ClawHub
Back to skill

Security audit

一人MCN---1个Skill撑起一家自媒体公司

Security checks across malware telemetry and agentic risk

Overview

This is a content-creator assistant skill that gives planning, writing, SEO, analytics, pricing, and scheduling advice without hidden execution, persistence, or account-control behavior.

Installers should expect this skill to respond to ordinary creator-related phrases and produce potentially long content-workflow advice. Review business-pricing, contract, platform-rule, and trend recommendations before relying on them, since the artifact itself notes that market numbers and platform rules may be outdated and should be verified.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Vague Triggers

Medium
Confidence
86% confidence
Finding
The README advertises very broad natural-language trigger phrases such as asking for topic ideas or a full content workflow, which increases the chance the skill will activate from ordinary user phrasing without clear intent boundaries. In an agent ecosystem, this can cause unintended invocation, over-broad automation, or accidental chaining of modules when the user did not explicitly mean to run the full skill.

Vague Triggers

High
Confidence
95% confidence
Finding
The skill description advertises activation via very broad everyday phrases such as '帮我想选题', '怎么拍视频', and '接广告多少钱', which are common user utterances that may appear in unrelated conversations. This creates a real risk of unintended invocation, context hijacking, and surprise behavior, especially because the skill claims it can auto-route and chain multiple modules from natural language alone.

Vague Triggers

High
Confidence
97% confidence
Finding
The routing logic explicitly prioritizes broad semantic matching and keyword fallbacks across many generic intents, allowing ordinary creator-related phrases to trigger the skill without strong disambiguation. Because it also supports automatic module chaining, an accidental match could steer the agent into producing substantial outputs or taking over the conversation flow when the user did not intend to invoke this skill.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.