Rv Measure
Security checks across malware telemetry and agentic risk
Overview
This skill is a non-executable placeholder for future AI model analysis functionality and shows no unsafe behavior in the reviewed artifacts.
Installation risk appears low because this version is only documentation, but it also does not provide the advertised measurement capability yet. Review any future version carefully once code, integrations, or model-introspection tooling are added.
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
64/64 vendors flagged this skill as clean.