Security audit

AI Debate Coach - 吵架助手

Security checks across malware telemetry and agentic risk

Overview

This is a text-only Chinese debate coaching skill that helps analyze arguments and draft replies, with no hidden system access or installation behavior.

This skill is technically low risk to install. Review generated replies before sending them, especially in personal or workplace conflicts, and avoid using it for harassment, intimidation, or escalating arguments.

SkillSpector

By NVIDIA

Vulnerability Patterns

Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
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: 88% confidence
Finding: The activation instruction is broad enough to trigger whenever a user wants to argue, rebut, or analyze someone else’s position, without clear guardrails on scope or safety boundaries. In practice, this can cause the skill to engage in emotionally charged interpersonal conflicts too readily and amplify adversarial behavior, especially because it explicitly helps craft stronger rebuttals and a 'key_line' with high persuasive force.

Natural-Language Policy Violations

Medium

Confidence: 81% confidence
Finding: The skill description and content are written to operate in Chinese by default, with no user opt-in or documented locale restriction. This is mainly a policy and usability risk: it can override user language expectations, reduce transparency, and cause the skill to produce output the user cannot understand, but it does not by itself create a direct security compromise.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.