ClawHub

Chaos pivot

Security checks across malware telemetry and agentic risk

Overview

This is an instruction-only reasoning aid that helps an agent stop repeating failed approaches and try structured alternatives, without requesting credentials, persistence, or system access.

Reasonable to install as a thinking aid. Use extra supervision on high-stakes or irreversible tasks because it may cause the agent to change strategy and test alternate approaches sooner than you expect.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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
Findings (2)

Vague Triggers

Medium
Confidence
86% confidence
Finding
The activation criteria are broad and subjective, such as detecting uncertainty, repetition, or 'more steps than expected,' which can cause the skill to trigger in many normal workflows. In an agent system, this can disrupt task execution, induce unnecessary strategy changes, and create looping or unstable behavior that degrades reliability and may bypass more narrowly scoped safeguards.

Vague Triggers

Medium
Confidence
84% confidence
Finding
The manifest description says the skill triggers when an agent is 'stuck, looping, or failing repeatedly' without defining scope, thresholds, or safe task boundaries. Because manifest text may influence orchestration or skill selection, this broad trigger language increases the chance of unintended invocation across unrelated contexts, making agent behavior less predictable and potentially unsafe in sensitive workflows.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal