Security audit

测评设计师

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed exam-design helper with limited markdown instructions and proportionate, mostly aggregate education-data handling.

Before installing, confirm you are comfortable with the skill using assessment context from related teaching skills and writing back aggregate score-rate or item-quality data. It should not be used to publish named student results, rankings, or copied copyrighted exam questions.

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 (2)

Vague Triggers

Medium

Confidence: 89% confidence
Finding: The metadata description uses very broad trigger phrases such as '帮我出一份试卷' and '设计一个单元测验' without clear gating conditions, so a router or agent may invoke this skill for many generic teacher requests. Over-invocation can cause unintended access to adjacent workflows and unnecessary data handling, especially because the skill is designed to read from and write to other education-analysis skills.

Vague Triggers

Medium

Confidence: 92% confidence
Finding: The trigger section enumerates many common teacher utterances but does not define negative examples or hard boundaries for when the skill should not activate. In a multi-skill environment, this ambiguity increases the chance that routine requests are routed here, which can trigger unnecessary exam-design behavior and downstream sharing of aggregated performance data with student-analyzer.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.