Security audit

Xiao Chuang You Literature

Security checks across malware telemetry and agentic risk

Overview

This is a low-impact Chinese literature response-style skill with broad triggers but no evidence of code execution, data access, persistence, or hidden behavior.

Safe to install if you want a Chinese literature-oriented style for poetry, reading, book recommendations, and calligraphy questions. Be aware that broad trigger words may make it activate during ordinary literature-related conversations, so narrow or disable it if that becomes distracting.

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: 88% confidence
Finding: The skill description activates on very broad literature-related topics such as reading, poetry, and book recommendations, which are common in ordinary conversation. This can cause the skill to trigger outside narrowly intended contexts, increasing the chance of irrelevant instruction injection or unintended takeover of general chat behavior, though the content itself is not directly harmful.

Vague Triggers

Medium

Confidence: 93% confidence
Finding: The trigger list includes highly generic words like “诗”, “词”, “文学”, “阅读”, and “读到”, which can appear in many unrelated conversations. Overbroad triggers can cause excessive skill activation, making the agent more susceptible to prompt-scope confusion and reducing reliability by applying this skill when it was not requested.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.