Security audit

Web Verify Protocol

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed web-research verification workflow with no code execution, persistence, credential use, or hidden data handling found.

Installers should expect this skill to make fact-checking tasks slower and more search-heavy. It is reasonable for publication-grade verification, but users should not treat the protocol's confidence claims as a guarantee that all facts are correct.

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 (2)

Vague Triggers

Medium

Confidence: 94% confidence
Finding: The skill metadata advertises a very broad set of triggers and an expansive description covering almost any task involving web verification, fact-checking, or data reliability. This can cause the skill to activate in many loosely related contexts, leading to unnecessary protocol enforcement, workflow hijacking, or suppression of more appropriate task-specific skills.

Vague Triggers

Medium

Confidence: 92% confidence
Finding: The 'When to Use' section uses broad language such as 'any AI-generated data task' and 'when users emphasize reliability,' which makes activation criteria subjective and difficult to bound. In a multi-skill agent, this can cause unintended invocation across common writing, research, and analysis flows, potentially overriding normal execution paths and creating denial-of-service-like friction by forcing a heavy verification process.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.