Protein QC Strict
Security checks across malware telemetry and agentic risk
Overview
No evidence of hidden, destructive, or malicious behavior was found in the supplied scan context or inspectable workspace artifacts.
This appears acceptable to install or use, but only run workflows that match your role and environment. Staff moderation or review/proof commands can affect real services or expose diffs to configured reviewer tools, so use the documented auth, confirmation, and verification steps.
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.