小红书长文发布
Security checks across malware telemetry and agentic risk
Overview
The skill’s Xiaohongshu publishing purpose is clear, but it can use a logged-in account to make public posts without a clear final approval step.
Only install or run this if you trust the skill and intentionally want it to publish to Xiaohongshu. Use a dedicated browser profile, verify the account and content first, and prefer requiring a final manual confirmation before any public post is made.
VirusTotal
65/65 vendors flagged this skill as clean.
Risk analysis
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
An agent invocation could publish content to the user’s Xiaohongshu creator account before the user has a final chance to review it.
This instructs the agent to perform the final public publishing action, not merely prepare a draft or preview, and no required human confirmation gate is described.
技能会自动执行以下操作流程... 7. **点击"发布"** - 完成最终发布操作
Make draft or preview the default, require explicit user confirmation immediately before clicking Publish, and limit activation to clear user requests to publish.
If run in the wrong browser profile or account, it may post under that logged-in Xiaohongshu account.
The skill relies on an existing logged-in Xiaohongshu creator account/session to perform publishing actions. This is expected for the stated purpose but grants account-level posting authority.
- 小红书创作者账号已登录
Use a dedicated browser profile or account, verify the logged-in identity before running, and avoid leaving sensitive accounts logged in when not needed.
A user or workflow may believe a post was successfully published or verified when the included implementation did not actually perform or confirm the browser action.
The included runnable code shows browser actions only as comments and an empty success check, yet it returns a successful publishing result. If this file is used by the runtime or CLI, it could falsely assure the user that publishing was completed.
function clickPublishButton() {
// browser.act(kind="click", ref="e990") // 发布按钮
}
...
return { success: true, message: '发布成功' };Implement real OpenClaw browser calls and success verification, or clearly mark this file as pseudocode and avoid returning success without observed confirmation.