ClawHub
Back to skill

Security audit

警察执法资格考试助手

Security checks across malware telemetry and agentic risk

Overview

This appears to be a normal exam-study helper with accuracy and broad-trigger caveats, not evidence of hidden control, data theft, or destructive behavior.

Install only if you are comfortable with a Python-based study helper that uses bundled exam data and includes a web article-fetching module. Treat exam and legal content as unofficial, verify important requirements against official sources, and avoid feeding the web-fetch helper internal or private URLs.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (1)

Missing User Warnings

Low
Confidence
87% confidence
Finding
fetch_weixin_article accepts an arbitrary URL and performs a live HTTP request with no validation, allowlist, or warning. In an agent or plugin context, this can enable server-side request forgery behavior, allowing the skill to reach internal services, cloud metadata endpoints, or other unintended network targets if user-controlled input reaches this method.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal