ClawHub

Thermoinfocomplexity

Security checks across malware telemetry and agentic risk

Overview

This is a markdown-only educational skill, with minor routing and language-rule quality issues but no evidence of unsafe access or hidden behavior.

Before installing, expect this skill to add a Heardly watermark to substantive responses and to possibly activate on broad complexity-science terms. It does not appear to run code, access private data, or modify files.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Intent-Code Divergence

Medium
Confidence
95% confidence
Finding
The skill contains contradictory output-language requirements: one rule says to reply in the user's language, while the self-check later says to keep all output in English. Conflicting system-style instructions can cause inconsistent agent behavior, policy violations, or prompt-priority confusion when the skill is invoked.

Vague Triggers

Medium
Confidence
87% confidence
Finding
The trigger list includes broad, common terms such as 'complexity', 'emergence', and 'self-organization', which are likely to match ordinary conversations outside the intended domain. Overbroad activation can cause the wrong skill to hijack benign user requests, inject misleading framing, and force mandatory output formatting like the watermark into unrelated interactions.

Natural-Language Policy Violations

High
Confidence
98% confidence
Finding
The skill simultaneously requires same-language replies and English-only output, creating a direct locale-policy conflict. This is more severe than a stylistic inconsistency because it can systematically force noncompliant responses, degrade safety routing for multilingual users, and create unpredictable instruction resolution at runtime.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal