ClawHub
Back to skill

Security audit

Public Opinion Scanner

Security checks across malware telemetry and agentic risk

Overview

This skill appears to be a legitimate public-opinion research tool, but it can profile named individuals and persist reports without clear privacy safeguards or explicit save control.

Review this skill carefully before installing. Use it only for lawful, legitimate public-interest or authorized reputation analysis, avoid targeting private individuals without a valid reason, and check where reports are saved so sensitive names, links, or summaries are not left in shared or long-lived workspace storage.

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 (2)

Missing User Warnings

Medium
Confidence
86% confidence
Finding
The skill instructs that reports are '默认保存' to a specific workspace path, which creates files as a side effect without making that behavior prominent at the start of the skill or clearly requiring user confirmation. This can surprise users, leak sensitive query subjects into persistent storage, and leave reputation-analysis artifacts on disk that other tools or users may later access.

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The skill explicitly supports scanning the public reputation of named individuals ('个人舆情扫描器') and aggregating potentially sensitive personal information from multiple sources, but provides no privacy warning, lawful-use limitation, or caution against misuse. In context, this makes profiling, harassment, doxxing-adjacent monitoring, and unfair reputational assessment easier, especially because the output is structured, persistent, and link-preserving.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.