Security audit
Security checks for vulnerabilities and agentic risk
This skill does what it says by publishing Xiaohongshu long-form posts, but it can perform a real public account action from a somewhat broad trigger without a clear user-controlled final confirmation.
Review this skill carefully before installing. It is not malicious, but users should only invoke it when they intend to publish to a real Xiaohongshu account, and they should require a final manual confirmation before the publish button is clicked.
The trigger phrases include broad language such as “写一篇小红书”, which can match ordinary content-writing requests rather than an intent to actually publish. In this skill’s context, activation causes navigation to a live publishing flow on the user’s Xiaohongshu account, so ambiguous triggering can lead to unintended external actions.
The skill does not clearly warn that it will perform a real posting action on the user’s external Xiaohongshu account. Because the workflow culminates in clicking a live “发布” button, a user may believe they are only drafting or formatting content when the skill is actually making an irreversible public-facing account action.
No suspicious patterns detected.
