Agent Loop
Security checks across malware telemetry and agentic risk
Overview
No security issue was identified from the supplied scanner signals, but the actual skill artifact was not available in the workspace for a full coherence review.
This appears acceptable based on the available scan signals, but install with normal caution because the underlying skill files were not available for direct review in this workspace. Review the skill's own files before granting credentials, filesystem access, background execution, or mutation authority.
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
66/66 vendors flagged this skill as clean.