Ai Friendly Architecture Design Skill
Security checks across malware telemetry and agentic risk
Overview
This skill is a coherent architecture-guidance package with a manual eval script and no evidence of hidden data access or unsafe behavior.
Installers should understand this is advisory architecture guidance, not a runtime AI framework. Review its recommendations before applying them to high-stakes systems, especially suggestions involving autonomous agents, persistent memory, or external model providers, and run the included eval script only if you want to test the skill behavior.
SkillSpector
By NVIDIA
Vulnerability Patterns
- 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
- Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
VirusTotal
63/63 vendors flagged this skill as clean.