Brake

Security checks across malware telemetry and agentic risk

Overview

This is a straightforward brake CAD assistant that uses a disclosed external website, with privacy and scoping caveats users should understand.

Install only if you are comfortable sending brake design parameters to jixietools.com and receiving a guest progress link there. Treat the generated URL and guest code like a private access link, review the product selection carefully because of the category ID inconsistency, and stop the workflow before production-sheet creation if you do not want remote job creation or polling.

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 (3)

Intent-Code Divergence

Medium
Confidence
96% confidence
Finding
The skill metadata states the brake category ID is 24, but the actual listing request uses category_id=8. This inconsistency can cause the agent to fetch and present the wrong product set, leading to unintended actions or disclosure of unrelated product data; because the rest of the workflow can proceed unauthenticated, the mismatch increases the chance of operating on the wrong backend objects without user awareness.

Vague Triggers

Medium
Confidence
85% confidence
Finding
The trigger phrases include very broad terms such as "制动器" and "brake", which can appear in normal conversation outside the intended CAD-generation workflow. That can cause accidental invocation of a skill that immediately begins collecting parameters and interacting with external systems, creating a risk of unintended data transmission or order creation.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill creates guest-accessible production sheets and repeatedly surfaces a public URL/guest code without clearly warning the user that possession of that link may grant access to sensitive design artifacts and status data. In a CAD/manufacturing context, those links may expose proprietary specifications, making the omission of privacy guidance materially risky.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal