时事日报分析

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:25 PM UTC
• Version:
  1.0.1

Overview

This skill is a disclosed news-report workflow that uses public web research guidance and local PDF generation, with quality caveats but no hidden or unsafe behavior evident.

Reasonable to install for guided news-report generation. Before using it, confirm when the assistant will search the web and where it will write Markdown/PDF files, install `reportlab` from a trusted source, and review the final report/PDF because the verification helper and PDF script have quality limitations.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• 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
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (3)

Intent-Code Divergence

Medium

• Confidence: 87% confidence
• Finding: The docstring states that key information requires confirmation from more than two Tier 1-2 sources, but the implementation marks Tier 3-4 items as verified when only one qualifying Tier 1-2 source matches keywords. In a news-reporting workflow, this weakens the stated verification standard and can cause unverified or weakly corroborated information to be treated as verified, increasing the risk of misinformation in generated reports.

Vague Triggers

Medium

• Confidence: 86% confidence
• Finding: The README includes a very broad example trigger phrase, "Generate today's international news report," which could cause the skill to activate in situations where the user did not intend external web research or report generation. In this skill's context, activation can lead to web searches and local file/PDF creation by an AI agent, so ambiguous invocation increases the risk of unintended data access and actions.

Missing User Warnings

Medium

• Confidence: 91% confidence
• Finding: The README instructs the AI agent to use web search and run a PDF generation script, but it does not present a clear user-facing warning that external data will be accessed and files may be created on disk. In an agentic environment, missing disclosure can cause users to unknowingly trigger network activity or local side effects, reducing informed consent and increasing the chance of unsafe execution.

VirusTotal

65/65 vendors flagged this skill as clean.