Security audit
Homelab Ai
Security checks across malware telemetry and agentic risk
Overview
This is a disclosed setup guide for a local home-lab AI router, with normal package-install and trusted-LAN cautions but no hidden or malicious behavior in the skill artifact.
Install this only on machines and a LAN you trust. Verify the PyPI/uv packages before running them, avoid exposing the local endpoint to untrusted networks, and check what ~/.fleet-manager logs contain before sending private prompts, audio, or documents through the cluster.
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
VirusTotal findings are pending for this skill version.
Static analysis
No suspicious patterns detected.