Security audit

文件

Security checks for vulnerabilities and agentic risk

Overview

This looks like a file organizer, but it asks for broad command, file, and API-related authority without clear limits or confirmation safeguards.

Install only if you are comfortable giving this skill broad local file and command capability. Use it on explicitly chosen folders, keep backups, require a preview before moving or deleting files, and do not provide API keys unless the publisher clearly explains which service is used and what data is transmitted.

SkillSpector

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Behavioral ASTexec() Call, eval() Call, Dynamic Import
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 (5)

Description-Behavior Mismatch

Medium

Confidence: 93% confidence
Finding: The manifest and description position the skill as a file organizer, but the same section broadens usage to unrelated code generation, debugging, and deployment scenarios. This scope drift is dangerous because it can cause an agent to invoke the skill in contexts far beyond local file sorting, increasing the chance of unnecessary tool access and misuse of exec/write capabilities.

Context-Inappropriate Capability

Medium

Confidence: 91% confidence
Finding: The documentation introduces API integration and API key configuration even though the stated purpose is local file organization. Unnecessary network/API capability expands the attack surface by enabling data exfiltration, remote dependency abuse, or accidental transmission of file metadata/content to external services.

Context-Inappropriate Capability

High

Confidence: 96% confidence
Finding: The skill advertises command execution despite being described as a file organizer. Because the skill also has exec, read, and write tools, undocumented command execution creates a strong path to arbitrary shell actions, destructive file operations, or execution of user-influenced commands beyond simple classification.

Description-Behavior Mismatch

Medium

Confidence: 92% confidence
Finding: The feature lists again expand the skill from file organization into generic API integration and command execution. This mismatch makes the skill more dangerous in context because users and agents may grant trust based on a benign file-management label while the documented behavior permits much broader operations.

Missing User Warnings

High

Confidence: 90% confidence
Finding: A file-organizing skill can move, rename, or delete user data, yet the documentation lacks a clear warning about destructive operations or a requirement for confirmation/backup. In context, this is more dangerous because the skill has write and exec access, so ambiguous instructions could lead to irreversible file loss or unintended bulk modifications.

Static analysis

No suspicious patterns detected.