ClawHub
Back to skill

Security audit

AI全链路科技资讯工厂

Security checks for vulnerabilities and agentic risk

Overview

This is a coherent publishing workflow, but it can post to real social accounts and schedule repeated runs without clear approval safeguards.

Install only if you intend to connect real Xiaohongshu and WeChat publishing accounts. Use it step by step or in draft mode first, review every generated post before publishing, change outputs to a user-scoped directory, avoid enabling cron for live posting, and separately audit the dependent local skills that handle cookies, sessions, authorization, and platform actions.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The README explicitly describes automated posting to Xiaohongshu as part of the pipeline but does not warn users that the skill may perform external account actions, publish content publicly, or incur reputational/privacy consequences. In an agent setting, omission of clear consent and action-boundary warnings can lead users to trigger unintended real-world posting behavior.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The usage example instructs the agent to generate and publish content across multiple platforms in one request, normalizing cross-platform posting without any notice about external side effects or review requirements. This increases the risk of accidental public posting, leakage of sensitive or unvetted content, and misuse of connected social-media accounts.

Vague Triggers

Medium
Confidence
88% confidence
Finding
The trigger phrases are broad, generic, and overlap with ordinary user requests such as AI news, daily news, and content writing. This increases the chance of accidental invocation of a high-capability workflow that can fetch external content, generate materials, write files, and potentially publish to third-party platforms without the user explicitly intending to run this skill.

Missing User Warnings

High
Confidence
95% confidence
Finding
The skill explicitly advertises one-click multi-platform publishing, saving output under /root/articles, and scheduled unattended execution, but does not present clear user-facing consent and side-effect warnings. In this context, accidental or implicit activation could lead to unauthorized posting to Xiaohongshu/WeChat, persistent filesystem writes, and recurring automated actions, making the skill materially more dangerous than a read-only content helper.

Static analysis

No suspicious patterns detected.