ClawHub
Back to skill

Security audit

AI新闻工具-专业版

Security checks for vulnerabilities and agentic risk

Overview

This news-intelligence skill is mostly purpose-aligned, but it asks agents to use broad command execution, recurring alerts, and outbound webhooks/email without enough scoping or privacy controls.

Install only if you trust the external news-pro CLI and can control where alerts, webhooks, email, reports, API keys, and local indexes go. Limit this skill to news-intelligence tasks, review any scheduled push configuration before enabling it, and avoid sending sensitive monitored topics or internal reports to unapproved destinations.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Intent-Code Divergence

Medium
Confidence
90% confidence
Finding
The skill’s own security guidance warns against concatenating user input into command-line parameters, yet the example constructs a search query from variable content and passes it to the CLI. Even though the Python example uses subprocess.run with an argument list rather than a shell string, untrusted query content can still alter downstream CLI behavior or trigger unsafe parsing in the news-pro tool if that tool treats leading dashes, operators, or special syntax as control input.

Vague Triggers

Medium
Confidence
80% confidence
Finding
The trigger condition says to use the skill whenever data analysis, report generation, statistical insight, or visualization is needed, which is far broader than a news aggregation skill’s stated purpose. Overbroad activation guidance can cause an agent to invoke this skill in unrelated contexts, increasing the chance of unnecessary external data access, command execution, or transmission of user data to third-party services.

Vague Triggers

Medium
Confidence
78% confidence
Finding
The description mixes broad usage guidance with product marketing language and does not establish precise activation boundaries, making it ambiguous when the agent should invoke the skill. This can lead to inappropriate tool selection and unnecessary exposure of user prompts or data to external APIs and execution paths not needed for the task.

Missing User Warnings

Medium
Confidence
87% confidence
Finding
The skill includes callback_url, webhook, and email push examples that send data outward, but the usage sections do not clearly warn that news content, summaries, report attachments, or filters may contain sensitive user or organizational information. In an enterprise intelligence context, outbound notifications materially increase privacy and data-leakage risk because they can exfiltrate monitored topics, internal interests, or generated reports to third-party endpoints.

Static analysis

No suspicious patterns detected.