Humanize Ai Text
Security checks across malware telemetry and agentic risk
Overview
This is a low-permission writing helper with no included executable code, network behavior, credential access, or persistence.
Before installing, treat this as an instruction-only text-editing skill unless the missing scripts are separately supplied and reviewed. Use it on copies or explicit output files, manually review any rewritten text for changed meaning, and follow school, workplace, or platform rules about AI-assisted writing disclosure.
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.