ClawHub
Back to skill

Security audit

对任意输入做结构化分析

Security checks across malware telemetry and agentic risk

Overview

The skill is mostly a Markdown analysis template, but its documentation overstates capabilities like file writing, external API use, and command execution without clear scope or controls.

Install only if you want a general structured-analysis prompt and you are comfortable policing its boundaries yourself. Do not let it run shell commands, write files, call external APIs, or process sensitive data unless you explicitly intend that for the task and have reviewed what will be sent or changed.

SkillSpector

Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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
Findings (4)

Intent-Code Divergence

High
Confidence
94% confidence
Finding
The manifest exposes only a read tool, but the documentation advertises file writing, API integration, and command execution. This mismatch can mislead users and downstream agents into believing the skill is authorized for higher-risk actions, weakening operator scrutiny and potentially encouraging unsafe capability escalation outside the declared manifest.

Context-Inappropriate Capability

Medium
Confidence
87% confidence
Finding
Claiming command execution capability in an analysis skill expands the implied authority of the skill beyond its stated purpose. Even if no exec tool is declared here, normalizing shell execution for broad 'analyze any input' workflows increases the risk that an agent or user may invoke dangerous commands on untrusted data under the assumption that such behavior is intended.

Context-Inappropriate Capability

Medium
Confidence
84% confidence
Finding
Advertising external API integration without clear scope or justification creates ambiguity about data flow and trust boundaries. In a broadly scoped analysis skill, this can lead users to provide sensitive inputs without understanding that data may be transmitted to third-party services.

Vague Triggers

Medium
Confidence
78% confidence
Finding
The skill is framed as suitable for 'any input' with very broad invocation language and minimal trigger constraints. Overbroad scope increases the chance that the skill will be applied to sensitive, adversarial, or out-of-scope tasks, which is especially risky when the same document also implies API and command-execution behavior.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.