V19 Trust Engine

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:44 PM UTC
• Version:
  1.0.0

Overview

This documentation-only skill explains a user-run external trust-scoring API and does not install code or perform hidden actions.

Install this only if you intend to use the V19 governance service. Before running the curl examples, verify the endpoint is one you trust, avoid submitting sensitive agent details, and keep any returned Pro key private.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (2)

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The skill embeds a real governance API key directly in example commands and does not explain its scope, rate limits, logging, or reuse implications. Even if intended as a public/read-only key, publishing reusable credentials in operational examples encourages blind reuse, normalizes secret leakage patterns, and can enable unauthorized consumption or telemetry generation against the remote service.

Missing User Warnings

Medium

• Confidence: 91% confidence
• Finding: The self-registration example instructs users to POST agent data to an external endpoint and states that a Pro key will be returned, but it omits any notice about data transmission, retention, trust boundaries, or the sensitivity of the returned credential. This can trick users or downstream agents into disclosing identifiers and mishandling newly issued keys without informed consent or secure storage practices.

VirusTotal

62/62 vendors flagged this skill as clean.