ClawHub
Back to skill

Security audit

Fanbozhe Clean

Security checks across malware telemetry and agentic risk

Overview

This is a prompt-only coaching skill that changes conversational style for business decision review and does not request system access or hidden capabilities.

Install this if you want a direct, skeptical decision-review persona. Be aware that phrases like “red team” or “second opinion” may make the assistant more blunt; invoke it explicitly when you want that mode.

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
92% confidence
Finding
The trigger list includes broad natural-language phrases such as '红队' and '第二意见' that can appear in ordinary conversation, making accidental activation plausible. In this skill, unintended activation matters because it switches the assistant into a harsher contrarian mode that can change tone, decision support style, and user experience without clear consent.

Vague Triggers

Medium
Confidence
95% confidence
Finding
The opponent-mode activation logic is underspecified: it lists example phrases but does not define matching rules, precedence, or whether explicit user intent is required. That ambiguity can cause the assistant to adopt an aggressive, high-certainty rebuttal style in contexts where the user was merely discussing criticism, red teaming, or a second opinion rather than requesting this mode.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.