Yuwen Publish Precheck | 发布前审

Publish with a clearer view of content risk

Yuwen Publish Precheck (发布前审) helps Chinese-language creators review scripts, articles, social posts, subtitles and cover text before posting to Douyin (抖音), Xiaohongshu / RedNote (小红书), or WeChat Channels (微信视频号). This is the free, MIT-licensed public release, v1.0.6.

What it does

• Runs a local Python scan to locate wording that needs attention, then guides your AI agent through contextual review. A keyword match is a candidate, not a verdict.
• Produces a separate conclusion for each target platform, with quoted text, rule references, severity and suggested replacement wording.
• When asked to revise, preserves the author's meaning and voice, then rechecks the revised draft.
• Supports commercial-content review and additional medical/health and financial-content rule references.
• Helps investigate removal or recommendation issues while distinguishing documented violations from unknown causes.
• Stores personal wordlists, false-positive exceptions, expression examples and review history in local Markdown files.

Start with one sentence

“用 yuwen-publish-precheck 审一下这篇口播稿能不能发抖音：（贴上稿子）”

“帮我把这篇改成小红书能发的版本。”

“我这条视频被限流了，帮我复盘一下为什么。”

Included

SKILL.md, the Python scanner and built-in patterns, platform and industry references, editable templates, documentation, MIT license and third-party notices. No third-party Python packages or scanner API key are required. The agent needs local file access and Python 3.10+.

Scope and boundaries

This release focuses on text. It does not guarantee approval or reach, predict hidden ranking systems, provide moderation-evasion tricks, or publish on your behalf. Official references carry verification dates and may need checking as platform policies change. The scanner runs locally; the AI agent you use has its own data-handling policies.

Source, updates and issues: https://github.com/yuwen-cool/yuwen-publish-precheck

中文：发布前先定位风险，再结合语境判定；需要改稿时保留原意和语气，改完复检，并把你的经验沉淀成本地规则库。