ClawHub
Back to skill

Security audit

port-trans组合包

Security checks for vulnerabilities and agentic risk

Overview

This bundle is mostly transparent marketing documentation, but it combines broad local command/file authority with external API and Telegram actions without enough scoping or data-boundary guidance.

Install only if you are comfortable granting the bundled skills local file access, write access, command execution, and external API or Telegram connectivity. Avoid sending secrets or broad local files through the example API workflow, prefer HTTPS endpoints, and review each member skill's own instructions before enabling automation.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Context-Inappropriate Capability

High
Confidence
84% confidence
Finding
A marketing/composition bundle should not be instructing or implying shell command execution as part of its own functionality, especially when the document also advertises broad tools like exec, write, glob, and grep. This can normalize unsafe use patterns and encourage operators or downstream agents to treat untrusted input as command arguments, increasing risk of command injection or unintended local actions when the bundle is consumed automatically.

External Transmission

Medium
Category
Data Exfiltration
Content
headers = {"Authorization": "Bearer YOUR_API_KEY"}

# 执行技能
response = requests.post(api_url, json={
    "input": "示例输入数据",
    "options": {"mode": "advanced"}
})
Confidence
78% confidence
Finding
The example shows outbound transmission of user-supplied data to an API endpoint with bearer-token authentication, but it does not document trust boundaries, transport guarantees beyond a default insecure HTTP URL, data classification, or restrictions on what may be sent. In a skill context that advertises file handling and automation, this increases the chance that sensitive local content could be exfiltrated to a remote service through copy-pasted example usage or automated interpretation.

Static analysis

No suspicious patterns detected.