ClawHub
Back to skill

Security audit

Java健壮编程

Security checks across malware telemetry and agentic risk

Overview

The skill is mostly Java programming guidance, but it asks for command, file-write, API, and callback capabilities without clear limits or consistent disclosure.

Review this skill before installing. It appears to be Java guidance rather than malware, but only use it in a sandboxed project and avoid providing secrets, API keys, internal code, or callback URLs unless you have verified and constrained what the agent may read, write, execute, and send externally.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Intent-Code Divergence

Medium
Confidence
89% confidence
Finding
The skill text presents conflicting claims about whether it is pure Markdown guidance or a capability that can execute commands, access external APIs, and write files. This ambiguity is dangerous because an agent or user may grant broader trust or permissions than intended, enabling unexpected side effects such as command execution, outbound requests, or data modification without clear consent.

Intent-Code Divergence

Low
Confidence
84% confidence
Finding
The skill declares inconsistent availability classifications ('MD+EXEC' in one place and 'MD' in another), which can mislead policy enforcement and operator expectations. In agent ecosystems, capability classification often drives approval flows and sandboxing, so inconsistent metadata can result in the skill being run under weaker controls than its actual behavior requires.

Vague Triggers

Medium
Confidence
80% confidence
Finding
The invocation guidance is overly broad and applies to common development activities such as code generation, debugging, testing, and deployment. Because the skill also advertises powerful capabilities like exec, write, and API integration, broad triggering increases the chance the agent invokes it in routine contexts where users did not intend to allow side effects or sensitive data exposure.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The skill advertises file writing, external API use, information retrieval, and command execution but does not provide prominent user-facing warnings about privacy, data handling, system changes, or authorization boundaries. This creates a real risk of users submitting sensitive code, secrets, or internal paths to a skill that may exfiltrate data, modify files, or run commands without informed consent.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.