← all repositories
aiworkskills/wechat-article-skills

WeChat content ops, packaged as nine agent skills

It automates the full WeChat article lifecycle by breaking it into nine agent skills for tools like Cursor and Claude Code.

★650 stars Python AgentsLLMOps · Eval
wechat-article-skills
Not currently ranked — collecting fresh signals.
star history

What it does

This project is a set of prompt-and-tool bundles that turn an AI agent into a WeChat Official Account editor. You tell it to write an article, and it handles topic research, drafting, sensitivity review, Markdown-to-HTML formatting, image generation, and API publishing, pausing at each step for your approval. Configuration happens through a web dashboard or by dropping YAML files and a .aws preset package into a local directory.

The interesting bit

Instead of building yet another monolithic SaaS, the authors split the workflow into discrete agent skills that plug into existing tools like Cursor, Claude Code, Codex, and the OpenClaw ecosystem. The .aws preset system is the clever glue: it bundles your brand voice, visual themes, and business context into a single file the agent can read, so the AI doesn’t need to re-interview you every session.

Key highlights

  • Nine modular skills cover the full editorial chain: topic selection, writing, review, formatting, image generation, publishing, and a business asset library.
  • Formats Markdown into WeChat-specific HTML with four built-in visual themes and fourteen image styles.
  • Publishes directly via the WeChat API, including multi-account support and automatic image compression.
  • Configuration and style presets are managed through a visual web dashboard or raw YAML for the terminal-inclined.
  • Compatible with a wide range of Chinese and Western LLMs via OpenAI-compatible APIs; API keys stay local.

Caveats

  • Image generation requires a separate model configuration; agents that only handle text won’t produce visuals out of the box.
  • The multi-image “sticker” post skill is noted as being in internal beta.
  • Much of the convenience relies on the aiworkskills.cn web dashboard for preset management.

Verdict

If you operate WeChat Official Accounts and already live inside Cursor or Claude Code, this saves you from copy-pasting between a dozen tools. If you don’t publish in Chinese or don’t trust an AI with your WeChat API credentials, look elsewhere.

Frequently asked

What is aiworkskills/wechat-article-skills?
It automates the full WeChat article lifecycle by breaking it into nine agent skills for tools like Cursor and Claude Code.
Is wechat-article-skills open source?
Yes — aiworkskills/wechat-article-skills is open source, released under the Apache-2.0 license.
What language is wechat-article-skills written in?
aiworkskills/wechat-article-skills is primarily written in Python.
How popular is wechat-article-skills?
aiworkskills/wechat-article-skills has 650 stars on GitHub.
Where can I find wechat-article-skills?
aiworkskills/wechat-article-skills is on GitHub at https://github.com/aiworkskills/wechat-article-skills.

heatdrop uses Google Analytics to see which pages get read — nothing else. Your call. How we handle data.