Security audit

文曲·文库

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed research-collection workflow that searches, downloads, and stores writing materials with user confirmation and bounded persistence.

Install this only if you want an agent to build a reusable writing-material library. Expect it to use web search, fetch public pages, install a pinned CLI after approval, and write material indexes under your project and home directory; confirm the collection scope and any browser-runtime download before proceeding.

SkillSpector

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: 87% confidence
Finding: The skill advertises activation for very broad phrases such as “web scraping”, “save this webpage”, and generic research-collection requests. That makes unintended invocation plausible, which is risky because this skill can install a CLI, perform network searches, download remote content, and write files after only loosely matching user intent. In this context, accidental activation increases the chance of unnecessary external access and side effects.

Vague Triggers

Medium

Confidence: 83% confidence
Finding: The activation table defines broad categories like “收集素材”, “整理资料”, and “抓取这个网页” without stating when the skill should not activate or when a lighter-weight skill/tool should be preferred. Because this skill’s workflow includes search, download, storage, and possible CLI setup, ambiguous triggering can escalate a simple user request into persistent storage and networked collection activity the user did not specifically intend.

Static analysis

No suspicious patterns detected.