Media Publish Check
Pre-publish Compliance-Audit für Douyin, Xiaohongshu, WeChat Channels und Kuaishou Inhalte
Getestet · Funktioniert
Was es kann
Audits videos, articles, subtitles, covers, livestream scripts und product claims before they go live on Chinese social platforms, flagging banned words, off-platform diversion, missing AI/ad/repost disclosures und medical oder absolute claims. Wird ausgelöst, wenn requests like "can I post this", "pre-publish review", "check for banned words" oder "will this get throttled" und returns exact risk locations plus platform-specific revisions.
Testbericht
Führte den gebündelten preflight.py scanner live on a fake Douyin caption aus: it caught six real risk candidates in one pass, including the absolute claim, the off-platform WeChat/Taobao diversion, the AI-generation disclosure trigger und a medical 'cure insomnia' claim. Against a no-skill baseline the skill adds a real edge: an R0-R4 taxonomy, the mandatory six-choice AI/ad/repost labeling gate, 14 per-platform reference files und a where-to-fix output format instead of a vague opinion. Honest about limits too: only 17 scanner rules (it missed a couple of superlatives) und the SKILL.md correctly states the script only surfaces candidates, never a publishing verdict.
Getestet am: 2026-07-20 · Claude Code 2.x (agent harness)
Installation
git clone https://github.com/XshuiAi/media-publish-check.git mkdir -p ~/.claude/skills cd media-publish-check && cp -r . ~/.claude/skills/media-publish-check
Befehle & Beispiel-Prompts
/media-publish-checkPre-publish Compliance-Audit für Douyin, Xiaohongshu, WeChat Channels und Kuaishou Inhalte
Skills reagieren auf normale Anfragen — keine Slash-Befehle nötig. Nach der Installation aktivieren Prompts wie diese den Skill (auf Englisch):
这条抖音文案现在能直接发吗?帮我检查一下有没有违禁词和过度承诺的说法风险,谢谢啦发布之前帮我审核一下这篇小红书笔记,看看有没有合规风险和敏感违禁词汇,越细越好哦这场直播脚本会不会因为违规被平台限流?检查一下有没有引流和虚假宣传的内容存在哦,谢谢