ClawHub
Back to skill

Security audit

Gejun Math Coach

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed, instruction-only math tutoring skill; its main issues are optional voice command use and a worked-example math inconsistency, not malicious behavior.

Reasonable to install for math study support. Verify important answers independently, especially worked examples, and only enable the optional voice feature if you are comfortable with the agent running the shown local TTS/playback commands for current tutoring text.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (3)

Context-Inappropriate Capability

Medium
Confidence
93% confidence
Finding
The skill embeds executable shell commands for TTS generation and local audio playback, which extends a math-coaching skill into command-execution behavior. Even though the stated purpose is accessibility/voice coaching, instructing an agent to run Python and `afplay` on the host increases attack surface and can be abused if any part of the spoken text, file path, or surrounding execution flow becomes user-influenced.

Intent-Code Divergence

Medium
Confidence
96% confidence
Finding
The file presents conflicting mathematical answers for the same problem: the S2 example states the tangent line is y = -a²x, while the S7 walkthrough correctly derives y = -ax. In an educational skill that claims answer-verification and correctness guarantees, inconsistent worked solutions can mislead users, degrade trust, and propagate incorrect learning outcomes.

Vague Triggers

Medium
Confidence
88% confidence
Finding
The trigger list is very broad and includes common educational terms such as '学习方法', '教教我', and '复习', which can cause the skill to activate outside the user’s intended scope. Over-broad activation is dangerous because it can unexpectedly steer unrelated conversations into this skill’s workflow, including optional external integrations and command-bearing voice behavior described elsewhere in the file.

VirusTotal

62/62 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.