规则、编辑器排版、诱导红线、群发频次与时机、原创机制、敏感行业)。写公众号务读。
  • references/compliance_checklist.md — 合规自查清单(广告法禁用词 + 微信诱导红线)。

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 Gzh Article AI skill do?

公众号图文创作 / 推文 / 官方号文章 (official account article) — 端到端:选题→搜集→成文→配图→去AI化→公众号内联样式排版→交付/草稿箱。honors user persona & style memory.

Why use Gzh Article on TypingMind?

Because you install it once and use it with any model. Gzh Article 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 Gzh Article in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/mateaix/mateclaw/tree/dev/mateclaw-server/src/main/resources/skills/gzh_article. 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 Gzh Article?

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 Gzh Article?

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

Is the Gzh Article AI skill free?

Yes. It is published on GitHub by mateaix under the Apache-2.0 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 👇