Agent Testing & QA Toolkit: Integration, Chaos, and Contract Testing for Multi-Agent Systems

• Outcome: Pass
• Latest audit: May 29, 2026, 8:37 PM
• Version:
  1.3.1

Overview

This is a non-executing testing guide, but copied examples could affect real GreenHelix accounts or funds if run with production credentials.

Treat this as a guide, not trusted automation. Keep examples in sandbox or staging by default, never record cassettes with production keys or secrets, review cassette diffs before committing, and require explicit approval before production deploys, canary rollouts, escrow releases, or wallet-affecting tests.

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 (1)

Missing User Warnings

Medium

• Confidence: 81% confidence
• Finding: The guide includes deployment and environment-key patterns that use live API keys and financially meaningful operations, but it does not place a strong, repeated warning at the point of use that non-sandbox execution may affect real funds. In a commerce/escrow context, readers may copy these snippets into CI/CD or staging with production-like credentials, causing unintended spend or live account changes.

VirusTotal

64/64 vendors flagged this skill as clean.