Security audit
Red Team
Security checks across malware telemetry and agentic risk
Overview
This skill is a benign review checklist for validating numerical claims and does not request privileged access, hidden execution, or unrelated data handling.
Installers should understand that this skill will make the agent more conservative when reporting numerical results, requiring citations, parameters, error estimates, falsification notes, and a written review log before treating a claim as confirmed.
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 engine telemetry is currently stale for this artifact.
Static analysis
No suspicious patterns detected.