ClawHub
Back to skill

Security audit

视觉文件整理专业版

Security checks for vulnerabilities and agentic risk

Overview

The skill has a clear file-sorting purpose, but it needs review because it can recursively reorganize many local files and enable scheduled/history behavior without tight safeguards.

Review before installing. Use it only on explicit target directories, run dry-run previews first, avoid enabling scheduled cleanup until paths and rules are confirmed, and make sure important folders are backed up before batch rename or move operations.

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Intent-Code Divergence

Medium
Confidence
93% confidence
Finding
The skill claims compatibility with a 'no deletion' safety rule, but elsewhere explicitly describes support for delete operations via create/query/modify/delete modes. This inconsistency can mislead users or an agent into trusting the skill's safety guarantees while invoking destructive filesystem behavior.

Vague Triggers

Medium
Confidence
88% confidence
Finding
The trigger conditions are broad enough to capture generic file handling, document conversion, and content extraction requests beyond the skill's narrowly defined sorting use case. In an agent environment with read/write/exec tools, overbroad routing increases the chance the skill is invoked on unintended tasks and performs unsafe filesystem operations.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill describes bulk organization, recursive directory processing, report export, and rollback over real filesystem paths, but it does not present a prominent warning that these operations can move, rename, overwrite, or otherwise affect large sets of user files. In a tool-enabled agent, that omission materially raises the risk of accidental data loss, corruption, or unintended modification at scale.

Static analysis

No suspicious patterns detected.