AI全链路科技资讯工厂

AdvisoryAudited by Static analysis on Apr 30, 2026.

Overview

No suspicious patterns detected.

Findings (0)

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.

What this means

The agent could post AI-generated content to the user's public social or official accounts, affecting reputation or compliance if the content is inaccurate or unwanted.

Why it was flagged

The workflow directs the agent to publish multiple public posts/articles, but the artifacts do not define a mandatory draft, preview, or user approval step before publication.

Skill content
用 xiaohongshutools 发布小红书笔记(3条:热点、应用、深度各1条) ... 发布到公众号
Recommendation

Add an explicit review-and-confirmation gate before every publish action, default to saving drafts, and require separate user approval for Xiaohongshu and WeChat publication.

What this means

If configured, the agent may act through the user's logged-in social accounts, including publishing content under that identity.

Why it was flagged

The skill expects access to platform session cookies and official-account authorization, while the supplied metadata declares no primary credential or required environment variables.

Skill content
确保已配置 cookies 和 session(参考 SKILL.md) ... 确认公众号授权状态
Recommendation

Declare required credentials clearly, document where sessions are stored, limit them to posting-only scopes where possible, and warn users before using account sessions.

What this means

The safety of the overall workflow depends on other installed skills that may have broader account or file access than this package discloses.

Why it was flagged

The workflow depends on local, unreviewed skills for high-impact functions such as social posting; their code and provenance are not included in the artifact set.

Skill content
"source": "local", "path": "/root/.openclaw/workspace/skills/xiaohongshutools"
Recommendation

Review and pin the dependent skills, especially publishing tools, before enabling the workflow or scheduling it.

NoteHigh Confidence
ASI10: Rogue Agents
What this means

If the user installs the cron job, the agent may continue generating and publishing content on a schedule without the user being present.

Why it was flagged

The cron command is a disclosed, user-directed setup example, but it would persistently trigger the full publishing workflow every day.

Skill content
配合 OpenClaw cron 实现每日自动执行 ... cron add --name "AI科技资讯工厂"
Recommendation

Use cron only after adding publish confirmations, draft mode, monitoring, and an easy disable path.