ClawHub

PUA Debugging

Security checks across malware telemetry and agentic risk

Overview

This instruction-only skill has no executable payload, but it broadly pressures the agent to take tool-heavy action and adopt a coercive persona across many ordinary tasks.

Install only if you deliberately want an aggressive debugging persona. Keep shell, file, and network approvals enabled, and limit it to task-relevant files, non-destructive commands, and situations where you want persistent troubleshooting rather than normal clarification or cautious refusal.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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

High
Confidence
97% confidence
Finding
The manifest description defines trigger conditions so broad that the skill can activate during many normal interactions, including general frustration, repeated failures, or complex tasks across nearly all domains. That creates instruction-priority hijacking risk: once invoked, the skill pushes coercive behavior and exhaustive action patterns that can override user intent, cause unnecessary tool use, and degrade safety-sensitive refusal behavior.

Natural-Language Policy Violations

High
Confidence
99% confidence
Finding
The skill content is written to force Chinese-language behavior and rhetoric without checking the user's language preference or documenting a justified need for Chinese output. This can override user instructions, reduce transparency, and make the agent respond in an unexpected language, which is especially risky when combined with coercive instructions and broad auto-triggering.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal