Auth

Security checks across malware telemetry and agentic risk

Overview

This is a documentation-only authentication guide with no executable behavior, though one OAuth example should not be copied into production as-is.

Safe to install as an authentication reference, but treat all snippets as templates. Do not use the OAuth account-linking example as-is; require trusted verified-email claims and issuer validation, or require the user to log in to the existing account before linking a provider.

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 (2)

Intent-Code Divergence

High

Confidence: 98% confidence
Finding: The account-linking example links an OAuth identity to an existing local account based only on email equality, while later treating provider email as implicitly verified. This is dangerous because some providers may return unverified or weakly verified email claims, allowing account takeover if an attacker authenticates with a provider account that presents the victim’s email address.

Missing User Warnings

Medium

Confidence: 96% confidence
Finding: The guidance shows automatic account linking by email without warning that email alone is not a sufficient proof of control across identity providers. In an authentication skill, this omission is especially dangerous because developers may copy the pattern directly, leading to unauthorized account linking and possible account takeover.

VirusTotal

62/62 vendors flagged this skill as clean.

View on VirusTotal