央视新闻抓取(免费版)

• Outcome: Review
• Latest audit: Aug 17, 2026 · 1:29 AM UTC
• Version: 1.0.4

Overview

The skill is mostly a CCTV news-fetching guide, but it asks for command execution and includes overbroad activation language plus an unsafe remote install command.

Review this skill before installing. Use it only for CCTV news queries, avoid vague invocations, prefer an already installed Node.js runtime or a verified package-manager install path, and do not run the curl-to-bash Bun installer unless you have independently reviewed and trust it. Treat local JSON/cache examples as persistent files you may need to manage or delete.

SkillSpector

Vulnerability Patterns

• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands

Findings (7)

Context-Inappropriate Capability

Medium

• Confidence: 89% confidence
• Finding: The caching example writes fetched data to arbitrary local paths under a caller-controlled cache directory, adding persistent file-write behavior beyond the skill's narrow stated purpose. In an agent environment with exec access, unnecessary write capabilities increase the blast radius by allowing unreviewed local persistence of retrieved content, accidental overwrites, or storage in sensitive/shared locations.

Intent-Code Divergence

Medium

• Confidence: 88% confidence
• Finding: The skill is presented as a narrowly scoped CCTV news fetcher, but the capability summary advertises generic file handling, API integration, and command execution. In agent ecosystems, overstating broad capabilities can cause over-invocation and justify use in contexts far beyond the intended task, increasing the chance of unnecessary exec/file operations on unrelated user requests.

Vague Triggers

Medium

• Confidence: 91% confidence
• Finding: The activation language is overly broad, covering generic marketing copy, writing, title optimization, and content creation tasks rather than a narrowly bounded news-fetch workflow. In an agent setting, such broad triggers can cause the skill to activate on unrelated prompts and gain access to exec/read tools in contexts where they are unnecessary, expanding attack surface.

Vague Triggers

Medium

• Confidence: 90% confidence
• Finding: The example trigger phrase '执行核心功能' is so vague that it provides no meaningful boundary for when the skill should run. Ambiguous invocation criteria increase the risk that an orchestrator or user will trigger exec-capable behavior unintentionally during unrelated conversations.

Missing User Warnings

Low

• Confidence: 77% confidence
• Finding: The quick-start instructions include shell redirection that creates a local JSON file, but the document does not clearly warn users that running the command will persist data to disk. In shared or sensitive environments, silent file creation can lead to unintended local data retention and operational surprises.

Missing User Warnings

Low

• Confidence: 86% confidence
• Finding: The caching example persistently stores fetched news data without clearly informing the user that content will be written to disk and reused later. Even if the data is not highly sensitive, unannounced persistence is risky in agent environments because it can create data retention, sharing, and cleanup issues.

External Script Fetching

High

• Category: Supply Chain

Content

| 依赖项 | 类型 | 是否必需 | 获取方式 |
|:-------|:-----|:---------|:---------|
| Node.js 16+ | 运行时 | 二选一 | 官网下载安装 |
| Bun 1.0+ | 运行时 | 二选一 | `curl -fsSL https://bun.sh/install \| bash` |
| node-html-parser | npm包 | 必需 | `npm install node-html-parser` 或 `bun add node-html-parser` |
| Python 3.8+ | 运行时 | 可选 | 辅助脚本使用 |
| LLM API | API | 必需 | 由Agent平台内置LLM提供 |

• Confidence: 98% confidence
• Finding: The documentation recommends piping a remote script directly into bash to install Bun. This is a well-known high-risk pattern because it executes code fetched over the network without pinning, integrity verification, or local review; if the remote source or connection is compromised, arbitrary code execution follows immediately.

Static analysis

No suspicious patterns detected.