ClawHub
Back to skill

Security audit

新闻聚合引擎(免费版)

Security checks across malware telemetry and agentic risk

Overview

This is a markdown-only news aggregation skill whose disclosed behavior fits its purpose, though its tool scope and input wording are broader than ideal.

Install only if you are comfortable with an agent using network search, and possibly command execution, to gather news. Prefer giving narrow topics, date ranges, and source constraints, and avoid sending private or sensitive information through the optional callback URL.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Vague Triggers

Medium
Confidence
90% confidence
Finding
The description explicitly says the skill can '自发搜索' and '自发提升工作效率', which implies autonomous activation or broad self-directed behavior without clear user-trigger boundaries. In an agent environment with read and exec tools, this can cause the skill to initiate network-like data gathering workflows or shell-assisted fetching beyond what the user specifically requested.

Vague Triggers

Medium
Confidence
92% confidence
Finding
The input schema defines `input` as arbitrary free-form '处理的输入数据或指令', which permits unbounded instruction injection through the skill interface. Because the skill has `exec` and `read` tools and is designed to search and process external content, broad free-form instructions can be abused to steer the agent into unintended command execution or unsafe retrieval behavior.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The front matter advertises automation and includes the `exec` tool, but the markdown does not warn users that the skill may execute commands or interact with external sites/sources as part of operation. This reduces informed consent and increases the chance that a user invokes the skill without understanding that it may perform higher-risk actions in the agent runtime.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.