Paper Workflow
Security checks across malware telemetry and agentic risk
Overview
This is a documentation-only academic paper workflow skill with no hidden execution, persistence, credential handling, or destructive behavior found.
Before installing, review the companion-skill installation commands because they fetch external skills from LobeHub and GitHub. The core skill itself appears to be a benign paper-writing workflow, but users should explicitly tell the agent if screenshots must preserve their original language or if project materials contain sensitive unpublished data.
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
57/57 vendors flagged this skill as clean.