Security audit
Skill 107
Security checks across malware telemetry and agentic risk
Overview
This skill is an educational distributed-systems reference with no executable code or requests for sensitive access.
Safe to install as reference material. Treat the design advice as general guidance and validate any distributed-systems choices against your own production requirements before applying them.
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
65/65 vendors flagged this skill as clean.