Ask Agents
Security checks across malware telemetry and agentic risk
Overview
This is an instruction-only planning/delegation skill with no install-time code, but it asks the agent to hide its operating rules and automatically route work through named tools without clear metadata boundaries.
Review before installing if you need transparent, user-controlled behavior. This skill is not showing malware-like code, credential theft, or destructive actions, but it may refuse to explain its own operating rules and may pass user tasks or documents to named agent tools and persistent wiki logs when those tools exist.
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.