Security audit

生产力增强工具（免费版）

Security checks for vulnerabilities and agentic risk

Overview

This is a generic productivity skill whose broad read/write/exec wording deserves caution, but the inspected artifact does not show hidden, deceptive, destructive, or purpose-mismatched behavior.

Install only if you want a general productivity automation helper. Before using it, give narrow instructions, specify which files or folders it may touch, and require confirmation before command execution, exports, deletions, or workflow changes.

SkillSpector

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Behavioral ASTexec() Call, eval() Call, Dynamic Import
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (2)

Vague Triggers

Medium

Confidence: 84% confidence
Finding: The skill’s examples and scope are broad enough to encourage the agent to perform loosely bounded planning, analysis, and automation actions without clear constraints on what data, files, or workflows may be touched. In an agent environment with read/exec/write permissions, ambiguous tasking can lead to unintended access, modification, or automation of sensitive local content even if the file does not contain explicitly malicious instructions.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The skill advertises use of read, exec, and write capabilities and repeatedly describes create/query/export style operations, but it does not provide strong warnings or approval requirements for system-impacting actions. In practice, this can normalize unsafe execution and file modification behavior, increasing the chance that an agent will run commands or alter files based on underspecified natural-language prompts.

Static analysis

No suspicious patterns detected.