Web Researcher

Security checks across malware telemetry and agentic risk

Overview

This is a simple web-research instruction skill that asks the agent to search and cite sources, with no hidden code or privileged behavior found.

Install only if you want the agent to perform live web research. Avoid including confidential information in research prompts, because search queries and fetched pages may reveal the topic to external providers.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Vague Triggers

Medium
Confidence
92% confidence
Finding
The skill recommends invocation on very broad phrases like "the latest," "news," and "research," which overlap with many ordinary user requests and can cause over-triggering. In practice, this can route unrelated prompts into a web-access workflow, increasing exposure to prompt injection, unreliable external content, and unintended tool use.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal