Agent Metacognition
Security checks across malware telemetry and agentic risk
Overview
This skill is a simple in-memory agent self-monitoring and reflection library with no evidence of hidden access, credential use, network activity, or destructive behavior.
Reasonable to install if you want an in-process metacognition helper. Be aware that task context, errors, decisions, and lessons you pass into it may remain in memory during the process, so avoid including secrets in monitored context or reports.
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.