academic-humanization-low-ai
Security checks across malware telemetry and agentic risk
Overview
This is an instruction-only academic rewriting skill with an integrity concern around AI-detection evasion, but it does not install code, access data, use credentials, persist, or perform hidden actions.
Install only if you are comfortable with a skill whose explicit goal is lowering AI-detection signals in academic writing. Follow school, publisher, employer, and platform disclosure rules; use ordinary clarity or style editing instead when transparency about AI assistance is required.
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.