ClawHub
Back to skill

Security audit

数据

Security checks for vulnerabilities and agentic risk

Overview

The skill is a broad data-processing helper, but it asks for powerful file and command access while making security and command-whitelist claims that are not backed by the artifact.

Review this before installing if you would use it on private, business, or regulated data. Only enable it in workspaces where broad file access and shell commands are acceptable, and do not rely on its encryption, redaction, access-control, or command-whitelist claims unless your agent platform independently enforces those protections.

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 makes strong security claims about encrypted transmission, encrypted storage, access control, and safe handling, but the file only documents behavior and exposes powerful capabilities such as exec/read/write without any concrete enforcement mechanism. This can mislead users or downstream agents into trusting the skill with sensitive data under false assumptions, increasing the chance of unsafe use and data exposure.

Intent-Code Divergence

High
Confidence
97% confidence
Finding
The document states that only whitelisted commands are executed, but the declared toolset includes generic exec and the file shows no whitelist, validation, or command restriction mechanism. In a skill that processes user-supplied content, this mismatch is dangerous because users or agents may assume command execution is constrained when it may actually permit arbitrary shell commands.

Vague Triggers

Medium
Confidence
79% confidence
Finding
The invocation guidance is broad and positions the skill for general data analysis, reporting, visualization, and automation without tight trigger constraints or limits on when exec should be used. Because the skill also exposes read/write/exec/glob, vague activation criteria can cause over-invocation in unrelated contexts and expand the chance that sensitive files are accessed or commands are run unnecessarily.

Static analysis

No suspicious patterns detected.