Hutian Opc Business Model

Security checks across malware telemetry and agentic risk

Overview

The skill is mostly coherent for business-plan analysis, but its generated HTML report loads remote JavaScript that could access sensitive report contents.

Install only if you are comfortable generating reports that may load third-party JavaScript when opened in a browser. For confidential business plans, prefer using the PNG output or modifying the skill to bundle ECharts locally and escape report fields before opening generated HTML.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • 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)

Intent-Code Divergence

Medium
Confidence
90% confidence
Finding
The skill simultaneously forbids fabrication and permits 'reasonable inference' when product-line data is missing, creating a contradictory instruction set. In a decision-support context for business plans, this can cause the agent to present speculative classifications as analysis, potentially misleading investment, strategy, or compliance decisions with unverified content.

Description-Behavior Mismatch

Medium
Confidence
90% confidence
Finding
The HTML report injects a remote script tag that loads ECharts from jsDelivr, creating an unexpected external network dependency in a tool that otherwise appears to do local file processing. If the CDN is unavailable, blocked, or the dependency is compromised, report rendering can fail or execute attacker-controlled JavaScript in the viewer's browser.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal