Security audit

Daily News Fetcher

Security checks across malware telemetry and agentic risk

Overview

This is a simple news-summary skill that fetches public RSS headlines when invoked, with some documentation and language-default caveats but no evidence of hidden data access or unsafe behavior.

Install this if you want a basic headline fetcher and are comfortable with it contacting BBC, Xinhua, and NYTimes RSS feeds when you ask for news. Treat it as a simple feed formatter, not a translator or Reuters-backed ranking tool, unless the publisher updates the documentation and code.

SkillSpector

By NVIDIA

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration

Findings (5)

Lp3

Medium

Category: MCP Least Privilege
Confidence: 95% confidence
Finding: The skill instructs the agent to run a network-capable script, but the manifest shown in SKILL.md does not declare any corresponding permission or capability boundary. This creates a transparency and policy-enforcement gap: users and the platform may not realize the skill performs external fetching, and permission checks may be bypassed or inconsistently applied.

Vague Triggers

Medium

Confidence: 90% confidence
Finding: The phrase "any request for current news updates" is an open-ended trigger condition that broadens activation beyond clearly defined user intents. This can cause overmatching and unexpected invocation of a networked script in contexts where the user may only be discussing news abstractly rather than requesting a fetch.

Vague Triggers

Medium

Confidence: 90% confidence
Finding: The phrase "any request for current news updates" is an open-ended trigger condition that broadens activation beyond clearly defined user intents. This can cause overmatching and unexpected invocation of a networked script in contexts where the user may only be discussing news abstractly rather than requesting a fetch.

Vague Triggers

Medium

Confidence: 93% confidence
Finding: The explicit trigger list again includes the generic term "news," which is too broad to safely distinguish intentional skill use from ordinary conversation. In a skill that performs external retrieval, broad triggers materially increase the risk of unintended execution.

Natural-Language Policy Violations

Medium

Confidence: 84% confidence
Finding: Forcing Chinese-language output without checking user preference can produce misleading or unusable results for users expecting another language, which is a quality and consent issue. While not a direct code-execution risk, it can cause confusion and degrade trust, especially for a broadly triggered skill.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.