// Publish Markdown articles to WeChat Official Account draft box using a local config file. Use when user wants to push blog posts to 微信公众号, convert Markdown to WeChat format, or create 公众号草稿.
[HINT] Download the complete skill directory including SKILL.md and all related files
| name | wechat-mp-publisher |
| description | Publish Markdown articles to WeChat Official Account draft box using a local config file. Use when user wants to push blog posts to 微信公众号, convert Markdown to WeChat format, or create 公众号草稿. |
| metadata | {"version":"0.6.0"} |
Publish Markdown articles to WeChat Official Account (微信公众号) draft box.
uv run <skill_dir>/scripts/publish.py -f <markdown-file> [-u <source-url>] [-c <cover-image>] [-t <title-override>] [-a <author>] [--config <config-file>]
The script:
draft/add API-f, --file — Markdown file path (required)-u, --url — Original article URL (set as "阅读原文" link)-c, --cover — Cover image path (defaults to first image in article)-t, --title — Override title (defaults to frontmatter title)-a, --author — Author name (defaults to "张昊辰")The script reads a local JSON config file. Default path:
/root/.openclaw/config/wechat-mp-publisher.json
Optional overrides: --config <path> or WECHAT_MP_CONFIG=<path> for one command. Do not store WeChat credentials in OpenClaw env.vars.
{
"app_id": "<wechat official account app id>",
"app_secret": "<wechat official account app secret>"
}
[^N] references become green superscript numbers; definitions at bottom become styled "参考文献" section with small gray text. External links in footnotes are converted to plain text (WeChat filters external URLs).**text**immediately-followed patterns that markdown-it can't parse.<section> instead of <pre>/<code> to bypass WeChat's forced pre-wrap.<p> + bullet/number characters (WeChat adds block-level wrappers inside <li>, breaking inline formatting).Optimized for both light and dark mode reading:
Draft created in WeChat MP backend (mp.weixin.qq.com). User confirms and publishes manually.
[^*\n]+? (not [^*]+?) to prevent matching across paragraph boundaries. Without the \n exclusion, **end of paragraph pairs with beginning of next**, corrupting entire sections.--- separators: If the source Markdown has a horizontal rule between content and footnote definitions, the footnote extraction leaves an orphan ---. The script cleans trailing \n---\n before rendering.[text](url) links in footnotes are stripped to plain text because WeChat silently drops external links from article content.<code> elements without explicit display:inline may be rendered as inline-block on WeChat mobile, causing surrounding text to break with large gaps. The style now includes display:inline;word-break:break-all;line-height:inherit;.**text**： added a space before CJK punctuation (**text** ：), causing WeChat to render extra visual gaps in list items. Fixed by excluding CJK punctuation (U+3000-303F, U+FF00-FFEF) from the lookahead.\*\*([^*\n]+?)\*\* matched across two bold boundaries when the closing ** of one bold and the opening ** of the next were on the same line, producing literal ** in output. Fixed with negative lookahead (?![。！？.!?\s]).white-space:pre-wrap (v0.5.0 fix): WeChat mobile (especially iOS) forcibly overrides white-space:pre to pre-wrap on <pre>/<code> elements, causing code to wrap instead of horizontal scroll. Fixed by replacing <pre><code> with <section> elements — WeChat preserves <section> styles as-is.<li> in block-level <section>, causing a line break between bold labels and their descriptions. Fixed by converting <ul>/<ol> to <p> tags with bullet/number characters.<a> tag rejection (v0.4.0 fix): WeChat errcode 45166 caused by <a> tags; all links stripped keeping text.<strong> block rendering (v0.4.0 fix): WeChat forces <strong> to block display; replaced with <span style="font-weight:600">.<pre> blocks lacked white-space:pre-wrap, causing indentation and line breaks to collapse in WeChat. Code block colors updated from Catppuccin to One Dark (#282c34) for better readability.