ClawHub
Back to skill

Security audit

Structs Reconnaissance

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed Structs game reconnaissance helper that queries game data and saves local intel notes, with no executable code or hidden sharing behavior found.

Install this if you want an agent to gather Structs game intelligence and maintain local memory/intel notes. Expect it to write or update those local files, and use it cautiously if you prefer reconnaissance results to remain temporary or if your local Guild Stack database contains sensitive game intelligence.

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 (3)

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The skill metadata explicitly states that reconnaissance findings will be persisted to `memory/intel/`, but it does not warn the user that running the skill will modify local persistent data. This can cause silent retention of sensitive intelligence, create privacy or operational-security issues, and surprise users who expected read-only scouting behavior.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The procedure instructs the agent to update multiple dossier and threat-tracking files as part of normal execution, again without disclosure or consent for persistent storage. Because this is operational reconnaissance data about players, guilds, and threats, silent writes may accumulate sensitive or strategically valuable information on disk and could alter user workspace state in unexpected ways.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The error-handling and persistence loop direct the agent to create directories and repeatedly write or update intelligence files, which increases the risk of unnoticed, ongoing modification of local data. This is more dangerous in context because the skill is framed as reconnaissance and competitive intelligence, so the persisted content may be sensitive and the repeated update loop normalizes continuous state changes without user awareness.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.