ClawHub
Back to skill

Security audit

Game Theory

Security checks across malware telemetry and agentic risk

Overview

This is a documentation-only game theory skill for crypto analysis, with dual-use attack examples framed mainly for understanding and defense.

Safe to install as a reference skill. Treat the governance and MEV attack material as defensive threat-modeling content for protocol review and hardening, not as guidance for manipulating real communities, markets, or protocols.

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

Medium
Confidence
94% confidence
Finding
This section gives explicit attacker guidance such as how to accumulate quietly, time proposals, divide opposition, shape narratives, and extract incrementally, without an adjacent warning that the content is for defensive analysis only. In a skill that may be surfaced to users for strategic advice, this can lower the barrier for governance abuse by turning abstract threat modeling into actionable offensive playbooks.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.