ClawHub
Back to skill

Security audit

提问质量优化师

Security checks across malware telemetry and agentic risk

Overview

This is an instruction-only question-improvement skill that broadly changes how answers are written, but it does not request code execution, credentials, file access, persistence, or external data transfer.

Install this only if you want the assistant to routinely critique, rewrite, expand, and answer questions in Chinese. Avoid or disable it when you want short direct answers, adaptive language behavior, or no automatic prompt enhancement.

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 (1)

Vague Triggers

High
Confidence
95% confidence
Finding
The skill is configured to trigger on 'any user question' with no domain, safety, or context boundaries, which makes it effectively intercept all interactions. In an agent environment, this broad activation can override normal routing, cause unintended processing of sensitive prompts, and amplify downstream unsafe behavior because the skill always rewrites and expands user input before answering.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.