ClawHub

Reply Styles

Security checks across malware telemetry and agentic risk

Overview

This is a coherent reply-drafting skill with optional local style helper scripts and no evidence of credential access, network activity, persistence, or hidden data handling.

Safe to install based on the inspected artifacts. Treat it as a drafting aid, review any customer-facing or public reply before sending, and specify the desired output language if you do not want Chinese output.

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 (3)

Vague Triggers

Medium
Confidence
89% confidence
Finding
The invocation triggers are broad enough to match common, everyday phrasing such as 'help me reply politely' or 'make this more tactful,' which can cause the skill to activate in situations where the user did not explicitly request this specialized behavior. That increases the risk of unintended routing, hidden prompt shaping, or overriding a user's preferred style/language, especially in multi-skill environments where trigger specificity matters.

Natural-Language Policy Violations

Medium
Confidence
95% confidence
Finding
The skill states its function in Chinese and frames behavior around generating replies in that language without offering user-language detection or choice. In practice, this can cause unwanted language coercion, degrade usability for non-Chinese users, and create prompt-conflict behavior where the skill output diverges from the user's requested language or system locale.

Natural-Language Policy Violations

Medium
Confidence
89% confidence
Finding
The file’s goal statement is written as a normative instruction in Chinese and frames behavior around Chinese-language output without any indication that this is conditional on the user’s language preference. In a reply-generation skill, hard-coding language behavior can cause undesired output, override user intent, and create prompt-level steering that propagates across channels.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal