Fast Fact-Check

Security checks across malware telemetry and agentic risk

Overview

This fact-checking skill is coherent and purpose-aligned, with no artifact evidence of hidden data access, persistence, exfiltration, or unsafe authority.

Install this if you want a bounded fact-checking workflow that uses web search and citations. Expect it to make external web queries for the facts you ask about, and treat its optional validator as a structural citation check rather than proof that a source semantically supports every claim.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Intent-Code Divergence

Medium
Confidence
98% confidence
Finding
This is a real integrity issue: the answer and evidence cite source [2], but only source [1] is actually provided, so the claimed verification cannot be checked. In a fact-checking skill whose purpose is to return fast, source-backed answers, dangling citations materially undermine trust and can let unsupported or incorrect claims appear verified.

VirusTotal

61/61 vendors flagged this skill as clean.

View on VirusTotal